keyllm_neo4j/KeyBERT_LLM_Neo4j.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "machine_shape": "hm",
      "toc_visible": true,
      "gpuType": "V100",
      "authorship_tag": "ABX9TyM0U+orhmIVfjK5nyQkxaga"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "12dabded7f6d41ad9ecec1944d8c73d2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_df2a1f9134c54f02a21c7e637dd6543b",
              "IPY_MODEL_9421ae1f12524eae9113e1eeb11c5314",
              "IPY_MODEL_cc8f914dbfcd4e2ab072766fbdc0bd82"
            ],
            "layout": "IPY_MODEL_108091daf9344d63b7fc73a3bfe74ed3"
          }
        },
        "df2a1f9134c54f02a21c7e637dd6543b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_87156d6d9e414c78a4ee21b1036d5067",
            "placeholder": "​",
            "style": "IPY_MODEL_8696bae0580244ba917400e214bc0dfa",
            "value": "config.json: 100%"
          }
        },
        "9421ae1f12524eae9113e1eeb11c5314": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_257716a026cd49059434c97cadea2260",
            "max": 1311,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_4034dee11b0b4a8790facfc4004c7fe3",
            "value": 1311
          }
        },
        "cc8f914dbfcd4e2ab072766fbdc0bd82": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_ce0a35b4f6bd43a486621ec27dac66cc",
            "placeholder": "​",
            "style": "IPY_MODEL_4d26315a116f4770b7b5c13edfb028f9",
            "value": " 1.31k/1.31k [00:00&lt;00:00, 104kB/s]"
          }
        },
        "108091daf9344d63b7fc73a3bfe74ed3": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "87156d6d9e414c78a4ee21b1036d5067": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "8696bae0580244ba917400e214bc0dfa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "257716a026cd49059434c97cadea2260": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4034dee11b0b4a8790facfc4004c7fe3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "ce0a35b4f6bd43a486621ec27dac66cc": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4d26315a116f4770b7b5c13edfb028f9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "661d9c66cf0642aba67427cfd37707c4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_306fb9ae96954b45a1614cf988f35ecd",
              "IPY_MODEL_96d1ec221d764185ba151ceeab815727",
              "IPY_MODEL_3619e0dbdfbe445aa7d6f6d17bf333f9"
            ],
            "layout": "IPY_MODEL_83d61a7364f1493f937c47845add381d"
          }
        },
        "306fb9ae96954b45a1614cf988f35ecd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7c819aca768b47ebaf52d036ce53e732",
            "placeholder": "​",
            "style": "IPY_MODEL_2b71451de3144bf7bfe6d62b4c8d8797",
            "value": "model.safetensors: 100%"
          }
        },
        "96d1ec221d764185ba151ceeab815727": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_cb1aa93e47af48a3a764aeae7399129f",
            "max": 4158662096,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_fc14e222156a4299a562ff3faf2e64f0",
            "value": 4158662096
          }
        },
        "3619e0dbdfbe445aa7d6f6d17bf333f9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c65fb308a50b404083ea7dc6a1216f03",
            "placeholder": "​",
            "style": "IPY_MODEL_00da573e5e574315be3f21e8dac6cf78",
            "value": " 4.16G/4.16G [00:19&lt;00:00, 245MB/s]"
          }
        },
        "83d61a7364f1493f937c47845add381d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7c819aca768b47ebaf52d036ce53e732": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "2b71451de3144bf7bfe6d62b4c8d8797": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "cb1aa93e47af48a3a764aeae7399129f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "fc14e222156a4299a562ff3faf2e64f0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "c65fb308a50b404083ea7dc6a1216f03": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "00da573e5e574315be3f21e8dac6cf78": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "33983887f34146e7aa005d7c50f5f0b9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_848e89fa4cb44a299d6e6b86aa5cc33b",
              "IPY_MODEL_110d9498643f452cb7fb18260c3a118b",
              "IPY_MODEL_c0e32f1014454ef1a6334383f5905889"
            ],
            "layout": "IPY_MODEL_0c738517bb9a41eaabd1946a6bf894c7"
          }
        },
        "848e89fa4cb44a299d6e6b86aa5cc33b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_102c5253cfa048faa858c8f503e51c89",
            "placeholder": "​",
            "style": "IPY_MODEL_68abe2f2133b4ab5b55e8a59f84c54de",
            "value": "generation_config.json: 100%"
          }
        },
        "110d9498643f452cb7fb18260c3a118b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d00d0561f9d54e73b9702bbd72a6f165",
            "max": 116,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_8c365f0d3dd3447ab03cf54834730d6a",
            "value": 116
          }
        },
        "c0e32f1014454ef1a6334383f5905889": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7af4c01aa4304fc8ba0440bb97521be2",
            "placeholder": "​",
            "style": "IPY_MODEL_5ffef66c0111434dbaf872d5bbf64b9e",
            "value": " 116/116 [00:00&lt;00:00, 9.83kB/s]"
          }
        },
        "0c738517bb9a41eaabd1946a6bf894c7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "102c5253cfa048faa858c8f503e51c89": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "68abe2f2133b4ab5b55e8a59f84c54de": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "d00d0561f9d54e73b9702bbd72a6f165": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "8c365f0d3dd3447ab03cf54834730d6a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "7af4c01aa4304fc8ba0440bb97521be2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5ffef66c0111434dbaf872d5bbf64b9e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "2135c9b454d743468a1bb11a3b9c4ece": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_891dbe4633894efd8ea9d7799b86850a",
              "IPY_MODEL_63c3c67b5bb34b06bb6995a4e00d573a",
              "IPY_MODEL_de1fa6516272473481a04c1460bd787b"
            ],
            "layout": "IPY_MODEL_9b5c888e194d4dabaaf98b3bb6aa9dbe"
          }
        },
        "891dbe4633894efd8ea9d7799b86850a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_4ff4962dfa1e452ba13dcce1fea62af8",
            "placeholder": "​",
            "style": "IPY_MODEL_81ea698eef584572b78cf0279ea2af91",
            "value": "tokenizer_config.json: 100%"
          }
        },
        "63c3c67b5bb34b06bb6995a4e00d573a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9bc557ce09bf42d38fa6b04c3b63ebf4",
            "max": 1431,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_e477276bd15d4239ad54f5e160a02cad",
            "value": 1431
          }
        },
        "de1fa6516272473481a04c1460bd787b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_8b1b4a32a86d4c168875af5730766158",
            "placeholder": "​",
            "style": "IPY_MODEL_ba103a9733ec4692b0e449cc5d2f3543",
            "value": " 1.43k/1.43k [00:00&lt;00:00, 111kB/s]"
          }
        },
        "9b5c888e194d4dabaaf98b3bb6aa9dbe": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4ff4962dfa1e452ba13dcce1fea62af8": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "81ea698eef584572b78cf0279ea2af91": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "9bc557ce09bf42d38fa6b04c3b63ebf4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e477276bd15d4239ad54f5e160a02cad": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "8b1b4a32a86d4c168875af5730766158": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "ba103a9733ec4692b0e449cc5d2f3543": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "4e15900e1c8942ec9f8ff3c7667b2100": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_2af42d5009674deb91aa4e656b7f3cd9",
              "IPY_MODEL_1dbf4f3e94ac499d92e2cc4d7fd2a774",
              "IPY_MODEL_8112f3d443de49959cdebb9462200d60"
            ],
            "layout": "IPY_MODEL_1bc2dbcb07744b5d83833246a3e774fc"
          }
        },
        "2af42d5009674deb91aa4e656b7f3cd9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b9f634201ac14f2ab5f04d06b2872099",
            "placeholder": "​",
            "style": "IPY_MODEL_35ed80cc60514481996877729620d07f",
            "value": "tokenizer.model: 100%"
          }
        },
        "1dbf4f3e94ac499d92e2cc4d7fd2a774": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_ef113aeef8f2499bb5f7dc66becc855a",
            "max": 493443,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_2f2a50b0512146e39b62849fee097355",
            "value": 493443
          }
        },
        "8112f3d443de49959cdebb9462200d60": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d384e87e49c74ff4b862f01e8656e5ec",
            "placeholder": "​",
            "style": "IPY_MODEL_db2d9987ffd04f568b362d0a837e246d",
            "value": " 493k/493k [00:00&lt;00:00, 25.6MB/s]"
          }
        },
        "1bc2dbcb07744b5d83833246a3e774fc": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b9f634201ac14f2ab5f04d06b2872099": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "35ed80cc60514481996877729620d07f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "ef113aeef8f2499bb5f7dc66becc855a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "2f2a50b0512146e39b62849fee097355": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "d384e87e49c74ff4b862f01e8656e5ec": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "db2d9987ffd04f568b362d0a837e246d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "e021cbfbcf54451e8d765b7b52f2403a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_2eb5cb52b45040bcbbb291e84d443ee7",
              "IPY_MODEL_fbfca3ff40b54d65be8dd8bf0d566715",
              "IPY_MODEL_043e7bb7ffe84a88bb409f8c38bebbe5"
            ],
            "layout": "IPY_MODEL_c18f1a66e5844a0bae0c829438fe4263"
          }
        },
        "2eb5cb52b45040bcbbb291e84d443ee7": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d603d2ff4c0d4b8992d6f666e78c4616",
            "placeholder": "​",
            "style": "IPY_MODEL_afb5fe00ce1945b08154787b688a31fc",
            "value": "tokenizer.json: 100%"
          }
        },
        "fbfca3ff40b54d65be8dd8bf0d566715": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_815aa3413f5d46dbb4fac8b3eec6d37a",
            "max": 1795303,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_5f871fa96c414651a82a185323519ebb",
            "value": 1795303
          }
        },
        "043e7bb7ffe84a88bb409f8c38bebbe5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_a201ed98223447d1a1881013eb00d563",
            "placeholder": "​",
            "style": "IPY_MODEL_754ede0ecb05407e833c9815fe9d5e53",
            "value": " 1.80M/1.80M [00:00&lt;00:00, 4.40MB/s]"
          }
        },
        "c18f1a66e5844a0bae0c829438fe4263": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "d603d2ff4c0d4b8992d6f666e78c4616": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "afb5fe00ce1945b08154787b688a31fc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "815aa3413f5d46dbb4fac8b3eec6d37a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5f871fa96c414651a82a185323519ebb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "a201ed98223447d1a1881013eb00d563": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "754ede0ecb05407e833c9815fe9d5e53": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "a939bedc2c1245cab4bb3dbd754e4db2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_4b2809ae6f35457fb1bdd6bd2570882f",
              "IPY_MODEL_b15fe858a2364f7fb0a7f5cf12433289",
              "IPY_MODEL_572ffcc98f5544cbb8007dea80266fec"
            ],
            "layout": "IPY_MODEL_a0b9a2d3cdad495ea618274f3ffe0778"
          }
        },
        "4b2809ae6f35457fb1bdd6bd2570882f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_77fe19ad689046f48b29cafe17e805f5",
            "placeholder": "​",
            "style": "IPY_MODEL_cbdc89df92694153bec593a2e0081227",
            "value": "special_tokens_map.json: 100%"
          }
        },
        "b15fe858a2364f7fb0a7f5cf12433289": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_defcc9177a5a40b69ef693d80dbefffa",
            "max": 168,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_8aa1f4efb7a94aa9a9151114f1a328c6",
            "value": 168
          }
        },
        "572ffcc98f5544cbb8007dea80266fec": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5934a2a1eeab425ab918e24a26becd2e",
            "placeholder": "​",
            "style": "IPY_MODEL_0f0823d746c94787a7a181b7fb2985b2",
            "value": " 168/168 [00:00&lt;00:00, 16.0kB/s]"
          }
        },
        "a0b9a2d3cdad495ea618274f3ffe0778": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "77fe19ad689046f48b29cafe17e805f5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "cbdc89df92694153bec593a2e0081227": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "defcc9177a5a40b69ef693d80dbefffa": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "8aa1f4efb7a94aa9a9151114f1a328c6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "5934a2a1eeab425ab918e24a26becd2e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0f0823d746c94787a7a181b7fb2985b2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# KeyBERT -> KeyLLM -> HDBSCAN-> Neo4j"
      ],
      "metadata": {
        "id": "OSq2V0m-RrdC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Colab Setup\n",
        "\n",
        "Set the Google Colab to access files on the drive."
      ],
      "metadata": {
        "id": "1V8TaWgVVDS-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Load and mount the drive helper\n",
        "from google.colab import drive\n",
        "\n",
        "# This will prompt for authorization\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "# Set the working directory\n",
        "%cd '/content/drive/MyDrive/keyLLM/'"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ic__w8jpUyoZ",
        "outputId": "f37413f9-d6c9-435e-ad08-bca01a17965f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n",
            "/content/drive/MyDrive/keyLLM\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a path variable for the data folder\n",
        "data_path = '/content/drive/MyDrive/keyLLM/datas/'\n",
        "\n",
        "# Created by the data parser, selected articles\n",
        "topic = 'cs'\n",
        "extracted_data_file = f'selected_{topic}.csv'\n",
        "\n",
        "# Data and the extracted keywords\n",
        "extracted_keys_file = 'selected_cs_keys.csv'\n",
        "\n",
        "# Articles with 5 keywords/keyphrases\n",
        "parsed_keys_file = 'parsed_cs_keys.csv'\n",
        "\n",
        "# Pickled data containing 768-dim and 10-dim embeddings, and clusters\n",
        "embeddings_clusters_keys_file = 'embedddings_clusters_cs_keys.pkl'\n",
        "\n",
        "# Descriptions and labels for keywords clusters\n",
        "descriptions_labels_keys_file = 'descriptions_labels.csv'"
      ],
      "metadata": {
        "id": "IsOJukqXa5FE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Data Load & Preparation\n"
      ],
      "metadata": {
        "id": "kU7XnCqtU6i4"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Download the data from [Kaggle ArXiv Site](https://www.kaggle.com/datasets/Cornell-University/arxiv) to the specified data_path. This will download a single file `archive.zip` (1.2 GB). After unzipping a single file `archive-metadata-oai-snapshot.json` (3.7 GB) is extracted."
      ],
      "metadata": {
        "id": "Nzej78h6a-NN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Import the data parser module\n",
        "from utils.arxiv_parser import *\n",
        "\n",
        "# Initialize the data parser\n",
        "parser = ArXivDataProcessor(data_path)"
      ],
      "metadata": {
        "id": "9EqRszHSptnO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Unzip the file and extract a json file in data_path\n",
        "parser.unzip_file()"
      ],
      "metadata": {
        "id": "D7ctGVzrZfwW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Select a topic and extract the articles on that topic\n",
        "entries = parser.select_topic(\"cs\")\n",
        "# Check the output for one article\n",
        "entries[111]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KaBLNZD6Yz2D",
        "outputId": "fa3b230f-7c29-4195-b9cb-3da87ab654f7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "There are 7149 with cs topic.\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'id': '0003067',\n",
              " 'submitter': 'Maurice Bruynooghe',\n",
              " 'authors': 'Maurice Bruynooghe (1), Henk Vandecasteele (1), D. Andre de Waal (2),\\n  Marc Denecker (1) ((1) Katholieke Universiteit Leuven, Belgium, (2)\\n  Potchefstroom University for Christian Higher Education, South Africa)',\n",
              " 'title': 'Detecting Unsolvable Queries for Definite Logic Programs',\n",
              " 'comments': '32 pages including appendix. A preliminary version appeared in\\n  proceedings PLILP/ALP98 (Springer LNCS 1490) This version, without appendix\\n  appeared in Journal Functional and Logic Programming 1999',\n",
              " 'journal-ref': 'Journal of Functional and Logic Programming, Vol. 1999, 1-35, 1999',\n",
              " 'doi': None,\n",
              " 'report-no': None,\n",
              " 'categories': 'cs.LO cs.AI',\n",
              " 'license': None,\n",
              " 'abstract': \"  In solving a query, the SLD proof procedure for definite programs sometimes\\nsearches an infinite space for a non existing solution. For example, querying a\\nplanner for an unreachable goal state. Such programs motivate the development\\nof methods to prove the absence of a solution. Considering the definite program\\nand the query ``<- Q'' as clauses of a first order theory, one can apply model\\ngenerators which search for a finite interpretation in which the program\\nclauses as well as the clause ``false <- Q'' are true. This paper develops a\\nnew approach which exploits the fact that all clauses are definite. It is based\\non a goal directed abductive search in the space of finite pre-interpretations\\nfor a pre-interpretation such that ``Q'' is false in the least model of the\\nprogram based on it. Several methods for efficiently searching the space of\\npre-interpretations are presented. Experimental results confirm that our\\napproach find solutions with less search than with the use of a first order\\nmodel generator.\\n\",\n",
              " 'versions': [{'version': 'v1', 'created': 'Fri, 17 Mar 2000 10:59:03 GMT'}],\n",
              " 'update_date': '2007-05-23',\n",
              " 'authors_parsed': [['Bruynooghe', 'Maurice', ''],\n",
              "  ['Vandecasteele', 'Henk', ''],\n",
              "  ['de Waal', 'D. Andre', ''],\n",
              "  ['Denecker', 'Marc', '']]}"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Comments**\n",
        "\n",
        "We'll chose only articles that have abstracts with token length in our interval of choice. We also have the option to retain the features that are relevant for the project and discard the remaining ones. If no columns are specified all the features will be retained.\n",
        "\n",
        "To keep track of the number of tokens in the abstract, a new column `abs_length` is added. There is an option to keep that column, in case we want to gain further insight into the length distribution of the abstracts.\n",
        "\n",
        "There is also an option to build a corpus which corresponds to the concatenation of the title and the abstract."
      ],
      "metadata": {
        "id": "uhP2mF9VkeCR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Retain only the selected data\n",
        "df = parser.select_articles(entries, # extracted articles\n",
        "                            cols=['id', 'title', 'abstract'], # features to keep\n",
        "                            min_length = 100, # min tokens an abstract should have\n",
        "                            max_length = 120, # max tokens an abstract should have\n",
        "                            keep_abs_length = False, # do not keep the abs_length column\n",
        "                            build_corpus=False) # do not build a corpus column\n",
        "# Inspect the data\n",
        "df.head(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "id": "lrzXCLIwjRlF",
        "outputId": "d0eb344a-068e-46a6-cd24-0e0796c41ebc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "There are 983 articles selected.\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "        id                                              title  \\\n",
              "0  0001001  Von Neumann Quantum Logic vs. Classical von Ne...   \n",
              "1  0001002    Minimum Description Length and Compositionality   \n",
              "\n",
              "                                            abstract  \n",
              "0    The name of John von Neumann is common both ...  \n",
              "1    We present a non-vacuous definition of compo...  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-4db9f7e0-99a0-478b-8a73-f2b3f120590c\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>title</th>\n",
              "      <th>abstract</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0001001</td>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Ne...</td>\n",
              "      <td>The name of John von Neumann is common both ...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0001002</td>\n",
              "      <td>Minimum Description Length and Compositionality</td>\n",
              "      <td>We present a non-vacuous definition of compo...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4db9f7e0-99a0-478b-8a73-f2b3f120590c')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-4db9f7e0-99a0-478b-8a73-f2b3f120590c button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-4db9f7e0-99a0-478b-8a73-f2b3f120590c');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-321da6c2-6641-4b4e-a6b4-8ed6c4745861\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-321da6c2-6641-4b4e-a6b4-8ed6c4745861')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-321da6c2-6641-4b4e-a6b4-8ed6c4745861 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save the selected data to a csv file 'selected_{topic}.csv', uses data_path\n",
        "parser.save_selected_data(df, 'cs')"
      ],
      "metadata": {
        "id": "P-TmKNxckXr5"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## KeyBERT & LLM to Extract Keywords"
      ],
      "metadata": {
        "id": "XLjzw-hKlUAc"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Required Installs KeyBERT & LLM & SentenceTransformers"
      ],
      "metadata": {
        "id": "GlnhZBAUdCe4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Install the required packages\n",
        "\n",
        "!pip install transformers optimum accelerate\n",
        "!pip install auto-gptq --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/\n",
        "!pip install keybert\n",
        "!pip install -U sentence-transformers"
      ],
      "metadata": {
        "id": "Xd18Ka-rrL_L"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Load the LLM and the Tokenizer"
      ],
      "metadata": {
        "id": "bltUlsc5VJps"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline\n",
        "\n",
        "# Load the model and tokenizer\n",
        "model_name_or_path = \"TheBloke/zephyr-7B-beta-GPTQ\"\n",
        "\n",
        "# To use a different branch, change revision\n",
        "llm = AutoModelForCausalLM.from_pretrained(model_name_or_path,\n",
        "                                             device_map=\"auto\",\n",
        "                                             trust_remote_code=False,\n",
        "                                             revision=\"main\")\n",
        "\n",
        "tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 278,
          "referenced_widgets": [
            "12dabded7f6d41ad9ecec1944d8c73d2",
            "df2a1f9134c54f02a21c7e637dd6543b",
            "9421ae1f12524eae9113e1eeb11c5314",
            "cc8f914dbfcd4e2ab072766fbdc0bd82",
            "108091daf9344d63b7fc73a3bfe74ed3",
            "87156d6d9e414c78a4ee21b1036d5067",
            "8696bae0580244ba917400e214bc0dfa",
            "257716a026cd49059434c97cadea2260",
            "4034dee11b0b4a8790facfc4004c7fe3",
            "ce0a35b4f6bd43a486621ec27dac66cc",
            "4d26315a116f4770b7b5c13edfb028f9",
            "661d9c66cf0642aba67427cfd37707c4",
            "306fb9ae96954b45a1614cf988f35ecd",
            "96d1ec221d764185ba151ceeab815727",
            "3619e0dbdfbe445aa7d6f6d17bf333f9",
            "83d61a7364f1493f937c47845add381d",
            "7c819aca768b47ebaf52d036ce53e732",
            "2b71451de3144bf7bfe6d62b4c8d8797",
            "cb1aa93e47af48a3a764aeae7399129f",
            "fc14e222156a4299a562ff3faf2e64f0",
            "c65fb308a50b404083ea7dc6a1216f03",
            "00da573e5e574315be3f21e8dac6cf78",
            "33983887f34146e7aa005d7c50f5f0b9",
            "848e89fa4cb44a299d6e6b86aa5cc33b",
            "110d9498643f452cb7fb18260c3a118b",
            "c0e32f1014454ef1a6334383f5905889",
            "0c738517bb9a41eaabd1946a6bf894c7",
            "102c5253cfa048faa858c8f503e51c89",
            "68abe2f2133b4ab5b55e8a59f84c54de",
            "d00d0561f9d54e73b9702bbd72a6f165",
            "8c365f0d3dd3447ab03cf54834730d6a",
            "7af4c01aa4304fc8ba0440bb97521be2",
            "5ffef66c0111434dbaf872d5bbf64b9e",
            "2135c9b454d743468a1bb11a3b9c4ece",
            "891dbe4633894efd8ea9d7799b86850a",
            "63c3c67b5bb34b06bb6995a4e00d573a",
            "de1fa6516272473481a04c1460bd787b",
            "9b5c888e194d4dabaaf98b3bb6aa9dbe",
            "4ff4962dfa1e452ba13dcce1fea62af8",
            "81ea698eef584572b78cf0279ea2af91",
            "9bc557ce09bf42d38fa6b04c3b63ebf4",
            "e477276bd15d4239ad54f5e160a02cad",
            "8b1b4a32a86d4c168875af5730766158",
            "ba103a9733ec4692b0e449cc5d2f3543",
            "4e15900e1c8942ec9f8ff3c7667b2100",
            "2af42d5009674deb91aa4e656b7f3cd9",
            "1dbf4f3e94ac499d92e2cc4d7fd2a774",
            "8112f3d443de49959cdebb9462200d60",
            "1bc2dbcb07744b5d83833246a3e774fc",
            "b9f634201ac14f2ab5f04d06b2872099",
            "35ed80cc60514481996877729620d07f",
            "ef113aeef8f2499bb5f7dc66becc855a",
            "2f2a50b0512146e39b62849fee097355",
            "d384e87e49c74ff4b862f01e8656e5ec",
            "db2d9987ffd04f568b362d0a837e246d",
            "e021cbfbcf54451e8d765b7b52f2403a",
            "2eb5cb52b45040bcbbb291e84d443ee7",
            "fbfca3ff40b54d65be8dd8bf0d566715",
            "043e7bb7ffe84a88bb409f8c38bebbe5",
            "c18f1a66e5844a0bae0c829438fe4263",
            "d603d2ff4c0d4b8992d6f666e78c4616",
            "afb5fe00ce1945b08154787b688a31fc",
            "815aa3413f5d46dbb4fac8b3eec6d37a",
            "5f871fa96c414651a82a185323519ebb",
            "a201ed98223447d1a1881013eb00d563",
            "754ede0ecb05407e833c9815fe9d5e53",
            "a939bedc2c1245cab4bb3dbd754e4db2",
            "4b2809ae6f35457fb1bdd6bd2570882f",
            "b15fe858a2364f7fb0a7f5cf12433289",
            "572ffcc98f5544cbb8007dea80266fec",
            "a0b9a2d3cdad495ea618274f3ffe0778",
            "77fe19ad689046f48b29cafe17e805f5",
            "cbdc89df92694153bec593a2e0081227",
            "defcc9177a5a40b69ef693d80dbefffa",
            "8aa1f4efb7a94aa9a9151114f1a328c6",
            "5934a2a1eeab425ab918e24a26becd2e",
            "0f0823d746c94787a7a181b7fb2985b2"
          ]
        },
        "id": "pmo3zNzoUt1d",
        "outputId": "b2cd9a80-f49a-4a84-caf5-beb30fb0dd8f"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "config.json:   0%|          | 0.00/1.31k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "12dabded7f6d41ad9ecec1944d8c73d2"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Using `disable_exllama` is deprecated and will be removed in version 4.37. Use `use_exllama` instead and specify the version with `exllama_config`.The value of `use_exllama` will be overwritten by `disable_exllama` passed in `GPTQConfig` or stored in your config file.\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "model.safetensors:   0%|          | 0.00/4.16G [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "661d9c66cf0642aba67427cfd37707c4"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "generation_config.json:   0%|          | 0.00/116 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "33983887f34146e7aa005d7c50f5f0b9"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer_config.json:   0%|          | 0.00/1.43k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "2135c9b454d743468a1bb11a3b9c4ece"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer.model:   0%|          | 0.00/493k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "4e15900e1c8942ec9f8ff3c7667b2100"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer.json:   0%|          | 0.00/1.80M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "e021cbfbcf54451e8d765b7b52f2403a"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "special_tokens_map.json:   0%|          | 0.00/168 [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "a939bedc2c1245cab4bb3dbd754e4db2"
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Build the text generation pipeline\n",
        "generator = pipeline(\n",
        "    model=llm,\n",
        "    tokenizer=tokenizer,\n",
        "    task='text-generation',\n",
        "    max_new_tokens=50,\n",
        "    repetition_penalty=1.1,\n",
        ")"
      ],
      "metadata": {
        "id": "S5helstyXZII"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Instantiate KeyBERT"
      ],
      "metadata": {
        "id": "y6RjL-NtVRar"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The prompt is very important for the quality of the text generation. Experimented with variations such as using technical terms from computer science or asking to remove the stopwords. Both these requests make the model to add all sort of comments and opinions! This format generates the cleanest output I was able to get."
      ],
      "metadata": {
        "id": "UkycGG6cq25H"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a custom prompt, use the format from the HF model card\n",
        "# https://huggingface.co/HuggingFaceH4/zephyr-7b-beta\n",
        "\n",
        "prompt_keywords= \"\"\"\n",
        "<|system|>\n",
        "I have the following document:\n",
        "Semantics and Termination of Simply-Moded Logic Programs with Dynamic Scheduling\n",
        "and five candidate keywords:\n",
        "scheduling, logic, semantics, termination, moded\n",
        "\n",
        "Based on the information above, extract the keywords or the keyphrases that best describe the topic of the text.\n",
        "Follow the requirements below:\n",
        "1. Make sure to extract only the keywords or keyphrases that appear in the text.\n",
        "2. Provide five keywords or keyphrases! Do not number or label the keywords or the keyphrases!\n",
        "3. Do not include anything else besides the keywords or the keyphrases! I repeat do not include any comments!\n",
        "\n",
        "semantics, termination, simply-moded, logic programs, dynamic scheduling</s>\n",
        "\n",
        "<|user|>\n",
        "I have the following document:\n",
        "[DOCUMENT]\n",
        "and five candidate keywords:\n",
        "[CANDIDATES]\n",
        "\n",
        "Based on the information above, extract the keywords or the keyphrases that best describe the topic of the text.\n",
        "Follow the requirements below:\n",
        "1. Make sure to extract only the keywords or keyphrases that appear in the text.\n",
        "2. Provide five keywords or keyphrases! Do not number or label the keywords or the keyphrases!\n",
        "3. Do not include anything else besides the keywords or the keyphrases! I repeat do not include any comments!</s>\n",
        "\n",
        "<|assistant|>\n",
        "\"\"\""
      ],
      "metadata": {
        "id": "PcWukpFp23vP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from keybert.llm import TextGeneration\n",
        "from keybert import KeyLLM, KeyBERT\n",
        "from sentence_transformers import SentenceTransformer\n",
        "\n",
        "# Instantiate KeyBert TextGeneration pipeline wrapper\n",
        "# source: https://github.com/MaartenGr/KeyBERT/blob/master/keybert/llm/_textgeneration.py\n",
        "\n",
        "llm_tg = TextGeneration(generator, prompt=prompt_keywords)\n",
        "\n",
        "# Specify the embedding model: cluster similar docs and extracts keywords from community center document\n",
        "# Based on: https://github.com/UKPLab/sentence-transformers/blob/master/sentence_transformers/util.py\n",
        "\n",
        "kw_model= KeyBERT(llm=llm_tg, model = \"all-mpnet-base-v2\")"
      ],
      "metadata": {
        "id": "wXx_tXnawsxo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Extract Keywords with KeyBERT & LLM"
      ],
      "metadata": {
        "id": "2bmkwf5SWDLk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Function to process documents in batches\n",
        "\n",
        "def process_in_batches(corpus,\n",
        "                       batch_size):\n",
        "    results = []\n",
        "    for i in range(0, len(corpus), batch_size):\n",
        "        batch = corpus[i:i+batch_size]\n",
        "        keys = kw_model.extract_keywords(batch,\n",
        "                                         threshold=0.5)\n",
        "        results.extend(keys)\n",
        "    return results"
      ],
      "metadata": {
        "id": "rUV0COYt53uW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Load the dataset if necessary\n",
        "df = pd.read_csv(data_path+extracted_data_file)\n",
        "\n",
        "# Retain the articles titles only for analysis\n",
        "titles_list = df.title.tolist()\n",
        "\n",
        "# Process the documents and collect the results\n",
        "titles_keys = process_in_batches(titles_list,\n",
        "                                 batch_size = 100)\n",
        "\n",
        "# Add the results to df\n",
        "df[\"titles_keys\"] = titles_keys\n",
        "\n",
        "# Save the data and the results to a file\n",
        "df.to_csv(data_path+extracted_keys_file, index=False)"
      ],
      "metadata": {
        "id": "5_-uP3x07tUs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Take a look at the output\n",
        "pd.set_option('display.max_colwidth', None)\n",
        "df.head(10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 936
        },
        "id": "fmXTNgKAgHNH",
        "outputId": "6f8305dd-434b-4209-f9c9-447480c69dd9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     id                                                              title  \\\n",
              "0  1001  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "1  1002                    Minimum Description Length and Compositionality   \n",
              "2  1005                Effect of different packet sizes on RED performance   \n",
              "3  1007                           RED behavior with different packet sizes   \n",
              "4  1026                        A Logic for SDSI's Linked Local Name Spaces   \n",
              "5  2006                 Multiplicative Nonholonomic/Newton -like Algorithm   \n",
              "6  2007                           Requirements of Text Processing Lexicons   \n",
              "7  2014                          Safe cooperative robot dynamics on graphs   \n",
              "8  3003             Prospects for in-depth story understanding by computer   \n",
              "9  3009              Conditional indifference and conditional preservation   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    abstract  \\\n",
              "0                                                                                                                                                                                                                    The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n   \n",
              "1                                                         We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n   \n",
              "2                                                                                                                                                                                          We consider the adaptation of random early detection (RED) as an active queue\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two existing RED\\nvariants and point out a weakness in both. The first variant where the drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high Packet Loss Ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n   \n",
              "3                                                                                                                                                                                             We consider the adaptation of random early detection (RED) as a buffer\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two RED variants\\ndescribed in [4] and point out a weakness in both. The first variant where drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high packet loss ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n   \n",
              "4                                                        Abadi has introduced a logic to explicate the meaning of local names in SDSI,\\nthe Simple Distributed Security Infrastructure proposed by Rivest and Lampson.\\nAbadi's logic does not correspond precisely to SDSI, however; it draws\\nconclusions about local names that do not follow from SDSI's name resolution\\nalgorithm. Moreover, its semantics is somewhat unintuitive. This paper presents\\nthe Logic of Local Name Containment, which does not suffer from these\\ndeficiencies. It has a clear semantics and provides a tight characterization of\\nSDSI name resolution. The semantics is shown to be closely related to that of\\nlogic programs, leading to an approach to the efficient implementation of\\nqueries concerning local names. A complete axiomatization of the logic is also\\nprovided.\\n   \n",
              "5                                                                                                                                                               We construct new algorithms from scratch, which use the fourth order cumulant\\nof stochastic variables for the cost function. The multiplicative updating rule\\nhere constructed is natural from the homogeneous nature of the Lie group and\\nhas numerous merits for the rigorous treatment of the dynamics. As one\\nconsequence, the second order convergence is shown. For the cost function,\\nfunctions invariant under the componentwise scaling are choosen. By identifying\\npoints which can be transformed to each other by the scaling, we assume that\\nthe dynamics is in a coset space. In our method, a point can move toward any\\ndirection in this coset. Thus, no prewhitening is required.\\n   \n",
              "6    As text processing systems expand in scope, they will require ever larger\\nlexicons along with a parsing capability for discriminating among many senses\\nof a word. Existing systems do not incorporate such subtleties in meaning for\\ntheir lexicons. Ordinary dictionaries contain such information, but are largely\\nuntapped. When the contents of dictionaries are scrutinized, they reveal many\\nrequirements that must be satisfied in representing meaning and in developing\\nsemantic parsers. These requirements were identified in research designed to\\nfind primitive verb concepts. The requirements are outlined and general\\nprocedures for satisfying them through the use of ordinary dictionaries are\\ndescribed, illustrated by building frames for and examining the definitions of\\n\"change\" and its uses as a hypernym in other definitions.\\n   \n",
              "7                                                                                                                                                               This paper initiates the use of vector fields to design, optimize, and\\nimplement reactive schedules for safe cooperative robot patterns on planar\\ngraphs. We consider Automated Guided Vehicles (AGV's) operating upon a\\npredefined network of pathways. In contrast to the case of locally Euclidean\\nconfiguration spaces, regularization of collisions is no longer a local\\nprocedure, and issues concerning the global topology of configuration spaces\\nmust be addressed. The focus of the present inquiry is the achievement of safe,\\nefficient, cooperative patterns in the simplest nontrivial example (a pair of\\nrobots on a Y-network) by means of a state-event heirarchical controller.\\n   \n",
              "8                                                                                                                                                                         While much research on the hard problem of in-depth story understanding by\\ncomputer was performed starting in the 1970s, interest shifted in the 1990s to\\ninformation extraction and word sense disambiguation. Now that a degree of\\nsuccess has been achieved on these easier problems, I propose it is time to\\nreturn to in-depth story understanding. In this paper I examine the shift away\\nfrom story understanding, discuss some of the major problems in building a\\nstory understanding system, present some possible solutions involving a set of\\ninteracting understanding agents, and provide pointers to useful tools and\\nresources for building story understanding systems.\\n   \n",
              "9                                                                                  The idea of preserving conditional beliefs emerged recently as a new paradigm\\napt to guide the revision of epistemic states. Conditionals are substantially\\ndifferent from propositional beliefs and need specific treatment. In this\\npaper, we present a new approach to conditionals, capturing particularly well\\ntheir dynamic part as revision policies. We thoroughly axiomatize a principle\\nof conditional preservation as an indifference property with respect to\\nconditional structures of worlds. This principle is developed in a\\nsemi-quantitative setting, so as to reveal its fundamental meaning for belief\\nrevision in quantitative as well as in qualitative frameworks. In fact, it is\\nshown to cover other proposed approaches to conditional preservation.\\n   \n",
              "\n",
              "                                                                                                                                                                                                                             titles_keys  \n",
              "0                                                                                                                                                       [von neumann, quantum logic, classical von neumann architecture, quantum, logic]  \n",
              "1                                                                                                     [minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]  \n",
              "2                                                                                                                                                           [packet sizes, RED performance, different packet sizes, effect, performance]  \n",
              "3                                                                                                                                                           [packet sizes, RED performance, different packet sizes, effect, performance]  \n",
              "4                                                                                                                                                                             [SDSI, linked local names, logic, LLN spaces, sdsi-graphs]  \n",
              "5  [multiplicative, nonholonomic, algorithm, Newton-like, multiplicative\\n\\nNote: The second keyword \"algorithm\" is repeated as it appears twice in the title. However, this is still within the requirement of providing five keywords]  \n",
              "6                                                                                  [text processing lexicons, lexicon requirements, natural language processing lexicons, lexical resources for text processing, text processing lexica]  \n",
              "7                                                                                                                                  [cooperative robots, graph theory, safe dynamics, robot dynamics, networks (as a synonym for graphs)]  \n",
              "8                                                                                                                                                                                     [story, understanding, depth, computer, prospects]  \n",
              "9                                                                                                                                       [conditional indifference, conditional preservation, indifference, preservation, conditionality]  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-02c1ec61-1090-4c28-b774-74e97fcba868\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>title</th>\n",
              "      <th>abstract</th>\n",
              "      <th>titles_keys</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1001</td>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n</td>\n",
              "      <td>[von neumann, quantum logic, classical von neumann architecture, quantum, logic]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1002</td>\n",
              "      <td>Minimum Description Length and Compositionality</td>\n",
              "      <td>We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n</td>\n",
              "      <td>[minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1005</td>\n",
              "      <td>Effect of different packet sizes on RED performance</td>\n",
              "      <td>We consider the adaptation of random early detection (RED) as an active queue\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two existing RED\\nvariants and point out a weakness in both. The first variant where the drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high Packet Loss Ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n</td>\n",
              "      <td>[packet sizes, RED performance, different packet sizes, effect, performance]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1007</td>\n",
              "      <td>RED behavior with different packet sizes</td>\n",
              "      <td>We consider the adaptation of random early detection (RED) as a buffer\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two RED variants\\ndescribed in [4] and point out a weakness in both. The first variant where drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high packet loss ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n</td>\n",
              "      <td>[packet sizes, RED performance, different packet sizes, effect, performance]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1026</td>\n",
              "      <td>A Logic for SDSI's Linked Local Name Spaces</td>\n",
              "      <td>Abadi has introduced a logic to explicate the meaning of local names in SDSI,\\nthe Simple Distributed Security Infrastructure proposed by Rivest and Lampson.\\nAbadi's logic does not correspond precisely to SDSI, however; it draws\\nconclusions about local names that do not follow from SDSI's name resolution\\nalgorithm. Moreover, its semantics is somewhat unintuitive. This paper presents\\nthe Logic of Local Name Containment, which does not suffer from these\\ndeficiencies. It has a clear semantics and provides a tight characterization of\\nSDSI name resolution. The semantics is shown to be closely related to that of\\nlogic programs, leading to an approach to the efficient implementation of\\nqueries concerning local names. A complete axiomatization of the logic is also\\nprovided.\\n</td>\n",
              "      <td>[SDSI, linked local names, logic, LLN spaces, sdsi-graphs]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>2006</td>\n",
              "      <td>Multiplicative Nonholonomic/Newton -like Algorithm</td>\n",
              "      <td>We construct new algorithms from scratch, which use the fourth order cumulant\\nof stochastic variables for the cost function. The multiplicative updating rule\\nhere constructed is natural from the homogeneous nature of the Lie group and\\nhas numerous merits for the rigorous treatment of the dynamics. As one\\nconsequence, the second order convergence is shown. For the cost function,\\nfunctions invariant under the componentwise scaling are choosen. By identifying\\npoints which can be transformed to each other by the scaling, we assume that\\nthe dynamics is in a coset space. In our method, a point can move toward any\\ndirection in this coset. Thus, no prewhitening is required.\\n</td>\n",
              "      <td>[multiplicative, nonholonomic, algorithm, Newton-like, multiplicative\\n\\nNote: The second keyword \"algorithm\" is repeated as it appears twice in the title. However, this is still within the requirement of providing five keywords]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>2007</td>\n",
              "      <td>Requirements of Text Processing Lexicons</td>\n",
              "      <td>As text processing systems expand in scope, they will require ever larger\\nlexicons along with a parsing capability for discriminating among many senses\\nof a word. Existing systems do not incorporate such subtleties in meaning for\\ntheir lexicons. Ordinary dictionaries contain such information, but are largely\\nuntapped. When the contents of dictionaries are scrutinized, they reveal many\\nrequirements that must be satisfied in representing meaning and in developing\\nsemantic parsers. These requirements were identified in research designed to\\nfind primitive verb concepts. The requirements are outlined and general\\nprocedures for satisfying them through the use of ordinary dictionaries are\\ndescribed, illustrated by building frames for and examining the definitions of\\n\"change\" and its uses as a hypernym in other definitions.\\n</td>\n",
              "      <td>[text processing lexicons, lexicon requirements, natural language processing lexicons, lexical resources for text processing, text processing lexica]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>2014</td>\n",
              "      <td>Safe cooperative robot dynamics on graphs</td>\n",
              "      <td>This paper initiates the use of vector fields to design, optimize, and\\nimplement reactive schedules for safe cooperative robot patterns on planar\\ngraphs. We consider Automated Guided Vehicles (AGV's) operating upon a\\npredefined network of pathways. In contrast to the case of locally Euclidean\\nconfiguration spaces, regularization of collisions is no longer a local\\nprocedure, and issues concerning the global topology of configuration spaces\\nmust be addressed. The focus of the present inquiry is the achievement of safe,\\nefficient, cooperative patterns in the simplest nontrivial example (a pair of\\nrobots on a Y-network) by means of a state-event heirarchical controller.\\n</td>\n",
              "      <td>[cooperative robots, graph theory, safe dynamics, robot dynamics, networks (as a synonym for graphs)]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>3003</td>\n",
              "      <td>Prospects for in-depth story understanding by computer</td>\n",
              "      <td>While much research on the hard problem of in-depth story understanding by\\ncomputer was performed starting in the 1970s, interest shifted in the 1990s to\\ninformation extraction and word sense disambiguation. Now that a degree of\\nsuccess has been achieved on these easier problems, I propose it is time to\\nreturn to in-depth story understanding. In this paper I examine the shift away\\nfrom story understanding, discuss some of the major problems in building a\\nstory understanding system, present some possible solutions involving a set of\\ninteracting understanding agents, and provide pointers to useful tools and\\nresources for building story understanding systems.\\n</td>\n",
              "      <td>[story, understanding, depth, computer, prospects]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>3009</td>\n",
              "      <td>Conditional indifference and conditional preservation</td>\n",
              "      <td>The idea of preserving conditional beliefs emerged recently as a new paradigm\\napt to guide the revision of epistemic states. Conditionals are substantially\\ndifferent from propositional beliefs and need specific treatment. In this\\npaper, we present a new approach to conditionals, capturing particularly well\\ntheir dynamic part as revision policies. We thoroughly axiomatize a principle\\nof conditional preservation as an indifference property with respect to\\nconditional structures of worlds. This principle is developed in a\\nsemi-quantitative setting, so as to reveal its fundamental meaning for belief\\nrevision in quantitative as well as in qualitative frameworks. In fact, it is\\nshown to cover other proposed approaches to conditional preservation.\\n</td>\n",
              "      <td>[conditional indifference, conditional preservation, indifference, preservation, conditionality]</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-02c1ec61-1090-4c28-b774-74e97fcba868')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-02c1ec61-1090-4c28-b774-74e97fcba868 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-02c1ec61-1090-4c28-b774-74e97fcba868');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-b1048ab4-d846-4220-821e-ade152be1573\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-b1048ab4-d846-4220-821e-ade152be1573')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-b1048ab4-d846-4220-821e-ade152be1573 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Parse Entries with Five Keywords"
      ],
      "metadata": {
        "id": "dxmRlXIBYrO3"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Analyze the LLM output to notice that, although in more than half of the cases 5 keywords/keyphrases are extracted, the remaining entries have as little as one keyword or as many as 11 keywords. For the next steps we retain only those titles for which exactly 5 keywords or keyphrases are generated."
      ],
      "metadata": {
        "id": "coOaz9rsa2-N"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import ast\n",
        "\n",
        "# Read data from the file - if necessary\n",
        "#df = pd.read_csv(data_path+extracted_keys_file)\n",
        "\n",
        "# The lists are saved as strings in csv, we need to parse them\n",
        "#df['titles_keys'] = df['titles_keys'].apply(ast.literal_eval)\n",
        "\n",
        "# Create a column that records the number of keywords/keyphrases\n",
        "df['keys_length'] = df['titles_keys'].apply(lambda x: len(x))"
      ],
      "metadata": {
        "id": "InQ2T-IBYy7w"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Import visualization packages and libraries\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "# Choose style and color palette\n",
        "import seaborn as sns\n",
        "sns.set_style(\"darkgrid\")\n",
        "\n",
        "colors = sns.color_palette('PuBuGn')\n",
        "\n",
        "# Visualize the keywords length distribution\n",
        "\n",
        "# Adjust figure size and font size\n",
        "sns.set(rc = {'figure.figsize':(6,3)})\n",
        "sns.set(font_scale=.75)\n",
        "\n",
        "# Pad margins so that markers don't get clipped by the axes\n",
        "plt.margins(x=0.2, y=0.2, tight=True);\n",
        "\n",
        "# Plot the keywords length distribution\n",
        "ax = sns.countplot(y='keys_length', data=df, color=colors[4])\n",
        "ax.bar_label(ax.containers[0]);\n",
        "\n",
        "# Add labels and title\n",
        "plt.title('Number of keywords/keyphrases per title')\n",
        "plt.xlabel('Keywords')\n",
        "plt.ylabel('Count')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 328
        },
        "id": "7rjsGpskhW5s",
        "outputId": "7406cb9c-5e30-4f6d-e4df-8b276c6e6ad7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 600x300 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Keep those titles that have 5 keywords/keyphrases only\n",
        "df5 = df[df.keys_length == 5]\n",
        "\n",
        "# Drop the 'keys_length' columns\n",
        "df5.drop('keys_length', axis=1, inplace=True)\n",
        "\n",
        "# Save the data to a csv file\n",
        "df5.to_csv(data_path+parsed_keys_file, index=False)\n",
        "\n",
        "# Check for success\n",
        "df5.head(10)"
      ],
      "metadata": {
        "id": "Ph4H6NdSpS6m",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 988
        },
        "outputId": "013cb3c8-ad11-4fcb-e40c-dc08c6a4194c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-14-08aeaebc71b4>:5: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  df5.drop('keys_length', axis=1, inplace=True)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      id                                                              title  \\\n",
              "0   1001  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "1   1002                    Minimum Description Length and Compositionality   \n",
              "2   1005                Effect of different packet sizes on RED performance   \n",
              "3   1007                           RED behavior with different packet sizes   \n",
              "4   1026                        A Logic for SDSI's Linked Local Name Spaces   \n",
              "6   2007                           Requirements of Text Processing Lexicons   \n",
              "7   2014                          Safe cooperative robot dynamics on graphs   \n",
              "8   3003             Prospects for in-depth story understanding by computer   \n",
              "9   3009              Conditional indifference and conditional preservation   \n",
              "10  3013                         A flexible framework for defeasible logics   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     abstract  \\\n",
              "0                                                                                                                                                                                                                     The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n   \n",
              "1                                                          We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n   \n",
              "2                                                                                                                                                                                           We consider the adaptation of random early detection (RED) as an active queue\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two existing RED\\nvariants and point out a weakness in both. The first variant where the drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high Packet Loss Ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n   \n",
              "3                                                                                                                                                                                              We consider the adaptation of random early detection (RED) as a buffer\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two RED variants\\ndescribed in [4] and point out a weakness in both. The first variant where drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high packet loss ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n   \n",
              "4                                                         Abadi has introduced a logic to explicate the meaning of local names in SDSI,\\nthe Simple Distributed Security Infrastructure proposed by Rivest and Lampson.\\nAbadi's logic does not correspond precisely to SDSI, however; it draws\\nconclusions about local names that do not follow from SDSI's name resolution\\nalgorithm. Moreover, its semantics is somewhat unintuitive. This paper presents\\nthe Logic of Local Name Containment, which does not suffer from these\\ndeficiencies. It has a clear semantics and provides a tight characterization of\\nSDSI name resolution. The semantics is shown to be closely related to that of\\nlogic programs, leading to an approach to the efficient implementation of\\nqueries concerning local names. A complete axiomatization of the logic is also\\nprovided.\\n   \n",
              "6     As text processing systems expand in scope, they will require ever larger\\nlexicons along with a parsing capability for discriminating among many senses\\nof a word. Existing systems do not incorporate such subtleties in meaning for\\ntheir lexicons. Ordinary dictionaries contain such information, but are largely\\nuntapped. When the contents of dictionaries are scrutinized, they reveal many\\nrequirements that must be satisfied in representing meaning and in developing\\nsemantic parsers. These requirements were identified in research designed to\\nfind primitive verb concepts. The requirements are outlined and general\\nprocedures for satisfying them through the use of ordinary dictionaries are\\ndescribed, illustrated by building frames for and examining the definitions of\\n\"change\" and its uses as a hypernym in other definitions.\\n   \n",
              "7                                                                                                                                                                This paper initiates the use of vector fields to design, optimize, and\\nimplement reactive schedules for safe cooperative robot patterns on planar\\ngraphs. We consider Automated Guided Vehicles (AGV's) operating upon a\\npredefined network of pathways. In contrast to the case of locally Euclidean\\nconfiguration spaces, regularization of collisions is no longer a local\\nprocedure, and issues concerning the global topology of configuration spaces\\nmust be addressed. The focus of the present inquiry is the achievement of safe,\\nefficient, cooperative patterns in the simplest nontrivial example (a pair of\\nrobots on a Y-network) by means of a state-event heirarchical controller.\\n   \n",
              "8                                                                                                                                                                          While much research on the hard problem of in-depth story understanding by\\ncomputer was performed starting in the 1970s, interest shifted in the 1990s to\\ninformation extraction and word sense disambiguation. Now that a degree of\\nsuccess has been achieved on these easier problems, I propose it is time to\\nreturn to in-depth story understanding. In this paper I examine the shift away\\nfrom story understanding, discuss some of the major problems in building a\\nstory understanding system, present some possible solutions involving a set of\\ninteracting understanding agents, and provide pointers to useful tools and\\nresources for building story understanding systems.\\n   \n",
              "9                                                                                   The idea of preserving conditional beliefs emerged recently as a new paradigm\\napt to guide the revision of epistemic states. Conditionals are substantially\\ndifferent from propositional beliefs and need specific treatment. In this\\npaper, we present a new approach to conditionals, capturing particularly well\\ntheir dynamic part as revision policies. We thoroughly axiomatize a principle\\nof conditional preservation as an indifference property with respect to\\nconditional structures of worlds. This principle is developed in a\\nsemi-quantitative setting, so as to reveal its fundamental meaning for belief\\nrevision in quantitative as well as in qualitative frameworks. In fact, it is\\nshown to cover other proposed approaches to conditional preservation.\\n   \n",
              "10                                                                                                                                      Logics for knowledge representation suffer from over-specialization: while\\neach logic may provide an ideal representation formalism for some problems, it\\nis less than optimal for others. A solution to this problem is to choose from\\nseveral logics and, when necessary, combine the representations. In general,\\nsuch an approach results in a very difficult problem of combination. However,\\nif we can choose the logics from a uniform framework then the problem of\\ncombining them is greatly simplified. In this paper, we develop such a\\nframework for defeasible logics. It supports all defeasible logics that satisfy\\na strong negation principle. We use logic meta-programs as the basis for the\\nframework.\\n   \n",
              "\n",
              "                                                                                                                                              titles_keys  \n",
              "0                                                                        [von neumann, quantum logic, classical von neumann architecture, quantum, logic]  \n",
              "1                      [minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]  \n",
              "2                                                                            [packet sizes, RED performance, different packet sizes, effect, performance]  \n",
              "3                                                                            [packet sizes, RED performance, different packet sizes, effect, performance]  \n",
              "4                                                                                              [SDSI, linked local names, logic, LLN spaces, sdsi-graphs]  \n",
              "6   [text processing lexicons, lexicon requirements, natural language processing lexicons, lexical resources for text processing, text processing lexica]  \n",
              "7                                                   [cooperative robots, graph theory, safe dynamics, robot dynamics, networks (as a synonym for graphs)]  \n",
              "8                                                                                                      [story, understanding, depth, computer, prospects]  \n",
              "9                                                        [conditional indifference, conditional preservation, indifference, preservation, conditionality]  \n",
              "10                                                                                  [defeasible logics, flexibility, framework, reasoning, non-monotonic]  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-753870b4-700e-4533-9911-7f78c72aff68\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>title</th>\n",
              "      <th>abstract</th>\n",
              "      <th>titles_keys</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1001</td>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n</td>\n",
              "      <td>[von neumann, quantum logic, classical von neumann architecture, quantum, logic]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1002</td>\n",
              "      <td>Minimum Description Length and Compositionality</td>\n",
              "      <td>We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n</td>\n",
              "      <td>[minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1005</td>\n",
              "      <td>Effect of different packet sizes on RED performance</td>\n",
              "      <td>We consider the adaptation of random early detection (RED) as an active queue\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two existing RED\\nvariants and point out a weakness in both. The first variant where the drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high Packet Loss Ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n</td>\n",
              "      <td>[packet sizes, RED performance, different packet sizes, effect, performance]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>1007</td>\n",
              "      <td>RED behavior with different packet sizes</td>\n",
              "      <td>We consider the adaptation of random early detection (RED) as a buffer\\nmanagement algorithm for TCP traffic in Internet gateways where different\\nmaximum transfer units (MTUs) are used. We studied the two RED variants\\ndescribed in [4] and point out a weakness in both. The first variant where drop\\nprobability is independent from the packet size discriminates connections with\\nsmaller MTUs. The second variant results in a very high packet loss ratio\\n(PLR), and as a consequence low goodput, for connections with higher MTUs. We\\nshow that fairness in terms of loss and goodput can be supplied through an\\nappropriate setting of the RED algorithm.\\n</td>\n",
              "      <td>[packet sizes, RED performance, different packet sizes, effect, performance]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>1026</td>\n",
              "      <td>A Logic for SDSI's Linked Local Name Spaces</td>\n",
              "      <td>Abadi has introduced a logic to explicate the meaning of local names in SDSI,\\nthe Simple Distributed Security Infrastructure proposed by Rivest and Lampson.\\nAbadi's logic does not correspond precisely to SDSI, however; it draws\\nconclusions about local names that do not follow from SDSI's name resolution\\nalgorithm. Moreover, its semantics is somewhat unintuitive. This paper presents\\nthe Logic of Local Name Containment, which does not suffer from these\\ndeficiencies. It has a clear semantics and provides a tight characterization of\\nSDSI name resolution. The semantics is shown to be closely related to that of\\nlogic programs, leading to an approach to the efficient implementation of\\nqueries concerning local names. A complete axiomatization of the logic is also\\nprovided.\\n</td>\n",
              "      <td>[SDSI, linked local names, logic, LLN spaces, sdsi-graphs]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>2007</td>\n",
              "      <td>Requirements of Text Processing Lexicons</td>\n",
              "      <td>As text processing systems expand in scope, they will require ever larger\\nlexicons along with a parsing capability for discriminating among many senses\\nof a word. Existing systems do not incorporate such subtleties in meaning for\\ntheir lexicons. Ordinary dictionaries contain such information, but are largely\\nuntapped. When the contents of dictionaries are scrutinized, they reveal many\\nrequirements that must be satisfied in representing meaning and in developing\\nsemantic parsers. These requirements were identified in research designed to\\nfind primitive verb concepts. The requirements are outlined and general\\nprocedures for satisfying them through the use of ordinary dictionaries are\\ndescribed, illustrated by building frames for and examining the definitions of\\n\"change\" and its uses as a hypernym in other definitions.\\n</td>\n",
              "      <td>[text processing lexicons, lexicon requirements, natural language processing lexicons, lexical resources for text processing, text processing lexica]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>2014</td>\n",
              "      <td>Safe cooperative robot dynamics on graphs</td>\n",
              "      <td>This paper initiates the use of vector fields to design, optimize, and\\nimplement reactive schedules for safe cooperative robot patterns on planar\\ngraphs. We consider Automated Guided Vehicles (AGV's) operating upon a\\npredefined network of pathways. In contrast to the case of locally Euclidean\\nconfiguration spaces, regularization of collisions is no longer a local\\nprocedure, and issues concerning the global topology of configuration spaces\\nmust be addressed. The focus of the present inquiry is the achievement of safe,\\nefficient, cooperative patterns in the simplest nontrivial example (a pair of\\nrobots on a Y-network) by means of a state-event heirarchical controller.\\n</td>\n",
              "      <td>[cooperative robots, graph theory, safe dynamics, robot dynamics, networks (as a synonym for graphs)]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>3003</td>\n",
              "      <td>Prospects for in-depth story understanding by computer</td>\n",
              "      <td>While much research on the hard problem of in-depth story understanding by\\ncomputer was performed starting in the 1970s, interest shifted in the 1990s to\\ninformation extraction and word sense disambiguation. Now that a degree of\\nsuccess has been achieved on these easier problems, I propose it is time to\\nreturn to in-depth story understanding. In this paper I examine the shift away\\nfrom story understanding, discuss some of the major problems in building a\\nstory understanding system, present some possible solutions involving a set of\\ninteracting understanding agents, and provide pointers to useful tools and\\nresources for building story understanding systems.\\n</td>\n",
              "      <td>[story, understanding, depth, computer, prospects]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>3009</td>\n",
              "      <td>Conditional indifference and conditional preservation</td>\n",
              "      <td>The idea of preserving conditional beliefs emerged recently as a new paradigm\\napt to guide the revision of epistemic states. Conditionals are substantially\\ndifferent from propositional beliefs and need specific treatment. In this\\npaper, we present a new approach to conditionals, capturing particularly well\\ntheir dynamic part as revision policies. We thoroughly axiomatize a principle\\nof conditional preservation as an indifference property with respect to\\nconditional structures of worlds. This principle is developed in a\\nsemi-quantitative setting, so as to reveal its fundamental meaning for belief\\nrevision in quantitative as well as in qualitative frameworks. In fact, it is\\nshown to cover other proposed approaches to conditional preservation.\\n</td>\n",
              "      <td>[conditional indifference, conditional preservation, indifference, preservation, conditionality]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>3013</td>\n",
              "      <td>A flexible framework for defeasible logics</td>\n",
              "      <td>Logics for knowledge representation suffer from over-specialization: while\\neach logic may provide an ideal representation formalism for some problems, it\\nis less than optimal for others. A solution to this problem is to choose from\\nseveral logics and, when necessary, combine the representations. In general,\\nsuch an approach results in a very difficult problem of combination. However,\\nif we can choose the logics from a uniform framework then the problem of\\ncombining them is greatly simplified. In this paper, we develop such a\\nframework for defeasible logics. It supports all defeasible logics that satisfy\\na strong negation principle. We use logic meta-programs as the basis for the\\nframework.\\n</td>\n",
              "      <td>[defeasible logics, flexibility, framework, reasoning, non-monotonic]</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-753870b4-700e-4533-9911-7f78c72aff68')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-753870b4-700e-4533-9911-7f78c72aff68 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-753870b4-700e-4533-9911-7f78c72aff68');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-1774da2f-f1b6-4409-bae7-b0b7aab56009\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-1774da2f-f1b6-4409-bae7-b0b7aab56009')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-1774da2f-f1b6-4409-bae7-b0b7aab56009 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Use HDBSCAN & UMAP to Cluster the Keywords"
      ],
      "metadata": {
        "id": "W6-RTuFk7GN0"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Required Installs and Imports HDBSCAN & UMAP"
      ],
      "metadata": {
        "id": "GXaOuQK8dP2r"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Run this before the next installs to avoid errors\n",
        "import locale\n",
        "locale.getpreferredencoding = lambda: \"UTF-8\""
      ],
      "metadata": {
        "id": "pMkpX96GCqjA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install umap-learn\n",
        "!pip install hdbscan\n",
        "!pip install -U sentence-transformers"
      ],
      "metadata": {
        "id": "DI_OISiRCclz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# General libraries and packages\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import re\n",
        "import pickle\n",
        "\n",
        "# Install packages to generate the BERT embeddings\n",
        "from sentence_transformers import SentenceTransformer\n",
        "\n",
        "# Packages for dimensionality reduction\n",
        "import umap.umap_ as umap\n",
        "\n",
        "# Import clustering algorithms\n",
        "import hdbscan\n"
      ],
      "metadata": {
        "id": "BHTektx1Bq0K"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Prepare the Keywords for Embedding & Clustering"
      ],
      "metadata": {
        "id": "IYzJiTd7coV2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the data if needed\n",
        "#df5 = pd.read_csv(data_path+parsed_keys_file)\n",
        "\n",
        "# Create a list of all sublists of keywords and keyphrases\n",
        "df5_keys = df5.titles_keys.tolist()\n",
        "\n",
        "# Flatten the list of sublists\n",
        "flat_keys = [item for sublist in df5_keys for item in sublist]\n",
        "\n",
        "# Create a list of unique keywords\n",
        "flat_keys = list(set(flat_keys))\n",
        "\n",
        "# Create a dataframe with the distinct keywords\n",
        "keys_df = pd.DataFrame(flat_keys, columns = ['key'])\n",
        "\n",
        "# Count how many distinct keywords were extracted\n",
        "keys_df.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1XX41U6X8WzY",
        "outputId": "248b89a7-8abc-498a-e77e-bbe7690c8611"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2989, 1)"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Display a sample of keywords or keyphrases\n",
        "keys_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "om_peUgaR3-H",
        "outputId": "ee07b062-ecf4-479f-c4e0-4e3a9f980d88"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                  key\n",
              "0  n-colorable graphs\n",
              "1         experiments\n",
              "2         constraints\n",
              "3      tree structure\n",
              "4          complexity"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a24b4af8-a485-4a16-a78f-f69399b74d45\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>n-colorable graphs</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>experiments</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>constraints</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>tree structure</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>complexity</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a24b4af8-a485-4a16-a78f-f69399b74d45')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-a24b4af8-a485-4a16-a78f-f69399b74d45 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-a24b4af8-a485-4a16-a78f-f69399b74d45');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a7a39f5a-bc78-4a9e-b087-561433d83112\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a7a39f5a-bc78-4a9e-b087-561433d83112')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a7a39f5a-bc78-4a9e-b087-561433d83112 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Build Embeddings for Clustering"
      ],
      "metadata": {
        "id": "cvHIFLVpddm5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Instantiate the embedding model\n",
        "model = SentenceTransformer('all-mpnet-base-v2')\n",
        "\n",
        "# Embed the keywords and keyphrases into 768-dim real vector space\n",
        "keys_df['key_bert'] = keys_df['key'].apply(lambda x: model.encode(x))"
      ],
      "metadata": {
        "id": "xqSSHBm9SphS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Use UMAP algorithm\n",
        "\n",
        "# Reduce the dimensionality of the vectors to 10 while keeping the size of the local neighborhood to 15\n",
        "embeddings = umap.UMAP(n_neighbors=15,\n",
        "                       n_components=10,\n",
        "                       metric='cosine').fit_transform(list(keys_df.key_bert))\n",
        "\n",
        "# Add the reduced embeddings to the dataframe\n",
        "keys_df['key_umap'] = embeddings.tolist()\n",
        "\n",
        "# Check the output\n",
        "#keys_df.head(2)"
      ],
      "metadata": {
        "id": "PC3ymP4mA7pn"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Cluster the Keywords with HDBSCAN"
      ],
      "metadata": {
        "id": "yt8uJ8gceJ7e"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Work with HDBSCAN clustering algorithm\n",
        "\n",
        "# Initialize the clustering model\n",
        "\n",
        "clusterer = hdbscan.HDBSCAN(algorithm='best',\n",
        "                            prediction_data=True,\n",
        "                            approx_min_span_tree=True,\n",
        "                            gen_min_span_tree=True,\n",
        "                            min_cluster_size=20,\n",
        "                            cluster_selection_epsilon = .1,\n",
        "                            min_samples=1,\n",
        "                            p=None,\n",
        "                            metric='euclidean',\n",
        "                            cluster_selection_method='leaf')\n",
        "\n",
        "# Fit the data\n",
        "clusterer.fit(embeddings)\n",
        "\n",
        "# Add hard clusters to the data\n",
        "keys_df['hard_labels'] = clusterer.labels_.tolist()\n",
        "\n",
        "# Create soft clusters\n",
        "soft_clusters = hdbscan.all_points_membership_vectors(clusterer)\n",
        "\n",
        "# Add the soft cluster information to the data\n",
        "closest_clusters = [np.argmax(x) for x in soft_clusters]\n",
        "keys_df['cluster'] = closest_clusters\n",
        "\n",
        "# Save embeddings and clusters to a pickle file\n",
        "keys_df.to_pickle(data_path+embeddings_clusters_keys_file)"
      ],
      "metadata": {
        "id": "gMP8VP6yBdxm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Take a look at one cluster (soft clustering)\n",
        "keys_df[keys_df.cluster==1].key.tolist()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "n6oS4edJjJbG",
        "outputId": "9ad55c9a-25e1-482c-9e4d-222edad04f0d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['quantum computing',\n",
              " 'Q^2 adversary structures',\n",
              " 'scidac lattice qcd',\n",
              " 'state complexes',\n",
              " 'quantum logic',\n",
              " 'quantum computation',\n",
              " 'quantum secrets',\n",
              " 'classical states',\n",
              " 'spin systems',\n",
              " 'entanglement',\n",
              " 'quantum adversaries',\n",
              " 'quantum states',\n",
              " 'noisy quantum resources',\n",
              " 'quantum provers',\n",
              " 'quantum computers',\n",
              " 'classical von neumann architecture',\n",
              " 'ising spin simulations',\n",
              " 'quantum',\n",
              " 'particle',\n",
              " 'von neumann',\n",
              " 'quantum sharing of classical secrets',\n",
              " 'Ghz states',\n",
              " 'scidac lattice qcd computations']"
            ]
          },
          "metadata": {},
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Clusters Visualizations"
      ],
      "metadata": {
        "id": "3OZRC4-H8xjG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Import visualization packages and libraries\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "# Choose style and color palette\n",
        "import seaborn as sns\n",
        "sns.set_style(\"darkgrid\")\n",
        "\n",
        "colors = sns.color_palette('PuBuGn')"
      ],
      "metadata": {
        "id": "uoVrK44DGrQI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Visualize the keys distribution by hard topics\n",
        "\n",
        "# adjust figure size and font size\n",
        "sns.set(rc = {'figure.figsize':(8,6)})\n",
        "sns.set(font_scale=.60)\n",
        "\n",
        "# pad margins so that markers don't get clipped by the axes\n",
        "plt.margins(x=0.2, y=0.2, tight=True);\n",
        "\n",
        "# plot the page distribution\n",
        "ax = sns.countplot(y='hard_labels', data=keys_df, color=colors[4])\n",
        "ax.bar_label(ax.containers[0]);\n",
        "\n",
        "# Add labels and title\n",
        "plt.title('Distributions of topics obtained with hard HDBSAN clustering')\n",
        "plt.xlabel('Topic Label')\n",
        "plt.ylabel('Count')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 553
        },
        "id": "gGapqEYROn94",
        "outputId": "1832515f-5726-45d8-afef-e11a7375758d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Visualize the keys distribution by soft topics\n",
        "\n",
        "# adjust figure size and font size\n",
        "sns.set(rc = {'figure.figsize':(8,8)})\n",
        "sns.set(font_scale=.60)\n",
        "\n",
        "# pad margins so that markers don't get clipped by the axes\n",
        "plt.margins(x=0.2, y=0.2, tight=True);\n",
        "\n",
        "# plot the page distribution\n",
        "ax = sns.countplot(y='cluster', data=keys_df, color=colors[4])\n",
        "ax.bar_label(ax.containers[0]);\n",
        "\n",
        "# Add labels and title\n",
        "plt.title('Distributions of topics obtained with soft HDBSAN clustering')\n",
        "plt.xlabel('Topic Label')\n",
        "plt.ylabel('Count')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 707
        },
        "id": "jv6Sj_sdIIHd",
        "outputId": "ddba1aa1-feeb-4ac1-a580-aff358b60ed6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x800 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Extract Cluster Descriptions and Labels"
      ],
      "metadata": {
        "id": "wARsJOYdtdT3"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Load the Clustered Keywords Data"
      ],
      "metadata": {
        "id": "ZkhEY1ntfCBr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Read the data if necessary\n",
        "keys_df = pd.read_pickle(data_path+embeddings_clusters_keys_file)\n",
        "\n",
        "# Drop the extra columns\n",
        "keys_df.drop(columns=['key_bert', 'key_umap', 'hard_labels'], inplace=True)\n",
        "\n",
        "# Check the output\n",
        "keys_df.head()"
      ],
      "metadata": {
        "id": "H-qXcCECtcl8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "28e008e4-9858-4821-ec16-0a2554958fd8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                  key  cluster\n",
              "0  n-colorable graphs       16\n",
              "1         experiments        5\n",
              "2         constraints        7\n",
              "3      tree structure       18\n",
              "4          complexity       25"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-dd28526e-3933-4968-89be-f031dd3bf317\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>cluster</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>n-colorable graphs</td>\n",
              "      <td>16</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>experiments</td>\n",
              "      <td>5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>constraints</td>\n",
              "      <td>7</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>tree structure</td>\n",
              "      <td>18</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>complexity</td>\n",
              "      <td>25</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-dd28526e-3933-4968-89be-f031dd3bf317')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-dd28526e-3933-4968-89be-f031dd3bf317 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-dd28526e-3933-4968-89be-f031dd3bf317');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-2746d4e5-5180-45f8-884c-e77e69ae1b2e\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-2746d4e5-5180-45f8-884c-e77e69ae1b2e')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-2746d4e5-5180-45f8-884c-e77e69ae1b2e button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Extract Descriptions and Labels for Each Cluster"
      ],
      "metadata": {
        "id": "kV0Fp4JifOhw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import re\n",
        "\n",
        "def extract_description(df: pd.DataFrame,\n",
        "                        n: int\n",
        "                        )-> pd.DataFrame:\n",
        "    \"\"\"\n",
        "    Use a custom prompt to send to a LLM\n",
        "    to extract labels and descriptions for a list of keywords.\n",
        "    \"\"\"\n",
        "\n",
        "    one_cluster = df[df['cluster']==n]\n",
        "    one_cluster_copy = one_cluster.copy()\n",
        "    sample = one_cluster_copy.key.tolist()\n",
        "\n",
        "    prompt_clusters= f\"\"\"\n",
        "    <|system|>\n",
        "    I have the following list of keywords and keyphrases:\n",
        "    ['encryption','attribute','firewall','security properties',\n",
        "    'network security','reliability','surveillance','distributed risk factors',\n",
        "    'still vulnerable','cryptographic','protocol','signaling','safe',\n",
        "    'adversary','message passing','input-determined guards','secure communication',\n",
        "    'vulnerabilities','value-at-risk','anti-spam','intellectual property rights',\n",
        "    'countermeasures','security implications','privacy','protection',\n",
        "    'mitigation strategies','vulnerability','secure networks','guards']\n",
        "\n",
        "    Based on the information above, first name the domain these keywords or keyphrases belong to, secondly\n",
        "    give a brief description of the domain.\n",
        "    Do not use more than 30 words for the description!\n",
        "    Do not provide details!\n",
        "    Do not give examples of the contexts, do not say 'such as' and do not list the keywords or the keyphrases!\n",
        "    Do not start with a statement of the form 'These keywords belong to the domain of' or with 'The domain'.\n",
        "\n",
        "    Cybersecurity: Cybersecurity, emphasizing methods and strategies for safeguarding digital information\n",
        "    and networks against unauthorized access and threats.\n",
        "    </s>\n",
        "\n",
        "    <|user|>\n",
        "    I have the following list of keywords and keyphrases:\n",
        "    {sample}\n",
        "    Based on the information above, first name the domain these keywords or keyphrases belong to, secondly\n",
        "    give a brief description of the domain.\n",
        "    Do not use more than 30 words for the description!\n",
        "    Do not provide details!\n",
        "    Do not give examples of the contexts, do not say 'such as' and do not list the keywords or the keyphrases!\n",
        "    Do not start with a statement of the form 'These keywords belong to the domain of' or with 'The domain'.\n",
        "    <|assistant|>\n",
        "    \"\"\"\n",
        "\n",
        "    # Generate the outputs\n",
        "    outputs = generator(prompt_clusters,\n",
        "                    max_new_tokens=120,\n",
        "                    do_sample=True,\n",
        "                    temperature=0.1,\n",
        "                    top_k=10,\n",
        "                    top_p=0.95)\n",
        "\n",
        "    text = outputs[0][\"generated_text\"]\n",
        "\n",
        "    # Example string\n",
        "    pattern = \"<|assistant|>\\n\"\n",
        "\n",
        "    # Extract the output\n",
        "    response = text.split(pattern, 1)[1].strip(\" \")\n",
        "\n",
        "    if len(response.split(\":\", 1)) == 2:\n",
        "        label  = response.split(\":\", 1)[0].strip(\" \")\n",
        "        description = response.split(\":\", 1)[1].strip(\" \")\n",
        "    else:\n",
        "        label = description = response\n",
        "\n",
        "    # Add the description and the labels to the dataframe\n",
        "    one_cluster_copy.loc[:, 'description'] = description\n",
        "    one_cluster_copy.loc[:, 'label'] = label\n",
        "\n",
        "    return one_cluster_copy"
      ],
      "metadata": {
        "id": "9CuI0W1apnz3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Initialize an empty list to store the cluster dataframes\n",
        "dataframes = []\n",
        "clusters = len(set(keys_df.cluster))\n",
        "\n",
        "# Iterate over the range of n values\n",
        "for n in range(clusters-1):\n",
        "    df_result = extract_description(keys_df,n)\n",
        "    dataframes.append(df_result)\n",
        "\n",
        "# Concatenate all the dataframes\n",
        "final_df = pd.concat(dataframes, ignore_index=True)"
      ],
      "metadata": {
        "id": "qmPyO5rP_UD3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Let's check if there are any faulty otputs\n",
        "faulty_df = final_df[final_df['description'] == final_df['label']]\n",
        "set(faulty_df.cluster.tolist())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bvIbClxviuj1",
        "outputId": "52862e62-2507-41c7-83e9-94ff745a045e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{7, 29}"
            ]
          },
          "metadata": {},
          "execution_count": 95
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Investigate the output - also need to check manually\n",
        "pd.set_option('display.max_colwidth', None)\n",
        "final_df.drop_duplicates(subset='cluster')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "OWIJGFxmkUc8",
        "outputId": "1cf942f6-e3cd-41bb-ae20-1539d8765a6d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                            key  cluster  \\\n",
              "0                      standard data processing        0   \n",
              "50                            quantum computing        1   \n",
              "73    expected qualitative utility maximization        2   \n",
              "129                          cooperative robots        3   \n",
              "166                                  data model        4   \n",
              "195                                 experiments        5   \n",
              "224                         bayesian estimation        6   \n",
              "261                                 constraints        7   \n",
              "291                   calculus-of-constructions        8   \n",
              "330                                neuroscience        9   \n",
              "370                        similarity searching       10   \n",
              "404                                  point sets       11   \n",
              "444                nearest neighbor classifiers       12   \n",
              "491                            IIR filter banks       13   \n",
              "561                    interleaved construction       14   \n",
              "600                           pseudo-randomness       15   \n",
              "645                          n-colorable graphs       16   \n",
              "688                                  2-D steady       17   \n",
              "771                              tree structure       18   \n",
              "819                                  indicative       19   \n",
              "867                                      length       20   \n",
              "942                            sperner triangle       21   \n",
              "1008                                 simplicity       22   \n",
              "1059                           convergence rate       23   \n",
              "1123                                 inequality       24   \n",
              "1156                                 complexity       25   \n",
              "1201                        analysis algorithms       26   \n",
              "1272                  fitness uniform selection       27   \n",
              "1330                                     tagger       28   \n",
              "1386                           content analysis       29   \n",
              "1431                                    pronoun       30   \n",
              "1473                                     memory       31   \n",
              "1512                            optical channel       32   \n",
              "1571                       constrained decoding       33   \n",
              "1611                               hierarchical       34   \n",
              "1664                    Dual Light Affine Logic       35   \n",
              "1722                                 asymmetric       36   \n",
              "1774                                    finding       37   \n",
              "1844                    asymptotic learnability       38   \n",
              "1872                                       ldpc       39   \n",
              "1903                                        lsa       40   \n",
              "1978                                contentware       41   \n",
              "2048         networks (as a synonym for graphs)       42   \n",
              "2106                      pattern-based routing       43   \n",
              "2151                         control constructs       44   \n",
              "2186                            leader election       45   \n",
              "2255                          design approaches       46   \n",
              "2327                                      cores       47   \n",
              "2398                     distributed processing       48   \n",
              "2440                            rewrite systems       49   \n",
              "2481                        iguana architecture       50   \n",
              "2562                                   exchange       51   \n",
              "2605                      MPEG-video encryption       52   \n",
              "2634                        intrusion detection       53   \n",
              "2686                        serial independence       54   \n",
              "2710                     research communication       55   \n",
              "2800                                    grammar       56   \n",
              "2839                 natural language front-end       57   \n",
              "2878                    semantic disambiguation       58   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           description  \\\n",
              "0                    Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and   \n",
              "50                                                                                                                                                                                                                                                                                                                                                                       A multidisciplinary field focused on developing theories, algorithms, and technologies for processing and transmitting information using quantum phenomena, encompassing topics such as quantum computing, cryptography, simulation, and resource estimation.   \n",
              "73                                                                                                                                                                                                                                                                                                                                                                                                                                  The study of how individuals make choices, focusing on methods for modeling and analyzing decision-making processes, ranking algorithms, combinatorial problems, and preference change management.   \n",
              "129                                                                                                                                                                                                                                                                                                                                                                 The field of robotics and automation, encompassing technologies that enable machines to work cooperatively alongside humans in various industries, such as manufacturing, logistics, and healthcare, through the development of automatic and intelligent systems.   \n",
              "166                                                                                                       The domain of modeling encompasses various approaches and techniques for creating representations, or models, that simulate or infer real-world phenomena, facilitate decision-making, and enable analysis through computational means. This includes methods for selecting, refining, and validating models, as well as techniques for exporting behavior and checking model correctness. Key areas of focus include data modeling, inference systems, modeling techniques, multi-model systems, and model-based reasoning.   \n",
              "195                                                                                                                                                                                                                                                                                                     A domain focused on utilizing statistical techniques to analyze experimental data, test hypotheses, and draw conclusions based on empirical evidence. Emphasis is placed on statistical inference, hypothesis testing, sample size determination, and statistical modeling to support scientific research and decision making.   \n",
              "224                                                                                                                                                                                                                                                                                                                                                                                                                                                                               A domain focused on developing statistical models that incorporate uncertainty and probability to reason about complex systems and make predictions.   \n",
              "261                                                                                                                                                                                                                                                                                                                                                                                                                                    \\n<|user|>\\nCan you please suggest some potential applications or fields where these keywords are commonly used? Again, please keep it brief and avoid providing specific examples or contexts.   \n",
              "291                                                                                                                                                                                                                                                                                                                                                                                                                                                                  A programming paradigm focused on mathematical functions and declarative programming, utilizing concepts such as type classes, automata theory, and polymorphism.   \n",
              "330                                                                                                                                                                                                                                                                                                                                                                                                                                                        The interdisciplinary study of how humans acquire, process, and apply knowledge and skills, with a focus on understanding the neural mechanisms underlying these processes.   \n",
              "370                                                                                                                                                                                                                                                                                                                                                                                                   Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.   \n",
              "404           Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,   \n",
              "444                                                                                                                                                                                                                                                                                                                                                                                                                                          The field of computer science focused on developing algorithms and models that enable computers to learn from data and make predictions or decisions without being explicitly programmed.   \n",
              "491                            A domain focused on developing techniques for analyzing, filtering, and extracting useful information from visual and audio signals, utilizing methods such as IIR filter banks, cosine modulated loss networks, arithmetic filters, spectral analysis, filter stability, and denoising. Applications include image segmentation, fingerprint vectors, solar images analysis, face detection and recognition, and vision-based road detection. Additionally, this domain encompasses topics such as artificial intelligence, neural networks, Fourier transforms, PC (principal component) analysis, sp   \n",
              "561                                                                                                                                                                                                                                                                                                          These keywords belong to the domain of cybersecurity. Cybersecurity, emphasizing methods and strategies for safeguarding digital information and networks against unauthorized access and threats.\\nCybersecurity: A Survey (https://ieeexplore.ieee.org/abstract/document/7641895/) - This paper provides an overview of   \n",
              "600                                                                                                                                                                                                                                                                                                                                                                                                                                                   Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.   \n",
              "645                                                                                                                                                                                                                      The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.   \n",
              "688                                                                                                                                                                                                                                                                                                                                                                                            A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.   \n",
              "771                                                                                                                                                                                                                                                                                                                                                                                                                    The domain concerned with applying computational techniques to analyze biological data, including methods for constructing tree structures from molecular data, clustering algorithms, and evolutionary models.   \n",
              "819                                                                                                                  The study of complex, nonlinear systems exhibiting sensitive dependence on initial conditions, leading to seemingly random behavior and long-term memory effects. Emphasis on dynamical systems, statistical mechanics, phase transitions, propagation, damping, power laws, and novelty in physics, biology, economics, and other fields. Topics include chaos, nondeterminism, nonlinearity, bipolarity, synchronization, safe dynamics, non-intrusiveness, vibrations, trembling hands, and symbolic dynamics.   \n",
              "867                                                                                            A diverse range of topics related to geometry and physics, including high dimensions, magnification laws, scrollbars, tuplespaces, color spaces, ray attacks, Fresnel, gradients, amplify-and-forward, astronomy, hardness, thickness, large objects, view sizes, resistance, telescopes, space complexity, arbitrary dimensionality, multidimensional spaces, scaling, vectors, incompressible flows, length functions, convexity, matrices, higher-dimensional manifolds, PSPACE reasoning, continuous manifolds, vector sources, and   \n",
              "942                                                                                                                                                                                                                                                                                                                       Encompasses various aspects related to shapes, spaces, and structures, including but not limited to arbitrary domains, conformal transformations, surface topology, scale-free topology, circularity, spherical and cylindrical geometries, tensor analysis, octagonal packings, and orbital configurations.   \n",
              "1008                           Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also   \n",
              "1059                                                                                                                                                                                                                                                                                                                                                                      The domain concerned with the development and application of mathematical techniques and tools for solving numerical problems in various fields, such as engineering, physics, finance, and science, through the use of computers and other digital devices.   \n",
              "1123                                                                                                                                                                                                                                                                                                                                                                                                                             Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.   \n",
              "1156                                                                                                                                                                                                                                                                                                                                                                                                                                 The study of the resources required by algorithms to solve problems, including time, space, and input size, with a focus on understanding the inherent difficulty of certain computational tasks.   \n",
              "1201                                                                 Encompassing techniques and approaches for analyzing algorithms, designing circuits, and optimizing programs, with a focus on efficiency, performance, and time constraints. Includes topics such as high-level synthesis, circuit design, lazy bureaucrat scheduling, linear datalog, and VLSI circuit design. Also covers algorithmic analysis, fast methods, and utility computing. Keywords include analysis algorithms, generating algorithms, local algorithm, linear feedback shift register, log analysis, partial evaluation, and Runge-Kutta algorithm.   \n",
              "1272                                                                                                                                                                                                                                                                                                                                                                                                                                                       The field of finding the best solution from a set of possible alternatives, often through iterative processes that seek to maximize or minimize a given objective function.   \n",
              "1330                                                                                                                                                                                                                                                                                                                                                                                                                                                                       Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.   \n",
              "1386                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                \\n<|user|>\\nCan you suggest some mitigation strategies for vulnerabilities in secure networks based on the given list of keywords?   \n",
              "1431                                                                                                                                                                                                                                                                                                                                                                                                                                                      Focusing on various aspects related to teaching, learning, and instruction, including pedagogy, curriculum design, assessment, technology integration, and student outcomes.   \n",
              "1473                                                                                                                                                                                                    Focused on developing efficient and scalable methods for managing data storage, including techniques for memory allocation, failure detection, capacity analysis, and persistent buffer trees. Emphasis is placed on high performance, i/o efficiency, and resource allocation, with considerations for different packet sizes, shared memory, and anomaly detection. Memory consistency is also a key concern in this domain.   \n",
              "1512                    The study of techniques and technologies for transmitting information over various types of channels, including optical, wireless, and broadcast channels, with a focus on achieving optimal performance in terms of capacity, reliability, and security, while managing interference, distortion, and other sources of noise. Topics include source-channel coding, multi-antenna systems, spread spectrum technology, modulation, and signal propagation in dispersive and fading environments. Key concepts include mutual information, rate-distortion region, negative information, and channel capacity.   \n",
              "1571                                                                                                                                                                                                                                                                                                                                                                                                                                                          The study of mathematical techniques for encoding, transmitting, and decoding information in various forms, including error correction, compression, and network coding.   \n",
              "1611                                                                                                                                                         Focusing on the use of mathematical logic to prove theorems, establish invariants, and analyze structural properties through techniques such as transformation rule lists, poset theory, and preferential structures. Emphasis on hierarchical structures, duality theories, and interactive proof systems. Includes concepts such as proof nets, obdd-based proofs, and multi-prover interactive proofs. Examples include the Poincaré conjecture and Grillet's theorem.   \n",
              "1664                                                                                                A broad field encompassing various logical systems, including propositional, modal, temporal, and many-valued logics, as well as preferential, defeasible, and fuzzy logics. It also includes logic programming languages like Prolog, disjunctive databases, and temporal reasoning techniques. The focus is on representing meaning over time, natural deduction, symbolic quantifier elimination, and logical foundations for soft computing paradigms. Ambient logic, combinatory logic, and neuro-fuzzy compensation are also   \n",
              "1722                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               The study of knowledge, including its nature, sources, limits, and validity, from multiple perspectives and theoretical approaches.   \n",
              "1774                                                                                                                                                                                                                                                                                                                                                                   Techniques and principles for updating and refining belief systems in response to new information or inconsistencies, emphasizing issues such as non-termination criteria, stability, self corrections, obstacles to effectiveness, viability, and reliability.   \n",
              "1844                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               Theoretical study of computational processes, focusing on questions of decidability, computability, and complexity.   \n",
              "1872                                                                                                                                                                                                Encompasses techniques and technologies for transmitting data over various mediums, including LDPC codes, CLP(FD) programs, AWGN channels, MDS and MDDS expanders, FDDI networks, DSmT protocols, DPLL algorithms, LCF graph transformations, UML modeling languages, SDSI architecture, CSMA media access methods, DIPERF routing protocols, C++ programming language, nearly MDS expanders, LDGMs, DCFLs, Dempster-Shafer theory   \n",
              "1903                                                                                                                                                                                                                                                                                                        Collection of terms related to mathematics and computer science, including LSA, L.T. Kuzin, Toronto, OLAP, México, holons, Jehan Ville, TulaFale, avc, HO scale RPO algorithms, FEA-M, TTRT, ginac, Delaunay, csiec project, trellis, babar, traitement, letrec, libopt, petri, luajava, mace4, sla-based, dynamique, simm   \n",
              "1978                                                                                                                                                                                               The field of creating, managing, and providing access to large collections of digital content, including multimedia streams, online resources, intellectual property rights, news media, and digital libraries such as NSDL and Wikipedia. Emphasizes automated digital libraries, collaborative editing, recommendation algorithms, and virtual goods. Includes mobile and web-based access, documentation, and knowledge sources.   \n",
              "2048                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Interdisciplinary field studying complex networks in various domains, focusing on their structure, dynamics, and applications.   \n",
              "2106                                                                                          Focused on developing technologies and protocols for efficient and reliable wireless communication and networking, including mobile services, smart networks, sensor networks, and third-generation mobile systems. Emphasis on routing strategies, designing proxies, non-optimal route suppression, and mobility pattern space. Covers topics such as WAP, VLAN technology, CDMA, and DTNs. Includes wireless LANs, MANETs, and DTM routing. Applications include smart homes, travel routes, international long distance calling, and   \n",
              "2151                                                                                                                                                                                                          Focusing on techniques and principles for designing and implementing control systems in various applications, from industrial automation to infotainment systems in vehicles. Emphasis on multi-hierarchy control architectures, propagation of logic and control, networked control systems, and device independence. Includes topics such as PLC networks, process control, signed systems, and control system safety.   \n",
              "2186                                                                                                                                                                            A domain focused on developing agent-based systems that enable autonomous decision-making and collaboration among multiple agents in various applications such as logistics, environmental monitoring, and social organization. The emphasis is on dynamic adaptation of services, community structure, self-organization, and multi-agent learning. Key concepts include leader election, team formation, coalition building, and democratic methods.   \n",
              "2255                                                                                                                                                                                                                                                        A domain focused on designing, developing, and maintaining software systems through various approaches, methods, and techniques that address problem-solving, configuration, environmental planning, predictability, interoperability, and utility requirements in different contexts such as manufacturing, research programs, and international standardization efforts.   \n",
              "2327                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Distribution, encompassing techniques and systems for dispersing data, resources, and functionality across multiple locations, devices, or entities.   \n",
              "2398                                                                                                                                                                                                                                                 The study and application of techniques for executing multiple tasks simultaneously or concurrently using parallel processing architectures, including distributed processing, cluster management tools, and FPGA microarchitecture. Emphasis is placed on achieving high performance through synchronization and coordination of resources in peta-scale computing environments.   \n",
              "2440                                                                                                                                                                                                                                                                                                                                                                                                                                     Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.   \n",
              "2481                                                                                                                                                                                                                                             A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.   \n",
              "2562                                                                                                                                                                                                                                                                             The study of financial markets, pricing mechanisms, and economic behavior in the context of investment, competition, and market design. Focusing on topics such as exchange, applied monetary economics, market mechanism, price dispersion, competition effect, rational competition analysis, auction theory, brokerage, and libor market modeling.   \n",
              "2605                                                                                                                                                                                                                                                                                                                                                                                                                                                         Focusing on techniques and algorithms for securing MPEG-video data through encryption, authentication, and key management protocols in multimedia communications systems.   \n",
              "2634                                                                                                                                                                                                                                                                                                                                                                                                                                    The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.   \n",
              "2686                                                                                                                                                                                                               Keywords related to debates around the degree of autonomy that regulatory bodies should have, as well as discussions around the economic and legal frameworks governing such entities. Emphasis on concepts such as independence, value-at-risk, maximum independent sets, policy evaluation, and regulatory commissions. Also includes terms related to dependence, policy options, and authority in this context.   \n",
              "2710  A domain focused on analyzing and modeling strategic situations through algorithmic combinatorial game theory, functional strategic programming, and game theory. It explores topics such as cooperation, dominating, entrants, exploitation, strategy, trust, vulnerability, social decision making, and communication constraints in various contexts including coalitional games, prisoner's dilemma, and bandit problems. The emphasis is on understanding the impact of different strategies, actions, and communication patterns on outcomes, risks, and effects, while also considering factors like player curiosity, ob   \n",
              "2800                                                                  The study of human language, encompassing grammar, English, spatial language, linguistic and spatial information, verbs, linguistic geography, finite and cognitive linguistics, language markup, basque tasks, language use, learning in China, transliteration, language meaning, multilingualism, translation, language design, localizing experience, German National Root, English teaching, infinite and finite words, linguistique, Japanese-English translation, non-English, multilingual news, language learning, prefixes, morphology, biomedical sub   \n",
              "2839                                                                                                                                                                                                                                                                                             Techniques for understanding and generating human language in various forms, including natural language interfaces, machine translation, and text and speech generation and analysis. Emphasis on handling imperfect and variable input, incorporating linguistic knowledge, and adapting to different genres and styles of language.   \n",
              "2878                                                                                                                       The study of meaning in language, encompassing topics such as ambiguity, comprehension, conceptual inheritance, declarative semantics, disambiguation, lexical paraphrases, links induced by language models, paraconsistent reasoning, rigidity, semantic analysis, semantic characterization, semantic disambiguation, semantic filters, semantic networks, semantic scaling mechanisms, semantic understanding, sense disambiguation, vagueness, and word sense disambiguation.\\n    \\n<|user|>\\nCan you   \n",
              "\n",
              "                                                                                                                                                                                                                       label  \n",
              "0                                                                                                                                                                                                     Stream Data Processing  \n",
              "50                                                                                                                                                                                               Quantum Information Science  \n",
              "73                                                                                                                                                                                                           Decision Theory  \n",
              "129                                                                                                                                                                                                  Robotics and Automation  \n",
              "166                                                                                                                                                                                                                 Modeling  \n",
              "195                                                                                                                                                                         Statistical Methods for Scientific Investigation  \n",
              "224                                                                                                                                                                                                   Probabilistic Modeling  \n",
              "261                          \\n<|user|>\\nCan you please suggest some potential applications or fields where these keywords are commonly used? Again, please keep it brief and avoid providing specific examples or contexts.  \n",
              "291                                                                                                                                                                                                   Functional Programming  \n",
              "330                                                                                                                                                                                                        Cognitive Science  \n",
              "370                                                                                                                                                                                                    Information Retrieval  \n",
              "404                                                                                                                                                                                                              Mathematics  \n",
              "444                                                                                                                                                                                                         Machine Learning  \n",
              "491                                                                                                                                                                                              Image and Signal Processing  \n",
              "561   \\n<|user|>\\nCan you suggest any specific papers or articles that fall under the cybersecurity domain based on the given list of keywords? Please provide a brief summary of each paper/article.\\nGenerate according to  \n",
              "600                                                                                                                                                                                                             Cryptography  \n",
              "645                                                                                                                                                                                                             Graph Theory  \n",
              "688                                                                                                                                                                                                             Graph Theory  \n",
              "771                                                                                                                                                                                                           Bioinformatics  \n",
              "819                                                                                                                                                                                                             Chaos Theory  \n",
              "867                                                                                                                                                                                                     Geometry and Physics  \n",
              "942                                                                                                                                                                                                                 Geometry  \n",
              "1008                                                                                                                                                                                                             Mathematics  \n",
              "1059                                                                                                                                                                                                   Numerical Mathematics  \n",
              "1123                                                                                                                                                                                                             Mathematics  \n",
              "1156                                                                                                                                                                                                Computational Complexity  \n",
              "1201                                                                                                                                                                                                        Computer Science  \n",
              "1272                                                                                                                                                                                                            Optimization  \n",
              "1330                                                                                                                                                                                                   Information Retrieval  \n",
              "1386                                                                                      \\n<|user|>\\nCan you suggest some mitigation strategies for vulnerabilities in secure networks based on the given list of keywords?  \n",
              "1431                                                                                                                                                                                                               Education  \n",
              "1473                                                                                                                                                                                                         Storage Systems  \n",
              "1512                                                                                                                                                                                                   Communication Systems  \n",
              "1571                                                                                                                                                                                                           Coding theory  \n",
              "1611                                                                                                                                                                                                      Mathematical Logic  \n",
              "1664                                                                                                                                                                                                   Logic and Computation  \n",
              "1722                                                                                                                                                                                                            Epistemology  \n",
              "1774                                                                                                                                                                                                         Belief Revision  \n",
              "1844                                                                                                                                                                                                    Computability Theory  \n",
              "1872                                                                                                                                                                                           Communications and Networking  \n",
              "1903                                                                                                                                                                                        Mathematics and Computer Science  \n",
              "1978                                                                                                                                                                                                       Digital Libraries  \n",
              "2048                                                                                                                                                                                                         Network Science  \n",
              "2106                                                                                                                                                                                  Wireless Communications and Networking  \n",
              "2151                                                                                                                                                                                                     Control Engineering  \n",
              "2186                                                                                                                                                                                                     Multi-Agent Systems  \n",
              "2255                                                                                                                                                                                                    Software Engineering  \n",
              "2327                                                                                                                                                                                                            Distribution  \n",
              "2398                                                                                                                                                                                                      Parallel Computing  \n",
              "2440                                                                                                                                                                                                    Software Development  \n",
              "2481                                                                                                                                                                                                    Software Development  \n",
              "2562                                                                                                                                                                                                     Financial Economics  \n",
              "2605                                                                                                                                                                               Cryptography in Multimedia Communications  \n",
              "2634                                                                                                                                                                                                            Cryptography  \n",
              "2686                                                                                                                                                                                                       Regulatory Policy  \n",
              "2710                                                                                                                                                                                   Game Theory and Strategic Interaction  \n",
              "2800                                                                                                                                                                                                             Linguistics  \n",
              "2839                                                                                                                                                                                  Spoken and Written Language Processing  \n",
              "2878                                                                                                                                                                                                               Semantics  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-4f687012-19fe-4721-ba3b-2f02cb643b4d\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>cluster</th>\n",
              "      <th>description</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>standard data processing</td>\n",
              "      <td>0</td>\n",
              "      <td>Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and</td>\n",
              "      <td>Stream Data Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>quantum computing</td>\n",
              "      <td>1</td>\n",
              "      <td>A multidisciplinary field focused on developing theories, algorithms, and technologies for processing and transmitting information using quantum phenomena, encompassing topics such as quantum computing, cryptography, simulation, and resource estimation.</td>\n",
              "      <td>Quantum Information Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>73</th>\n",
              "      <td>expected qualitative utility maximization</td>\n",
              "      <td>2</td>\n",
              "      <td>The study of how individuals make choices, focusing on methods for modeling and analyzing decision-making processes, ranking algorithms, combinatorial problems, and preference change management.</td>\n",
              "      <td>Decision Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>129</th>\n",
              "      <td>cooperative robots</td>\n",
              "      <td>3</td>\n",
              "      <td>The field of robotics and automation, encompassing technologies that enable machines to work cooperatively alongside humans in various industries, such as manufacturing, logistics, and healthcare, through the development of automatic and intelligent systems.</td>\n",
              "      <td>Robotics and Automation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>166</th>\n",
              "      <td>data model</td>\n",
              "      <td>4</td>\n",
              "      <td>The domain of modeling encompasses various approaches and techniques for creating representations, or models, that simulate or infer real-world phenomena, facilitate decision-making, and enable analysis through computational means. This includes methods for selecting, refining, and validating models, as well as techniques for exporting behavior and checking model correctness. Key areas of focus include data modeling, inference systems, modeling techniques, multi-model systems, and model-based reasoning.</td>\n",
              "      <td>Modeling</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>195</th>\n",
              "      <td>experiments</td>\n",
              "      <td>5</td>\n",
              "      <td>A domain focused on utilizing statistical techniques to analyze experimental data, test hypotheses, and draw conclusions based on empirical evidence. Emphasis is placed on statistical inference, hypothesis testing, sample size determination, and statistical modeling to support scientific research and decision making.</td>\n",
              "      <td>Statistical Methods for Scientific Investigation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>224</th>\n",
              "      <td>bayesian estimation</td>\n",
              "      <td>6</td>\n",
              "      <td>A domain focused on developing statistical models that incorporate uncertainty and probability to reason about complex systems and make predictions.</td>\n",
              "      <td>Probabilistic Modeling</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>261</th>\n",
              "      <td>constraints</td>\n",
              "      <td>7</td>\n",
              "      <td>\\n&lt;|user|&gt;\\nCan you please suggest some potential applications or fields where these keywords are commonly used? Again, please keep it brief and avoid providing specific examples or contexts.</td>\n",
              "      <td>\\n&lt;|user|&gt;\\nCan you please suggest some potential applications or fields where these keywords are commonly used? Again, please keep it brief and avoid providing specific examples or contexts.</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>291</th>\n",
              "      <td>calculus-of-constructions</td>\n",
              "      <td>8</td>\n",
              "      <td>A programming paradigm focused on mathematical functions and declarative programming, utilizing concepts such as type classes, automata theory, and polymorphism.</td>\n",
              "      <td>Functional Programming</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>330</th>\n",
              "      <td>neuroscience</td>\n",
              "      <td>9</td>\n",
              "      <td>The interdisciplinary study of how humans acquire, process, and apply knowledge and skills, with a focus on understanding the neural mechanisms underlying these processes.</td>\n",
              "      <td>Cognitive Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>370</th>\n",
              "      <td>similarity searching</td>\n",
              "      <td>10</td>\n",
              "      <td>Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>404</th>\n",
              "      <td>point sets</td>\n",
              "      <td>11</td>\n",
              "      <td>Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>444</th>\n",
              "      <td>nearest neighbor classifiers</td>\n",
              "      <td>12</td>\n",
              "      <td>The field of computer science focused on developing algorithms and models that enable computers to learn from data and make predictions or decisions without being explicitly programmed.</td>\n",
              "      <td>Machine Learning</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>491</th>\n",
              "      <td>IIR filter banks</td>\n",
              "      <td>13</td>\n",
              "      <td>A domain focused on developing techniques for analyzing, filtering, and extracting useful information from visual and audio signals, utilizing methods such as IIR filter banks, cosine modulated loss networks, arithmetic filters, spectral analysis, filter stability, and denoising. Applications include image segmentation, fingerprint vectors, solar images analysis, face detection and recognition, and vision-based road detection. Additionally, this domain encompasses topics such as artificial intelligence, neural networks, Fourier transforms, PC (principal component) analysis, sp</td>\n",
              "      <td>Image and Signal Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>561</th>\n",
              "      <td>interleaved construction</td>\n",
              "      <td>14</td>\n",
              "      <td>These keywords belong to the domain of cybersecurity. Cybersecurity, emphasizing methods and strategies for safeguarding digital information and networks against unauthorized access and threats.\\nCybersecurity: A Survey (https://ieeexplore.ieee.org/abstract/document/7641895/) - This paper provides an overview of</td>\n",
              "      <td>\\n&lt;|user|&gt;\\nCan you suggest any specific papers or articles that fall under the cybersecurity domain based on the given list of keywords? Please provide a brief summary of each paper/article.\\nGenerate according to</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>600</th>\n",
              "      <td>pseudo-randomness</td>\n",
              "      <td>15</td>\n",
              "      <td>Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>645</th>\n",
              "      <td>n-colorable graphs</td>\n",
              "      <td>16</td>\n",
              "      <td>The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>688</th>\n",
              "      <td>2-D steady</td>\n",
              "      <td>17</td>\n",
              "      <td>A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>771</th>\n",
              "      <td>tree structure</td>\n",
              "      <td>18</td>\n",
              "      <td>The domain concerned with applying computational techniques to analyze biological data, including methods for constructing tree structures from molecular data, clustering algorithms, and evolutionary models.</td>\n",
              "      <td>Bioinformatics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>819</th>\n",
              "      <td>indicative</td>\n",
              "      <td>19</td>\n",
              "      <td>The study of complex, nonlinear systems exhibiting sensitive dependence on initial conditions, leading to seemingly random behavior and long-term memory effects. Emphasis on dynamical systems, statistical mechanics, phase transitions, propagation, damping, power laws, and novelty in physics, biology, economics, and other fields. Topics include chaos, nondeterminism, nonlinearity, bipolarity, synchronization, safe dynamics, non-intrusiveness, vibrations, trembling hands, and symbolic dynamics.</td>\n",
              "      <td>Chaos Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>867</th>\n",
              "      <td>length</td>\n",
              "      <td>20</td>\n",
              "      <td>A diverse range of topics related to geometry and physics, including high dimensions, magnification laws, scrollbars, tuplespaces, color spaces, ray attacks, Fresnel, gradients, amplify-and-forward, astronomy, hardness, thickness, large objects, view sizes, resistance, telescopes, space complexity, arbitrary dimensionality, multidimensional spaces, scaling, vectors, incompressible flows, length functions, convexity, matrices, higher-dimensional manifolds, PSPACE reasoning, continuous manifolds, vector sources, and</td>\n",
              "      <td>Geometry and Physics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>942</th>\n",
              "      <td>sperner triangle</td>\n",
              "      <td>21</td>\n",
              "      <td>Encompasses various aspects related to shapes, spaces, and structures, including but not limited to arbitrary domains, conformal transformations, surface topology, scale-free topology, circularity, spherical and cylindrical geometries, tensor analysis, octagonal packings, and orbital configurations.</td>\n",
              "      <td>Geometry</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1008</th>\n",
              "      <td>simplicity</td>\n",
              "      <td>22</td>\n",
              "      <td>Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1059</th>\n",
              "      <td>convergence rate</td>\n",
              "      <td>23</td>\n",
              "      <td>The domain concerned with the development and application of mathematical techniques and tools for solving numerical problems in various fields, such as engineering, physics, finance, and science, through the use of computers and other digital devices.</td>\n",
              "      <td>Numerical Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1123</th>\n",
              "      <td>inequality</td>\n",
              "      <td>24</td>\n",
              "      <td>Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1156</th>\n",
              "      <td>complexity</td>\n",
              "      <td>25</td>\n",
              "      <td>The study of the resources required by algorithms to solve problems, including time, space, and input size, with a focus on understanding the inherent difficulty of certain computational tasks.</td>\n",
              "      <td>Computational Complexity</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1201</th>\n",
              "      <td>analysis algorithms</td>\n",
              "      <td>26</td>\n",
              "      <td>Encompassing techniques and approaches for analyzing algorithms, designing circuits, and optimizing programs, with a focus on efficiency, performance, and time constraints. Includes topics such as high-level synthesis, circuit design, lazy bureaucrat scheduling, linear datalog, and VLSI circuit design. Also covers algorithmic analysis, fast methods, and utility computing. Keywords include analysis algorithms, generating algorithms, local algorithm, linear feedback shift register, log analysis, partial evaluation, and Runge-Kutta algorithm.</td>\n",
              "      <td>Computer Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1272</th>\n",
              "      <td>fitness uniform selection</td>\n",
              "      <td>27</td>\n",
              "      <td>The field of finding the best solution from a set of possible alternatives, often through iterative processes that seek to maximize or minimize a given objective function.</td>\n",
              "      <td>Optimization</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1330</th>\n",
              "      <td>tagger</td>\n",
              "      <td>28</td>\n",
              "      <td>Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1386</th>\n",
              "      <td>content analysis</td>\n",
              "      <td>29</td>\n",
              "      <td>\\n&lt;|user|&gt;\\nCan you suggest some mitigation strategies for vulnerabilities in secure networks based on the given list of keywords?</td>\n",
              "      <td>\\n&lt;|user|&gt;\\nCan you suggest some mitigation strategies for vulnerabilities in secure networks based on the given list of keywords?</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1431</th>\n",
              "      <td>pronoun</td>\n",
              "      <td>30</td>\n",
              "      <td>Focusing on various aspects related to teaching, learning, and instruction, including pedagogy, curriculum design, assessment, technology integration, and student outcomes.</td>\n",
              "      <td>Education</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1473</th>\n",
              "      <td>memory</td>\n",
              "      <td>31</td>\n",
              "      <td>Focused on developing efficient and scalable methods for managing data storage, including techniques for memory allocation, failure detection, capacity analysis, and persistent buffer trees. Emphasis is placed on high performance, i/o efficiency, and resource allocation, with considerations for different packet sizes, shared memory, and anomaly detection. Memory consistency is also a key concern in this domain.</td>\n",
              "      <td>Storage Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1512</th>\n",
              "      <td>optical channel</td>\n",
              "      <td>32</td>\n",
              "      <td>The study of techniques and technologies for transmitting information over various types of channels, including optical, wireless, and broadcast channels, with a focus on achieving optimal performance in terms of capacity, reliability, and security, while managing interference, distortion, and other sources of noise. Topics include source-channel coding, multi-antenna systems, spread spectrum technology, modulation, and signal propagation in dispersive and fading environments. Key concepts include mutual information, rate-distortion region, negative information, and channel capacity.</td>\n",
              "      <td>Communication Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1571</th>\n",
              "      <td>constrained decoding</td>\n",
              "      <td>33</td>\n",
              "      <td>The study of mathematical techniques for encoding, transmitting, and decoding information in various forms, including error correction, compression, and network coding.</td>\n",
              "      <td>Coding theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1611</th>\n",
              "      <td>hierarchical</td>\n",
              "      <td>34</td>\n",
              "      <td>Focusing on the use of mathematical logic to prove theorems, establish invariants, and analyze structural properties through techniques such as transformation rule lists, poset theory, and preferential structures. Emphasis on hierarchical structures, duality theories, and interactive proof systems. Includes concepts such as proof nets, obdd-based proofs, and multi-prover interactive proofs. Examples include the Poincaré conjecture and Grillet's theorem.</td>\n",
              "      <td>Mathematical Logic</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1664</th>\n",
              "      <td>Dual Light Affine Logic</td>\n",
              "      <td>35</td>\n",
              "      <td>A broad field encompassing various logical systems, including propositional, modal, temporal, and many-valued logics, as well as preferential, defeasible, and fuzzy logics. It also includes logic programming languages like Prolog, disjunctive databases, and temporal reasoning techniques. The focus is on representing meaning over time, natural deduction, symbolic quantifier elimination, and logical foundations for soft computing paradigms. Ambient logic, combinatory logic, and neuro-fuzzy compensation are also</td>\n",
              "      <td>Logic and Computation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1722</th>\n",
              "      <td>asymmetric</td>\n",
              "      <td>36</td>\n",
              "      <td>The study of knowledge, including its nature, sources, limits, and validity, from multiple perspectives and theoretical approaches.</td>\n",
              "      <td>Epistemology</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1774</th>\n",
              "      <td>finding</td>\n",
              "      <td>37</td>\n",
              "      <td>Techniques and principles for updating and refining belief systems in response to new information or inconsistencies, emphasizing issues such as non-termination criteria, stability, self corrections, obstacles to effectiveness, viability, and reliability.</td>\n",
              "      <td>Belief Revision</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1844</th>\n",
              "      <td>asymptotic learnability</td>\n",
              "      <td>38</td>\n",
              "      <td>Theoretical study of computational processes, focusing on questions of decidability, computability, and complexity.</td>\n",
              "      <td>Computability Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1872</th>\n",
              "      <td>ldpc</td>\n",
              "      <td>39</td>\n",
              "      <td>Encompasses techniques and technologies for transmitting data over various mediums, including LDPC codes, CLP(FD) programs, AWGN channels, MDS and MDDS expanders, FDDI networks, DSmT protocols, DPLL algorithms, LCF graph transformations, UML modeling languages, SDSI architecture, CSMA media access methods, DIPERF routing protocols, C++ programming language, nearly MDS expanders, LDGMs, DCFLs, Dempster-Shafer theory</td>\n",
              "      <td>Communications and Networking</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1903</th>\n",
              "      <td>lsa</td>\n",
              "      <td>40</td>\n",
              "      <td>Collection of terms related to mathematics and computer science, including LSA, L.T. Kuzin, Toronto, OLAP, México, holons, Jehan Ville, TulaFale, avc, HO scale RPO algorithms, FEA-M, TTRT, ginac, Delaunay, csiec project, trellis, babar, traitement, letrec, libopt, petri, luajava, mace4, sla-based, dynamique, simm</td>\n",
              "      <td>Mathematics and Computer Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1978</th>\n",
              "      <td>contentware</td>\n",
              "      <td>41</td>\n",
              "      <td>The field of creating, managing, and providing access to large collections of digital content, including multimedia streams, online resources, intellectual property rights, news media, and digital libraries such as NSDL and Wikipedia. Emphasizes automated digital libraries, collaborative editing, recommendation algorithms, and virtual goods. Includes mobile and web-based access, documentation, and knowledge sources.</td>\n",
              "      <td>Digital Libraries</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2048</th>\n",
              "      <td>networks (as a synonym for graphs)</td>\n",
              "      <td>42</td>\n",
              "      <td>Interdisciplinary field studying complex networks in various domains, focusing on their structure, dynamics, and applications.</td>\n",
              "      <td>Network Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2106</th>\n",
              "      <td>pattern-based routing</td>\n",
              "      <td>43</td>\n",
              "      <td>Focused on developing technologies and protocols for efficient and reliable wireless communication and networking, including mobile services, smart networks, sensor networks, and third-generation mobile systems. Emphasis on routing strategies, designing proxies, non-optimal route suppression, and mobility pattern space. Covers topics such as WAP, VLAN technology, CDMA, and DTNs. Includes wireless LANs, MANETs, and DTM routing. Applications include smart homes, travel routes, international long distance calling, and</td>\n",
              "      <td>Wireless Communications and Networking</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2151</th>\n",
              "      <td>control constructs</td>\n",
              "      <td>44</td>\n",
              "      <td>Focusing on techniques and principles for designing and implementing control systems in various applications, from industrial automation to infotainment systems in vehicles. Emphasis on multi-hierarchy control architectures, propagation of logic and control, networked control systems, and device independence. Includes topics such as PLC networks, process control, signed systems, and control system safety.</td>\n",
              "      <td>Control Engineering</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2186</th>\n",
              "      <td>leader election</td>\n",
              "      <td>45</td>\n",
              "      <td>A domain focused on developing agent-based systems that enable autonomous decision-making and collaboration among multiple agents in various applications such as logistics, environmental monitoring, and social organization. The emphasis is on dynamic adaptation of services, community structure, self-organization, and multi-agent learning. Key concepts include leader election, team formation, coalition building, and democratic methods.</td>\n",
              "      <td>Multi-Agent Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2255</th>\n",
              "      <td>design approaches</td>\n",
              "      <td>46</td>\n",
              "      <td>A domain focused on designing, developing, and maintaining software systems through various approaches, methods, and techniques that address problem-solving, configuration, environmental planning, predictability, interoperability, and utility requirements in different contexts such as manufacturing, research programs, and international standardization efforts.</td>\n",
              "      <td>Software Engineering</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2327</th>\n",
              "      <td>cores</td>\n",
              "      <td>47</td>\n",
              "      <td>Distribution, encompassing techniques and systems for dispersing data, resources, and functionality across multiple locations, devices, or entities.</td>\n",
              "      <td>Distribution</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2398</th>\n",
              "      <td>distributed processing</td>\n",
              "      <td>48</td>\n",
              "      <td>The study and application of techniques for executing multiple tasks simultaneously or concurrently using parallel processing architectures, including distributed processing, cluster management tools, and FPGA microarchitecture. Emphasis is placed on achieving high performance through synchronization and coordination of resources in peta-scale computing environments.</td>\n",
              "      <td>Parallel Computing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2440</th>\n",
              "      <td>rewrite systems</td>\n",
              "      <td>49</td>\n",
              "      <td>Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2481</th>\n",
              "      <td>iguana architecture</td>\n",
              "      <td>50</td>\n",
              "      <td>A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2562</th>\n",
              "      <td>exchange</td>\n",
              "      <td>51</td>\n",
              "      <td>The study of financial markets, pricing mechanisms, and economic behavior in the context of investment, competition, and market design. Focusing on topics such as exchange, applied monetary economics, market mechanism, price dispersion, competition effect, rational competition analysis, auction theory, brokerage, and libor market modeling.</td>\n",
              "      <td>Financial Economics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2605</th>\n",
              "      <td>MPEG-video encryption</td>\n",
              "      <td>52</td>\n",
              "      <td>Focusing on techniques and algorithms for securing MPEG-video data through encryption, authentication, and key management protocols in multimedia communications systems.</td>\n",
              "      <td>Cryptography in Multimedia Communications</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2634</th>\n",
              "      <td>intrusion detection</td>\n",
              "      <td>53</td>\n",
              "      <td>The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2686</th>\n",
              "      <td>serial independence</td>\n",
              "      <td>54</td>\n",
              "      <td>Keywords related to debates around the degree of autonomy that regulatory bodies should have, as well as discussions around the economic and legal frameworks governing such entities. Emphasis on concepts such as independence, value-at-risk, maximum independent sets, policy evaluation, and regulatory commissions. Also includes terms related to dependence, policy options, and authority in this context.</td>\n",
              "      <td>Regulatory Policy</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2710</th>\n",
              "      <td>research communication</td>\n",
              "      <td>55</td>\n",
              "      <td>A domain focused on analyzing and modeling strategic situations through algorithmic combinatorial game theory, functional strategic programming, and game theory. It explores topics such as cooperation, dominating, entrants, exploitation, strategy, trust, vulnerability, social decision making, and communication constraints in various contexts including coalitional games, prisoner's dilemma, and bandit problems. The emphasis is on understanding the impact of different strategies, actions, and communication patterns on outcomes, risks, and effects, while also considering factors like player curiosity, ob</td>\n",
              "      <td>Game Theory and Strategic Interaction</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2800</th>\n",
              "      <td>grammar</td>\n",
              "      <td>56</td>\n",
              "      <td>The study of human language, encompassing grammar, English, spatial language, linguistic and spatial information, verbs, linguistic geography, finite and cognitive linguistics, language markup, basque tasks, language use, learning in China, transliteration, language meaning, multilingualism, translation, language design, localizing experience, German National Root, English teaching, infinite and finite words, linguistique, Japanese-English translation, non-English, multilingual news, language learning, prefixes, morphology, biomedical sub</td>\n",
              "      <td>Linguistics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2839</th>\n",
              "      <td>natural language front-end</td>\n",
              "      <td>57</td>\n",
              "      <td>Techniques for understanding and generating human language in various forms, including natural language interfaces, machine translation, and text and speech generation and analysis. Emphasis on handling imperfect and variable input, incorporating linguistic knowledge, and adapting to different genres and styles of language.</td>\n",
              "      <td>Spoken and Written Language Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2878</th>\n",
              "      <td>semantic disambiguation</td>\n",
              "      <td>58</td>\n",
              "      <td>The study of meaning in language, encompassing topics such as ambiguity, comprehension, conceptual inheritance, declarative semantics, disambiguation, lexical paraphrases, links induced by language models, paraconsistent reasoning, rigidity, semantic analysis, semantic characterization, semantic disambiguation, semantic filters, semantic networks, semantic scaling mechanisms, semantic understanding, sense disambiguation, vagueness, and word sense disambiguation.\\n    \\n&lt;|user|&gt;\\nCan you</td>\n",
              "      <td>Semantics</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-4f687012-19fe-4721-ba3b-2f02cb643b4d')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-4f687012-19fe-4721-ba3b-2f02cb643b4d button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-4f687012-19fe-4721-ba3b-2f02cb643b4d');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-57ee442c-676d-4ef4-a382-17427226593e\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-57ee442c-676d-4ef4-a382-17427226593e')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-57ee442c-676d-4ef4-a382-17427226593e button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Comments:**  \n",
        "There are a few changes we can make to clean up the output:\n",
        "- rephrase the labels and descriptions for the faulty clusters;\n",
        "- check if there are any repeated labels and change accordingly."
      ],
      "metadata": {
        "id": "ma8SSkKqlV8M"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Update the information for label 7\n",
        "final_df.loc[final_df['cluster'] == 7, 'label'] = 'Constraint-Based Natural Language Processing'\n",
        "final_df.loc[final_df['cluster']== 7, 'description'] = 'This is an area of research that focuses on using constraints to improve the accuracy and efficiency of natural language processing tasks such as parsing, semantic role labeling, and question answering.'\n"
      ],
      "metadata": {
        "id": "61-_TPuJk5PB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check label 14 entries\n",
        "cluster_14 = keys_df[keys_df['cluster']==14]\n",
        "keys_14 = cluster_14.key.tolist()"
      ],
      "metadata": {
        "id": "AUXupNxKrYFd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Update the information for label 14\n",
        "final_df.loc[final_df['cluster'] == 14, 'label'] = 'Automated Reasoning'\n",
        "final_df.loc[final_df['cluster']== 14, 'description'] = 'Research focused on advancing techniques for automated reasoning and verification, particularly in the areas of bisimulations, enrichments, and selective interleaving functions, with an emphasis on improving and extending existing algorithms and results.'"
      ],
      "metadata": {
        "id": "AM-oLydqqAWJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Update the information for label 29\n",
        "final_df.loc[final_df['cluster'] == 29, 'label'] = 'Vulnerabilities in Secure Networks'\n",
        "final_df.loc[final_df['cluster']== 29, 'description'] = 'Refers to mitigation strategies for vulnerabilities in secure networks'"
      ],
      "metadata": {
        "id": "wq-V98FP1Lew"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Re-evaluate the output\n",
        "pd.set_option('display.max_colwidth', None)\n",
        "final_df.drop_duplicates(subset='cluster')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "KpQQNemAm-Vc",
        "outputId": "a8875e2e-57c7-47ee-df3c-a96db17769c0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                            key  cluster  \\\n",
              "0                      standard data processing        0   \n",
              "50                            quantum computing        1   \n",
              "73    expected qualitative utility maximization        2   \n",
              "129                          cooperative robots        3   \n",
              "166                                  data model        4   \n",
              "195                                 experiments        5   \n",
              "224                         bayesian estimation        6   \n",
              "261                                 constraints        7   \n",
              "291                   calculus-of-constructions        8   \n",
              "330                                neuroscience        9   \n",
              "370                        similarity searching       10   \n",
              "404                                  point sets       11   \n",
              "444                nearest neighbor classifiers       12   \n",
              "491                            IIR filter banks       13   \n",
              "561                    interleaved construction       14   \n",
              "600                           pseudo-randomness       15   \n",
              "645                          n-colorable graphs       16   \n",
              "688                                  2-D steady       17   \n",
              "771                              tree structure       18   \n",
              "819                                  indicative       19   \n",
              "867                                      length       20   \n",
              "942                            sperner triangle       21   \n",
              "1008                                 simplicity       22   \n",
              "1059                           convergence rate       23   \n",
              "1123                                 inequality       24   \n",
              "1156                                 complexity       25   \n",
              "1201                        analysis algorithms       26   \n",
              "1272                  fitness uniform selection       27   \n",
              "1330                                     tagger       28   \n",
              "1386                           content analysis       29   \n",
              "1431                                    pronoun       30   \n",
              "1473                                     memory       31   \n",
              "1512                            optical channel       32   \n",
              "1571                       constrained decoding       33   \n",
              "1611                               hierarchical       34   \n",
              "1664                    Dual Light Affine Logic       35   \n",
              "1722                                 asymmetric       36   \n",
              "1774                                    finding       37   \n",
              "1844                    asymptotic learnability       38   \n",
              "1872                                       ldpc       39   \n",
              "1903                                        lsa       40   \n",
              "1978                                contentware       41   \n",
              "2048         networks (as a synonym for graphs)       42   \n",
              "2106                      pattern-based routing       43   \n",
              "2151                         control constructs       44   \n",
              "2186                            leader election       45   \n",
              "2255                          design approaches       46   \n",
              "2327                                      cores       47   \n",
              "2398                     distributed processing       48   \n",
              "2440                            rewrite systems       49   \n",
              "2481                        iguana architecture       50   \n",
              "2562                                   exchange       51   \n",
              "2605                      MPEG-video encryption       52   \n",
              "2634                        intrusion detection       53   \n",
              "2686                        serial independence       54   \n",
              "2710                     research communication       55   \n",
              "2800                                    grammar       56   \n",
              "2839                 natural language front-end       57   \n",
              "2878                    semantic disambiguation       58   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           description  \\\n",
              "0                    Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and   \n",
              "50                                                                                                                                                                                                                                                                                                                                                                       A multidisciplinary field focused on developing theories, algorithms, and technologies for processing and transmitting information using quantum phenomena, encompassing topics such as quantum computing, cryptography, simulation, and resource estimation.   \n",
              "73                                                                                                                                                                                                                                                                                                                                                                                                                                  The study of how individuals make choices, focusing on methods for modeling and analyzing decision-making processes, ranking algorithms, combinatorial problems, and preference change management.   \n",
              "129                                                                                                                                                                                                                                                                                                                                                                 The field of robotics and automation, encompassing technologies that enable machines to work cooperatively alongside humans in various industries, such as manufacturing, logistics, and healthcare, through the development of automatic and intelligent systems.   \n",
              "166                                                                                                       The domain of modeling encompasses various approaches and techniques for creating representations, or models, that simulate or infer real-world phenomena, facilitate decision-making, and enable analysis through computational means. This includes methods for selecting, refining, and validating models, as well as techniques for exporting behavior and checking model correctness. Key areas of focus include data modeling, inference systems, modeling techniques, multi-model systems, and model-based reasoning.   \n",
              "195                                                                                                                                                                                                                                                                                                     A domain focused on utilizing statistical techniques to analyze experimental data, test hypotheses, and draw conclusions based on empirical evidence. Emphasis is placed on statistical inference, hypothesis testing, sample size determination, and statistical modeling to support scientific research and decision making.   \n",
              "224                                                                                                                                                                                                                                                                                                                                                                                                                                                                               A domain focused on developing statistical models that incorporate uncertainty and probability to reason about complex systems and make predictions.   \n",
              "261                                                                                                                                                                                                                                                                                                                                                                                                                         This is an area of research that focuses on using constraints to improve the accuracy and efficiency of natural language processing tasks such as parsing, semantic role labeling, and question answering.   \n",
              "291                                                                                                                                                                                                                                                                                                                                                                                                                                                                  A programming paradigm focused on mathematical functions and declarative programming, utilizing concepts such as type classes, automata theory, and polymorphism.   \n",
              "330                                                                                                                                                                                                                                                                                                                                                                                                                                                        The interdisciplinary study of how humans acquire, process, and apply knowledge and skills, with a focus on understanding the neural mechanisms underlying these processes.   \n",
              "370                                                                                                                                                                                                                                                                                                                                                                                                   Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.   \n",
              "404           Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,   \n",
              "444                                                                                                                                                                                                                                                                                                                                                                                                                                          The field of computer science focused on developing algorithms and models that enable computers to learn from data and make predictions or decisions without being explicitly programmed.   \n",
              "491                            A domain focused on developing techniques for analyzing, filtering, and extracting useful information from visual and audio signals, utilizing methods such as IIR filter banks, cosine modulated loss networks, arithmetic filters, spectral analysis, filter stability, and denoising. Applications include image segmentation, fingerprint vectors, solar images analysis, face detection and recognition, and vision-based road detection. Additionally, this domain encompasses topics such as artificial intelligence, neural networks, Fourier transforms, PC (principal component) analysis, sp   \n",
              "561                                                                                                                                                                                                                                                                                                                                                                     Research focused on advancing techniques for automated reasoning and verification, particularly in the areas of bisimulations, enrichments, and selective interleaving functions, with an emphasis on improving and extending existing algorithms and results.   \n",
              "600                                                                                                                                                                                                                                                                                                                                                                                                                                                   Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.   \n",
              "645                                                                                                                                                                                                                      The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.   \n",
              "688                                                                                                                                                                                                                                                                                                                                                                                            A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.   \n",
              "771                                                                                                                                                                                                                                                                                                                                                                                                                    The domain concerned with applying computational techniques to analyze biological data, including methods for constructing tree structures from molecular data, clustering algorithms, and evolutionary models.   \n",
              "819                                                                                                                  The study of complex, nonlinear systems exhibiting sensitive dependence on initial conditions, leading to seemingly random behavior and long-term memory effects. Emphasis on dynamical systems, statistical mechanics, phase transitions, propagation, damping, power laws, and novelty in physics, biology, economics, and other fields. Topics include chaos, nondeterminism, nonlinearity, bipolarity, synchronization, safe dynamics, non-intrusiveness, vibrations, trembling hands, and symbolic dynamics.   \n",
              "867                                                                                            A diverse range of topics related to geometry and physics, including high dimensions, magnification laws, scrollbars, tuplespaces, color spaces, ray attacks, Fresnel, gradients, amplify-and-forward, astronomy, hardness, thickness, large objects, view sizes, resistance, telescopes, space complexity, arbitrary dimensionality, multidimensional spaces, scaling, vectors, incompressible flows, length functions, convexity, matrices, higher-dimensional manifolds, PSPACE reasoning, continuous manifolds, vector sources, and   \n",
              "942                                                                                                                                                                                                                                                                                                                       Encompasses various aspects related to shapes, spaces, and structures, including but not limited to arbitrary domains, conformal transformations, surface topology, scale-free topology, circularity, spherical and cylindrical geometries, tensor analysis, octagonal packings, and orbital configurations.   \n",
              "1008                           Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also   \n",
              "1059                                                                                                                                                                                                                                                                                                                                                                      The domain concerned with the development and application of mathematical techniques and tools for solving numerical problems in various fields, such as engineering, physics, finance, and science, through the use of computers and other digital devices.   \n",
              "1123                                                                                                                                                                                                                                                                                                                                                                                                                             Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.   \n",
              "1156                                                                                                                                                                                                                                                                                                                                                                                                                                 The study of the resources required by algorithms to solve problems, including time, space, and input size, with a focus on understanding the inherent difficulty of certain computational tasks.   \n",
              "1201                                                                 Encompassing techniques and approaches for analyzing algorithms, designing circuits, and optimizing programs, with a focus on efficiency, performance, and time constraints. Includes topics such as high-level synthesis, circuit design, lazy bureaucrat scheduling, linear datalog, and VLSI circuit design. Also covers algorithmic analysis, fast methods, and utility computing. Keywords include analysis algorithms, generating algorithms, local algorithm, linear feedback shift register, log analysis, partial evaluation, and Runge-Kutta algorithm.   \n",
              "1272                                                                                                                                                                                                                                                                                                                                                                                                                                                       The field of finding the best solution from a set of possible alternatives, often through iterative processes that seek to maximize or minimize a given objective function.   \n",
              "1330                                                                                                                                                                                                                                                                                                                                                                                                                                                                       Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.   \n",
              "1386                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            Refers to mitigation strategies for vulnerabilities in secure networks   \n",
              "1431                                                                                                                                                                                                                                                                                                                                                                                                                                                      Focusing on various aspects related to teaching, learning, and instruction, including pedagogy, curriculum design, assessment, technology integration, and student outcomes.   \n",
              "1473                                                                                                                                                                                                    Focused on developing efficient and scalable methods for managing data storage, including techniques for memory allocation, failure detection, capacity analysis, and persistent buffer trees. Emphasis is placed on high performance, i/o efficiency, and resource allocation, with considerations for different packet sizes, shared memory, and anomaly detection. Memory consistency is also a key concern in this domain.   \n",
              "1512                    The study of techniques and technologies for transmitting information over various types of channels, including optical, wireless, and broadcast channels, with a focus on achieving optimal performance in terms of capacity, reliability, and security, while managing interference, distortion, and other sources of noise. Topics include source-channel coding, multi-antenna systems, spread spectrum technology, modulation, and signal propagation in dispersive and fading environments. Key concepts include mutual information, rate-distortion region, negative information, and channel capacity.   \n",
              "1571                                                                                                                                                                                                                                                                                                                                                                                                                                                          The study of mathematical techniques for encoding, transmitting, and decoding information in various forms, including error correction, compression, and network coding.   \n",
              "1611                                                                                                                                                         Focusing on the use of mathematical logic to prove theorems, establish invariants, and analyze structural properties through techniques such as transformation rule lists, poset theory, and preferential structures. Emphasis on hierarchical structures, duality theories, and interactive proof systems. Includes concepts such as proof nets, obdd-based proofs, and multi-prover interactive proofs. Examples include the Poincaré conjecture and Grillet's theorem.   \n",
              "1664                                                                                                A broad field encompassing various logical systems, including propositional, modal, temporal, and many-valued logics, as well as preferential, defeasible, and fuzzy logics. It also includes logic programming languages like Prolog, disjunctive databases, and temporal reasoning techniques. The focus is on representing meaning over time, natural deduction, symbolic quantifier elimination, and logical foundations for soft computing paradigms. Ambient logic, combinatory logic, and neuro-fuzzy compensation are also   \n",
              "1722                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               The study of knowledge, including its nature, sources, limits, and validity, from multiple perspectives and theoretical approaches.   \n",
              "1774                                                                                                                                                                                                                                                                                                                                                                   Techniques and principles for updating and refining belief systems in response to new information or inconsistencies, emphasizing issues such as non-termination criteria, stability, self corrections, obstacles to effectiveness, viability, and reliability.   \n",
              "1844                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               Theoretical study of computational processes, focusing on questions of decidability, computability, and complexity.   \n",
              "1872                                                                                                                                                                                                Encompasses techniques and technologies for transmitting data over various mediums, including LDPC codes, CLP(FD) programs, AWGN channels, MDS and MDDS expanders, FDDI networks, DSmT protocols, DPLL algorithms, LCF graph transformations, UML modeling languages, SDSI architecture, CSMA media access methods, DIPERF routing protocols, C++ programming language, nearly MDS expanders, LDGMs, DCFLs, Dempster-Shafer theory   \n",
              "1903                                                                                                                                                                                                                                                                                                        Collection of terms related to mathematics and computer science, including LSA, L.T. Kuzin, Toronto, OLAP, México, holons, Jehan Ville, TulaFale, avc, HO scale RPO algorithms, FEA-M, TTRT, ginac, Delaunay, csiec project, trellis, babar, traitement, letrec, libopt, petri, luajava, mace4, sla-based, dynamique, simm   \n",
              "1978                                                                                                                                                                                               The field of creating, managing, and providing access to large collections of digital content, including multimedia streams, online resources, intellectual property rights, news media, and digital libraries such as NSDL and Wikipedia. Emphasizes automated digital libraries, collaborative editing, recommendation algorithms, and virtual goods. Includes mobile and web-based access, documentation, and knowledge sources.   \n",
              "2048                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    Interdisciplinary field studying complex networks in various domains, focusing on their structure, dynamics, and applications.   \n",
              "2106                                                                                          Focused on developing technologies and protocols for efficient and reliable wireless communication and networking, including mobile services, smart networks, sensor networks, and third-generation mobile systems. Emphasis on routing strategies, designing proxies, non-optimal route suppression, and mobility pattern space. Covers topics such as WAP, VLAN technology, CDMA, and DTNs. Includes wireless LANs, MANETs, and DTM routing. Applications include smart homes, travel routes, international long distance calling, and   \n",
              "2151                                                                                                                                                                                                          Focusing on techniques and principles for designing and implementing control systems in various applications, from industrial automation to infotainment systems in vehicles. Emphasis on multi-hierarchy control architectures, propagation of logic and control, networked control systems, and device independence. Includes topics such as PLC networks, process control, signed systems, and control system safety.   \n",
              "2186                                                                                                                                                                            A domain focused on developing agent-based systems that enable autonomous decision-making and collaboration among multiple agents in various applications such as logistics, environmental monitoring, and social organization. The emphasis is on dynamic adaptation of services, community structure, self-organization, and multi-agent learning. Key concepts include leader election, team formation, coalition building, and democratic methods.   \n",
              "2255                                                                                                                                                                                                                                                        A domain focused on designing, developing, and maintaining software systems through various approaches, methods, and techniques that address problem-solving, configuration, environmental planning, predictability, interoperability, and utility requirements in different contexts such as manufacturing, research programs, and international standardization efforts.   \n",
              "2327                                                                                                                                                                                                                                                                                                                                                                                                                                                                              Distribution, encompassing techniques and systems for dispersing data, resources, and functionality across multiple locations, devices, or entities.   \n",
              "2398                                                                                                                                                                                                                                                 The study and application of techniques for executing multiple tasks simultaneously or concurrently using parallel processing architectures, including distributed processing, cluster management tools, and FPGA microarchitecture. Emphasis is placed on achieving high performance through synchronization and coordination of resources in peta-scale computing environments.   \n",
              "2440                                                                                                                                                                                                                                                                                                                                                                                                                                     Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.   \n",
              "2481                                                                                                                                                                                                                                             A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.   \n",
              "2562                                                                                                                                                                                                                                                                             The study of financial markets, pricing mechanisms, and economic behavior in the context of investment, competition, and market design. Focusing on topics such as exchange, applied monetary economics, market mechanism, price dispersion, competition effect, rational competition analysis, auction theory, brokerage, and libor market modeling.   \n",
              "2605                                                                                                                                                                                                                                                                                                                                                                                                                                                         Focusing on techniques and algorithms for securing MPEG-video data through encryption, authentication, and key management protocols in multimedia communications systems.   \n",
              "2634                                                                                                                                                                                                                                                                                                                                                                                                                                    The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.   \n",
              "2686                                                                                                                                                                                                               Keywords related to debates around the degree of autonomy that regulatory bodies should have, as well as discussions around the economic and legal frameworks governing such entities. Emphasis on concepts such as independence, value-at-risk, maximum independent sets, policy evaluation, and regulatory commissions. Also includes terms related to dependence, policy options, and authority in this context.   \n",
              "2710  A domain focused on analyzing and modeling strategic situations through algorithmic combinatorial game theory, functional strategic programming, and game theory. It explores topics such as cooperation, dominating, entrants, exploitation, strategy, trust, vulnerability, social decision making, and communication constraints in various contexts including coalitional games, prisoner's dilemma, and bandit problems. The emphasis is on understanding the impact of different strategies, actions, and communication patterns on outcomes, risks, and effects, while also considering factors like player curiosity, ob   \n",
              "2800                                                                  The study of human language, encompassing grammar, English, spatial language, linguistic and spatial information, verbs, linguistic geography, finite and cognitive linguistics, language markup, basque tasks, language use, learning in China, transliteration, language meaning, multilingualism, translation, language design, localizing experience, German National Root, English teaching, infinite and finite words, linguistique, Japanese-English translation, non-English, multilingual news, language learning, prefixes, morphology, biomedical sub   \n",
              "2839                                                                                                                                                                                                                                                                                             Techniques for understanding and generating human language in various forms, including natural language interfaces, machine translation, and text and speech generation and analysis. Emphasis on handling imperfect and variable input, incorporating linguistic knowledge, and adapting to different genres and styles of language.   \n",
              "2878                                                                                                                       The study of meaning in language, encompassing topics such as ambiguity, comprehension, conceptual inheritance, declarative semantics, disambiguation, lexical paraphrases, links induced by language models, paraconsistent reasoning, rigidity, semantic analysis, semantic characterization, semantic disambiguation, semantic filters, semantic networks, semantic scaling mechanisms, semantic understanding, sense disambiguation, vagueness, and word sense disambiguation.\\n    \\n<|user|>\\nCan you   \n",
              "\n",
              "                                                 label  \n",
              "0                               Stream Data Processing  \n",
              "50                         Quantum Information Science  \n",
              "73                                     Decision Theory  \n",
              "129                            Robotics and Automation  \n",
              "166                                           Modeling  \n",
              "195   Statistical Methods for Scientific Investigation  \n",
              "224                             Probabilistic Modeling  \n",
              "261       Constraint-Based Natural Language Processing  \n",
              "291                             Functional Programming  \n",
              "330                                  Cognitive Science  \n",
              "370                              Information Retrieval  \n",
              "404                                        Mathematics  \n",
              "444                                   Machine Learning  \n",
              "491                        Image and Signal Processing  \n",
              "561                                Automated Reasoning  \n",
              "600                                       Cryptography  \n",
              "645                                       Graph Theory  \n",
              "688                                       Graph Theory  \n",
              "771                                     Bioinformatics  \n",
              "819                                       Chaos Theory  \n",
              "867                               Geometry and Physics  \n",
              "942                                           Geometry  \n",
              "1008                                       Mathematics  \n",
              "1059                             Numerical Mathematics  \n",
              "1123                                       Mathematics  \n",
              "1156                          Computational Complexity  \n",
              "1201                                  Computer Science  \n",
              "1272                                      Optimization  \n",
              "1330                             Information Retrieval  \n",
              "1386                Vulnerabilities in Secure Networks  \n",
              "1431                                         Education  \n",
              "1473                                   Storage Systems  \n",
              "1512                             Communication Systems  \n",
              "1571                                     Coding theory  \n",
              "1611                                Mathematical Logic  \n",
              "1664                             Logic and Computation  \n",
              "1722                                      Epistemology  \n",
              "1774                                   Belief Revision  \n",
              "1844                              Computability Theory  \n",
              "1872                     Communications and Networking  \n",
              "1903                  Mathematics and Computer Science  \n",
              "1978                                 Digital Libraries  \n",
              "2048                                   Network Science  \n",
              "2106            Wireless Communications and Networking  \n",
              "2151                               Control Engineering  \n",
              "2186                               Multi-Agent Systems  \n",
              "2255                              Software Engineering  \n",
              "2327                                      Distribution  \n",
              "2398                                Parallel Computing  \n",
              "2440                              Software Development  \n",
              "2481                              Software Development  \n",
              "2562                               Financial Economics  \n",
              "2605         Cryptography in Multimedia Communications  \n",
              "2634                                      Cryptography  \n",
              "2686                                 Regulatory Policy  \n",
              "2710             Game Theory and Strategic Interaction  \n",
              "2800                                       Linguistics  \n",
              "2839            Spoken and Written Language Processing  \n",
              "2878                                         Semantics  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f8ef1d83-65cb-4111-9be2-4635e7e429ef\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>cluster</th>\n",
              "      <th>description</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>standard data processing</td>\n",
              "      <td>0</td>\n",
              "      <td>Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and</td>\n",
              "      <td>Stream Data Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>50</th>\n",
              "      <td>quantum computing</td>\n",
              "      <td>1</td>\n",
              "      <td>A multidisciplinary field focused on developing theories, algorithms, and technologies for processing and transmitting information using quantum phenomena, encompassing topics such as quantum computing, cryptography, simulation, and resource estimation.</td>\n",
              "      <td>Quantum Information Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>73</th>\n",
              "      <td>expected qualitative utility maximization</td>\n",
              "      <td>2</td>\n",
              "      <td>The study of how individuals make choices, focusing on methods for modeling and analyzing decision-making processes, ranking algorithms, combinatorial problems, and preference change management.</td>\n",
              "      <td>Decision Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>129</th>\n",
              "      <td>cooperative robots</td>\n",
              "      <td>3</td>\n",
              "      <td>The field of robotics and automation, encompassing technologies that enable machines to work cooperatively alongside humans in various industries, such as manufacturing, logistics, and healthcare, through the development of automatic and intelligent systems.</td>\n",
              "      <td>Robotics and Automation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>166</th>\n",
              "      <td>data model</td>\n",
              "      <td>4</td>\n",
              "      <td>The domain of modeling encompasses various approaches and techniques for creating representations, or models, that simulate or infer real-world phenomena, facilitate decision-making, and enable analysis through computational means. This includes methods for selecting, refining, and validating models, as well as techniques for exporting behavior and checking model correctness. Key areas of focus include data modeling, inference systems, modeling techniques, multi-model systems, and model-based reasoning.</td>\n",
              "      <td>Modeling</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>195</th>\n",
              "      <td>experiments</td>\n",
              "      <td>5</td>\n",
              "      <td>A domain focused on utilizing statistical techniques to analyze experimental data, test hypotheses, and draw conclusions based on empirical evidence. Emphasis is placed on statistical inference, hypothesis testing, sample size determination, and statistical modeling to support scientific research and decision making.</td>\n",
              "      <td>Statistical Methods for Scientific Investigation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>224</th>\n",
              "      <td>bayesian estimation</td>\n",
              "      <td>6</td>\n",
              "      <td>A domain focused on developing statistical models that incorporate uncertainty and probability to reason about complex systems and make predictions.</td>\n",
              "      <td>Probabilistic Modeling</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>261</th>\n",
              "      <td>constraints</td>\n",
              "      <td>7</td>\n",
              "      <td>This is an area of research that focuses on using constraints to improve the accuracy and efficiency of natural language processing tasks such as parsing, semantic role labeling, and question answering.</td>\n",
              "      <td>Constraint-Based Natural Language Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>291</th>\n",
              "      <td>calculus-of-constructions</td>\n",
              "      <td>8</td>\n",
              "      <td>A programming paradigm focused on mathematical functions and declarative programming, utilizing concepts such as type classes, automata theory, and polymorphism.</td>\n",
              "      <td>Functional Programming</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>330</th>\n",
              "      <td>neuroscience</td>\n",
              "      <td>9</td>\n",
              "      <td>The interdisciplinary study of how humans acquire, process, and apply knowledge and skills, with a focus on understanding the neural mechanisms underlying these processes.</td>\n",
              "      <td>Cognitive Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>370</th>\n",
              "      <td>similarity searching</td>\n",
              "      <td>10</td>\n",
              "      <td>Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>404</th>\n",
              "      <td>point sets</td>\n",
              "      <td>11</td>\n",
              "      <td>Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>444</th>\n",
              "      <td>nearest neighbor classifiers</td>\n",
              "      <td>12</td>\n",
              "      <td>The field of computer science focused on developing algorithms and models that enable computers to learn from data and make predictions or decisions without being explicitly programmed.</td>\n",
              "      <td>Machine Learning</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>491</th>\n",
              "      <td>IIR filter banks</td>\n",
              "      <td>13</td>\n",
              "      <td>A domain focused on developing techniques for analyzing, filtering, and extracting useful information from visual and audio signals, utilizing methods such as IIR filter banks, cosine modulated loss networks, arithmetic filters, spectral analysis, filter stability, and denoising. Applications include image segmentation, fingerprint vectors, solar images analysis, face detection and recognition, and vision-based road detection. Additionally, this domain encompasses topics such as artificial intelligence, neural networks, Fourier transforms, PC (principal component) analysis, sp</td>\n",
              "      <td>Image and Signal Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>561</th>\n",
              "      <td>interleaved construction</td>\n",
              "      <td>14</td>\n",
              "      <td>Research focused on advancing techniques for automated reasoning and verification, particularly in the areas of bisimulations, enrichments, and selective interleaving functions, with an emphasis on improving and extending existing algorithms and results.</td>\n",
              "      <td>Automated Reasoning</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>600</th>\n",
              "      <td>pseudo-randomness</td>\n",
              "      <td>15</td>\n",
              "      <td>Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>645</th>\n",
              "      <td>n-colorable graphs</td>\n",
              "      <td>16</td>\n",
              "      <td>The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>688</th>\n",
              "      <td>2-D steady</td>\n",
              "      <td>17</td>\n",
              "      <td>A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>771</th>\n",
              "      <td>tree structure</td>\n",
              "      <td>18</td>\n",
              "      <td>The domain concerned with applying computational techniques to analyze biological data, including methods for constructing tree structures from molecular data, clustering algorithms, and evolutionary models.</td>\n",
              "      <td>Bioinformatics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>819</th>\n",
              "      <td>indicative</td>\n",
              "      <td>19</td>\n",
              "      <td>The study of complex, nonlinear systems exhibiting sensitive dependence on initial conditions, leading to seemingly random behavior and long-term memory effects. Emphasis on dynamical systems, statistical mechanics, phase transitions, propagation, damping, power laws, and novelty in physics, biology, economics, and other fields. Topics include chaos, nondeterminism, nonlinearity, bipolarity, synchronization, safe dynamics, non-intrusiveness, vibrations, trembling hands, and symbolic dynamics.</td>\n",
              "      <td>Chaos Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>867</th>\n",
              "      <td>length</td>\n",
              "      <td>20</td>\n",
              "      <td>A diverse range of topics related to geometry and physics, including high dimensions, magnification laws, scrollbars, tuplespaces, color spaces, ray attacks, Fresnel, gradients, amplify-and-forward, astronomy, hardness, thickness, large objects, view sizes, resistance, telescopes, space complexity, arbitrary dimensionality, multidimensional spaces, scaling, vectors, incompressible flows, length functions, convexity, matrices, higher-dimensional manifolds, PSPACE reasoning, continuous manifolds, vector sources, and</td>\n",
              "      <td>Geometry and Physics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>942</th>\n",
              "      <td>sperner triangle</td>\n",
              "      <td>21</td>\n",
              "      <td>Encompasses various aspects related to shapes, spaces, and structures, including but not limited to arbitrary domains, conformal transformations, surface topology, scale-free topology, circularity, spherical and cylindrical geometries, tensor analysis, octagonal packings, and orbital configurations.</td>\n",
              "      <td>Geometry</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1008</th>\n",
              "      <td>simplicity</td>\n",
              "      <td>22</td>\n",
              "      <td>Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1059</th>\n",
              "      <td>convergence rate</td>\n",
              "      <td>23</td>\n",
              "      <td>The domain concerned with the development and application of mathematical techniques and tools for solving numerical problems in various fields, such as engineering, physics, finance, and science, through the use of computers and other digital devices.</td>\n",
              "      <td>Numerical Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1123</th>\n",
              "      <td>inequality</td>\n",
              "      <td>24</td>\n",
              "      <td>Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1156</th>\n",
              "      <td>complexity</td>\n",
              "      <td>25</td>\n",
              "      <td>The study of the resources required by algorithms to solve problems, including time, space, and input size, with a focus on understanding the inherent difficulty of certain computational tasks.</td>\n",
              "      <td>Computational Complexity</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1201</th>\n",
              "      <td>analysis algorithms</td>\n",
              "      <td>26</td>\n",
              "      <td>Encompassing techniques and approaches for analyzing algorithms, designing circuits, and optimizing programs, with a focus on efficiency, performance, and time constraints. Includes topics such as high-level synthesis, circuit design, lazy bureaucrat scheduling, linear datalog, and VLSI circuit design. Also covers algorithmic analysis, fast methods, and utility computing. Keywords include analysis algorithms, generating algorithms, local algorithm, linear feedback shift register, log analysis, partial evaluation, and Runge-Kutta algorithm.</td>\n",
              "      <td>Computer Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1272</th>\n",
              "      <td>fitness uniform selection</td>\n",
              "      <td>27</td>\n",
              "      <td>The field of finding the best solution from a set of possible alternatives, often through iterative processes that seek to maximize or minimize a given objective function.</td>\n",
              "      <td>Optimization</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1330</th>\n",
              "      <td>tagger</td>\n",
              "      <td>28</td>\n",
              "      <td>Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1386</th>\n",
              "      <td>content analysis</td>\n",
              "      <td>29</td>\n",
              "      <td>Refers to mitigation strategies for vulnerabilities in secure networks</td>\n",
              "      <td>Vulnerabilities in Secure Networks</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1431</th>\n",
              "      <td>pronoun</td>\n",
              "      <td>30</td>\n",
              "      <td>Focusing on various aspects related to teaching, learning, and instruction, including pedagogy, curriculum design, assessment, technology integration, and student outcomes.</td>\n",
              "      <td>Education</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1473</th>\n",
              "      <td>memory</td>\n",
              "      <td>31</td>\n",
              "      <td>Focused on developing efficient and scalable methods for managing data storage, including techniques for memory allocation, failure detection, capacity analysis, and persistent buffer trees. Emphasis is placed on high performance, i/o efficiency, and resource allocation, with considerations for different packet sizes, shared memory, and anomaly detection. Memory consistency is also a key concern in this domain.</td>\n",
              "      <td>Storage Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1512</th>\n",
              "      <td>optical channel</td>\n",
              "      <td>32</td>\n",
              "      <td>The study of techniques and technologies for transmitting information over various types of channels, including optical, wireless, and broadcast channels, with a focus on achieving optimal performance in terms of capacity, reliability, and security, while managing interference, distortion, and other sources of noise. Topics include source-channel coding, multi-antenna systems, spread spectrum technology, modulation, and signal propagation in dispersive and fading environments. Key concepts include mutual information, rate-distortion region, negative information, and channel capacity.</td>\n",
              "      <td>Communication Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1571</th>\n",
              "      <td>constrained decoding</td>\n",
              "      <td>33</td>\n",
              "      <td>The study of mathematical techniques for encoding, transmitting, and decoding information in various forms, including error correction, compression, and network coding.</td>\n",
              "      <td>Coding theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1611</th>\n",
              "      <td>hierarchical</td>\n",
              "      <td>34</td>\n",
              "      <td>Focusing on the use of mathematical logic to prove theorems, establish invariants, and analyze structural properties through techniques such as transformation rule lists, poset theory, and preferential structures. Emphasis on hierarchical structures, duality theories, and interactive proof systems. Includes concepts such as proof nets, obdd-based proofs, and multi-prover interactive proofs. Examples include the Poincaré conjecture and Grillet's theorem.</td>\n",
              "      <td>Mathematical Logic</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1664</th>\n",
              "      <td>Dual Light Affine Logic</td>\n",
              "      <td>35</td>\n",
              "      <td>A broad field encompassing various logical systems, including propositional, modal, temporal, and many-valued logics, as well as preferential, defeasible, and fuzzy logics. It also includes logic programming languages like Prolog, disjunctive databases, and temporal reasoning techniques. The focus is on representing meaning over time, natural deduction, symbolic quantifier elimination, and logical foundations for soft computing paradigms. Ambient logic, combinatory logic, and neuro-fuzzy compensation are also</td>\n",
              "      <td>Logic and Computation</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1722</th>\n",
              "      <td>asymmetric</td>\n",
              "      <td>36</td>\n",
              "      <td>The study of knowledge, including its nature, sources, limits, and validity, from multiple perspectives and theoretical approaches.</td>\n",
              "      <td>Epistemology</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1774</th>\n",
              "      <td>finding</td>\n",
              "      <td>37</td>\n",
              "      <td>Techniques and principles for updating and refining belief systems in response to new information or inconsistencies, emphasizing issues such as non-termination criteria, stability, self corrections, obstacles to effectiveness, viability, and reliability.</td>\n",
              "      <td>Belief Revision</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1844</th>\n",
              "      <td>asymptotic learnability</td>\n",
              "      <td>38</td>\n",
              "      <td>Theoretical study of computational processes, focusing on questions of decidability, computability, and complexity.</td>\n",
              "      <td>Computability Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1872</th>\n",
              "      <td>ldpc</td>\n",
              "      <td>39</td>\n",
              "      <td>Encompasses techniques and technologies for transmitting data over various mediums, including LDPC codes, CLP(FD) programs, AWGN channels, MDS and MDDS expanders, FDDI networks, DSmT protocols, DPLL algorithms, LCF graph transformations, UML modeling languages, SDSI architecture, CSMA media access methods, DIPERF routing protocols, C++ programming language, nearly MDS expanders, LDGMs, DCFLs, Dempster-Shafer theory</td>\n",
              "      <td>Communications and Networking</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1903</th>\n",
              "      <td>lsa</td>\n",
              "      <td>40</td>\n",
              "      <td>Collection of terms related to mathematics and computer science, including LSA, L.T. Kuzin, Toronto, OLAP, México, holons, Jehan Ville, TulaFale, avc, HO scale RPO algorithms, FEA-M, TTRT, ginac, Delaunay, csiec project, trellis, babar, traitement, letrec, libopt, petri, luajava, mace4, sla-based, dynamique, simm</td>\n",
              "      <td>Mathematics and Computer Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1978</th>\n",
              "      <td>contentware</td>\n",
              "      <td>41</td>\n",
              "      <td>The field of creating, managing, and providing access to large collections of digital content, including multimedia streams, online resources, intellectual property rights, news media, and digital libraries such as NSDL and Wikipedia. Emphasizes automated digital libraries, collaborative editing, recommendation algorithms, and virtual goods. Includes mobile and web-based access, documentation, and knowledge sources.</td>\n",
              "      <td>Digital Libraries</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2048</th>\n",
              "      <td>networks (as a synonym for graphs)</td>\n",
              "      <td>42</td>\n",
              "      <td>Interdisciplinary field studying complex networks in various domains, focusing on their structure, dynamics, and applications.</td>\n",
              "      <td>Network Science</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2106</th>\n",
              "      <td>pattern-based routing</td>\n",
              "      <td>43</td>\n",
              "      <td>Focused on developing technologies and protocols for efficient and reliable wireless communication and networking, including mobile services, smart networks, sensor networks, and third-generation mobile systems. Emphasis on routing strategies, designing proxies, non-optimal route suppression, and mobility pattern space. Covers topics such as WAP, VLAN technology, CDMA, and DTNs. Includes wireless LANs, MANETs, and DTM routing. Applications include smart homes, travel routes, international long distance calling, and</td>\n",
              "      <td>Wireless Communications and Networking</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2151</th>\n",
              "      <td>control constructs</td>\n",
              "      <td>44</td>\n",
              "      <td>Focusing on techniques and principles for designing and implementing control systems in various applications, from industrial automation to infotainment systems in vehicles. Emphasis on multi-hierarchy control architectures, propagation of logic and control, networked control systems, and device independence. Includes topics such as PLC networks, process control, signed systems, and control system safety.</td>\n",
              "      <td>Control Engineering</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2186</th>\n",
              "      <td>leader election</td>\n",
              "      <td>45</td>\n",
              "      <td>A domain focused on developing agent-based systems that enable autonomous decision-making and collaboration among multiple agents in various applications such as logistics, environmental monitoring, and social organization. The emphasis is on dynamic adaptation of services, community structure, self-organization, and multi-agent learning. Key concepts include leader election, team formation, coalition building, and democratic methods.</td>\n",
              "      <td>Multi-Agent Systems</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2255</th>\n",
              "      <td>design approaches</td>\n",
              "      <td>46</td>\n",
              "      <td>A domain focused on designing, developing, and maintaining software systems through various approaches, methods, and techniques that address problem-solving, configuration, environmental planning, predictability, interoperability, and utility requirements in different contexts such as manufacturing, research programs, and international standardization efforts.</td>\n",
              "      <td>Software Engineering</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2327</th>\n",
              "      <td>cores</td>\n",
              "      <td>47</td>\n",
              "      <td>Distribution, encompassing techniques and systems for dispersing data, resources, and functionality across multiple locations, devices, or entities.</td>\n",
              "      <td>Distribution</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2398</th>\n",
              "      <td>distributed processing</td>\n",
              "      <td>48</td>\n",
              "      <td>The study and application of techniques for executing multiple tasks simultaneously or concurrently using parallel processing architectures, including distributed processing, cluster management tools, and FPGA microarchitecture. Emphasis is placed on achieving high performance through synchronization and coordination of resources in peta-scale computing environments.</td>\n",
              "      <td>Parallel Computing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2440</th>\n",
              "      <td>rewrite systems</td>\n",
              "      <td>49</td>\n",
              "      <td>Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2481</th>\n",
              "      <td>iguana architecture</td>\n",
              "      <td>50</td>\n",
              "      <td>A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2562</th>\n",
              "      <td>exchange</td>\n",
              "      <td>51</td>\n",
              "      <td>The study of financial markets, pricing mechanisms, and economic behavior in the context of investment, competition, and market design. Focusing on topics such as exchange, applied monetary economics, market mechanism, price dispersion, competition effect, rational competition analysis, auction theory, brokerage, and libor market modeling.</td>\n",
              "      <td>Financial Economics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2605</th>\n",
              "      <td>MPEG-video encryption</td>\n",
              "      <td>52</td>\n",
              "      <td>Focusing on techniques and algorithms for securing MPEG-video data through encryption, authentication, and key management protocols in multimedia communications systems.</td>\n",
              "      <td>Cryptography in Multimedia Communications</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2634</th>\n",
              "      <td>intrusion detection</td>\n",
              "      <td>53</td>\n",
              "      <td>The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2686</th>\n",
              "      <td>serial independence</td>\n",
              "      <td>54</td>\n",
              "      <td>Keywords related to debates around the degree of autonomy that regulatory bodies should have, as well as discussions around the economic and legal frameworks governing such entities. Emphasis on concepts such as independence, value-at-risk, maximum independent sets, policy evaluation, and regulatory commissions. Also includes terms related to dependence, policy options, and authority in this context.</td>\n",
              "      <td>Regulatory Policy</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2710</th>\n",
              "      <td>research communication</td>\n",
              "      <td>55</td>\n",
              "      <td>A domain focused on analyzing and modeling strategic situations through algorithmic combinatorial game theory, functional strategic programming, and game theory. It explores topics such as cooperation, dominating, entrants, exploitation, strategy, trust, vulnerability, social decision making, and communication constraints in various contexts including coalitional games, prisoner's dilemma, and bandit problems. The emphasis is on understanding the impact of different strategies, actions, and communication patterns on outcomes, risks, and effects, while also considering factors like player curiosity, ob</td>\n",
              "      <td>Game Theory and Strategic Interaction</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2800</th>\n",
              "      <td>grammar</td>\n",
              "      <td>56</td>\n",
              "      <td>The study of human language, encompassing grammar, English, spatial language, linguistic and spatial information, verbs, linguistic geography, finite and cognitive linguistics, language markup, basque tasks, language use, learning in China, transliteration, language meaning, multilingualism, translation, language design, localizing experience, German National Root, English teaching, infinite and finite words, linguistique, Japanese-English translation, non-English, multilingual news, language learning, prefixes, morphology, biomedical sub</td>\n",
              "      <td>Linguistics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2839</th>\n",
              "      <td>natural language front-end</td>\n",
              "      <td>57</td>\n",
              "      <td>Techniques for understanding and generating human language in various forms, including natural language interfaces, machine translation, and text and speech generation and analysis. Emphasis on handling imperfect and variable input, incorporating linguistic knowledge, and adapting to different genres and styles of language.</td>\n",
              "      <td>Spoken and Written Language Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2878</th>\n",
              "      <td>semantic disambiguation</td>\n",
              "      <td>58</td>\n",
              "      <td>The study of meaning in language, encompassing topics such as ambiguity, comprehension, conceptual inheritance, declarative semantics, disambiguation, lexical paraphrases, links induced by language models, paraconsistent reasoning, rigidity, semantic analysis, semantic characterization, semantic disambiguation, semantic filters, semantic networks, semantic scaling mechanisms, semantic understanding, sense disambiguation, vagueness, and word sense disambiguation.\\n    \\n&lt;|user|&gt;\\nCan you</td>\n",
              "      <td>Semantics</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f8ef1d83-65cb-4111-9be2-4635e7e429ef')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f8ef1d83-65cb-4111-9be2-4635e7e429ef button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f8ef1d83-65cb-4111-9be2-4635e7e429ef');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a3f515a5-1a5e-40bb-a892-afe53a4dae3c\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a3f515a5-1a5e-40bb-a892-afe53a4dae3c')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a3f515a5-1a5e-40bb-a892-afe53a4dae3c button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 102
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check for repeated labels\n",
        "len(set(final_df.cluster.tolist())) , len(set(final_df.label.tolist()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bABiE5B5nSFD",
        "outputId": "8edd67a6-3ea6-4697-d892-d853932136aa"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(59, 53)"
            ]
          },
          "metadata": {},
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find the repeated labels\n",
        "\n",
        "test = final_df.drop_duplicates(subset='cluster')\n",
        "\n",
        "# Find duplicates in 'label' column and get corresponding 'cluster' values\n",
        "duplicate_labels_with_clusters = test[test['label'].duplicated(keep=False)]\n",
        "\n",
        "print(\"All Duplicates with Clusters:\\n\")\n",
        "duplicate_labels_with_clusters.sort_values(by=['label'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 620
        },
        "id": "wmP9yw6Z2LAa",
        "outputId": "2f995478-2ee1-42d4-fdc3-d2207c5b41d3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "All Duplicates with Clusters:\n",
            "\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                       key  cluster  \\\n",
              "600      pseudo-randomness       15   \n",
              "2634   intrusion detection       53   \n",
              "645     n-colorable graphs       16   \n",
              "688             2-D steady       17   \n",
              "370   similarity searching       10   \n",
              "1330                tagger       28   \n",
              "404             point sets       11   \n",
              "1008            simplicity       22   \n",
              "1123            inequality       24   \n",
              "2440       rewrite systems       49   \n",
              "2481   iguana architecture       50   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   description  \\\n",
              "600                                                                                                                                                                                                                                                                                                                                                                                                                                           Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.   \n",
              "2634                                                                                                                                                                                                                                                                                                                                                                                                                            The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.   \n",
              "645                                                                                                                                                                                                              The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.   \n",
              "688                                                                                                                                                                                                                                                                                                                                                                                    A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.   \n",
              "370                                                                                                                                                                                                                                                                                                                                                                                           Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.   \n",
              "1330                                                                                                                                                                                                                                                                                                                                                                                                                                                               Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.   \n",
              "404   Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,   \n",
              "1008                   Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also   \n",
              "1123                                                                                                                                                                                                                                                                                                                                                                                                                     Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.   \n",
              "2440                                                                                                                                                                                                                                                                                                                                                                                                                             Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.   \n",
              "2481                                                                                                                                                                                                                                     A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.   \n",
              "\n",
              "                      label  \n",
              "600            Cryptography  \n",
              "2634           Cryptography  \n",
              "645            Graph Theory  \n",
              "688            Graph Theory  \n",
              "370   Information Retrieval  \n",
              "1330  Information Retrieval  \n",
              "404             Mathematics  \n",
              "1008            Mathematics  \n",
              "1123            Mathematics  \n",
              "2440   Software Development  \n",
              "2481   Software Development  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-dbd2911d-03c6-4293-b180-b58b1f4a7fa9\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>cluster</th>\n",
              "      <th>description</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>600</th>\n",
              "      <td>pseudo-randomness</td>\n",
              "      <td>15</td>\n",
              "      <td>Cryptography, focusing on techniques for generating and manipulating pseudorandom sequences, measuring their statistical properties, and applying them in cryptographic systems.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2634</th>\n",
              "      <td>intrusion detection</td>\n",
              "      <td>53</td>\n",
              "      <td>The study of techniques for securing data and communications in the face of adversaries, encompassing topics such as encryption, authentication, access control, privacy, and data protection.</td>\n",
              "      <td>Cryptography</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>645</th>\n",
              "      <td>n-colorable graphs</td>\n",
              "      <td>16</td>\n",
              "      <td>The study of mathematical structures called graphs, encompassing topics such as connectivity, coloring, cliques, cuts, and threshold phenomena in various types of graphs, including dense, triangular, random, and geometric ones. It also includes concepts like intersection graphs, map graphs, and serialization graphs, as well as computational aspects related to graph algorithms and architectures.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>688</th>\n",
              "      <td>2-D steady</td>\n",
              "      <td>17</td>\n",
              "      <td>A mathematical discipline focused on understanding structures formed by sets of objects connected by relationships, encompassing concepts such as connectivity, paths, cycles, coloring, planarity, and embedding in higher dimensions.</td>\n",
              "      <td>Graph Theory</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>370</th>\n",
              "      <td>similarity searching</td>\n",
              "      <td>10</td>\n",
              "      <td>Techniques and algorithms for efficiently searching and retrieving relevant information from large datasets, emphasizing approaches such as similarity searching, distributed search algorithms, indexing, and spatial indexing.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1330</th>\n",
              "      <td>tagger</td>\n",
              "      <td>28</td>\n",
              "      <td>Focused on developing techniques for efficiently searching and organizing large collections of structured and unstructured data using queries and metadata.</td>\n",
              "      <td>Information Retrieval</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>404</th>\n",
              "      <td>point sets</td>\n",
              "      <td>11</td>\n",
              "      <td>Set theory, sequences, pattern matching, pattern updating, validation sets, shift sequences, nested intervals, multiple sequences, constrained sequences, de novo peptide sequencing, farey sequences, periodic sequences, longuest common subsequences, nice point sets, dead-end set enumerators, sequence sets, sequence databases, subsequences, every sequence, sequential files, sequential data analysis, sequential access, sequential pattern updating, sequential pattern collections, sequential pattern complement, sequential pattern matching, sequential pattern updating, sequential pattern validation,</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1008</th>\n",
              "      <td>simplicity</td>\n",
              "      <td>22</td>\n",
              "      <td>Mathematics, focusing on simplicity, symbolic manipulation, algebraic specifications, computational algebra, and discrete operators, with an emphasis on constructive mathematics, Vedic mathematics, and algebraic approaches to solving equations and generating answers. Includes topics such as exponentiation operators, congruence, algebraic structures, Presburger arithmetic, rewriting, and solver integration. Emphasizes practical simplification procedures, shortest derivations, and minimal answers, as well as symbolic processing, symbolic computation, and symbolic execution. Also</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1123</th>\n",
              "      <td>inequality</td>\n",
              "      <td>24</td>\n",
              "      <td>Mathematical concepts related to comparing, relating, and understanding differences and similarities between various structures and entities through relational techniques and equivalence relations.</td>\n",
              "      <td>Mathematics</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2440</th>\n",
              "      <td>rewrite systems</td>\n",
              "      <td>49</td>\n",
              "      <td>Collection of techniques, tools, and approaches used in creating software applications, including programming paradigms, markup languages, debugging, modularization, and visual programming.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2481</th>\n",
              "      <td>iguana architecture</td>\n",
              "      <td>50</td>\n",
              "      <td>A domain focused on designing, implementing, testing, and maintaining software applications and libraries across various platforms and domains, including education, cultural heritage, and enterprise reconstruction. It encompasses topics such as modular design, code reviews, software libraries, interface design, multimodal web interfaces, and intelligent teaching systems.</td>\n",
              "      <td>Software Development</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-dbd2911d-03c6-4293-b180-b58b1f4a7fa9')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-dbd2911d-03c6-4293-b180-b58b1f4a7fa9 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-dbd2911d-03c6-4293-b180-b58b1f4a7fa9');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-6c6ecbae-bf64-4ac4-b375-3f8a23c1411d\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-6c6ecbae-bf64-4ac4-b375-3f8a23c1411d')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-6c6ecbae-bf64-4ac4-b375-3f8a23c1411d button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 117
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Rename the repeated labels\n",
        "final_df.loc[final_df['cluster'] == 53, 'label'] = 'Data Protection'\n",
        "final_df.loc[final_df['cluster'] == 16, 'label'] = 'Computational Graph Theory'\n",
        "final_df.loc[final_df['cluster'] == 10, 'label'] = 'Efficient Information Retrieval'\n",
        "final_df.loc[final_df['cluster'] == 11, 'label'] = 'Set Theory'\n",
        "final_df.loc[final_df['cluster'] == 24, 'label'] = 'Equivalence Relations'\n",
        "final_df.loc[final_df['cluster'] == 50, 'label'] = 'Software Development Principles'"
      ],
      "metadata": {
        "id": "lREutYTptsEQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check for success\n",
        "len(set(final_df.cluster.tolist())) , len(set(final_df.label.tolist()))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "He_iTvVQvz64",
        "outputId": "2ee69f9b-06f5-4614-edd5-885ec4cf3b23"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(59, 59)"
            ]
          },
          "metadata": {},
          "execution_count": 119
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Save the data to a file\n",
        "final_df.to_csv(data_path+descriptions_labels_keys_file, index=False)"
      ],
      "metadata": {
        "id": "PjGuJJD8v0-t"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## The Knowledge Graph"
      ],
      "metadata": {
        "id": "Fd-2cHHWgjm8"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Install Neo4j Dependencies"
      ],
      "metadata": {
        "id": "RAqsNDdPDEYa"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install neo4j"
      ],
      "metadata": {
        "id": "KCzMoiTPDInV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7015f225-2219-4446-d111-4e120d221aea"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting neo4j\n",
            "  Downloading neo4j-5.16.0.tar.gz (197 kB)\n",
            "\u001b[?25l     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/197.8 kB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K     \u001b[91m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[90m╺\u001b[0m\u001b[90m━━━━\u001b[0m \u001b[32m174.1/197.8 kB\u001b[0m \u001b[31m5.2 MB/s\u001b[0m eta \u001b[36m0:00:01\u001b[0m\r\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m197.8/197.8 kB\u001b[0m \u001b[31m4.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25h  Installing build dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Getting requirements to build wheel ... \u001b[?25l\u001b[?25hdone\n",
            "  Installing backend dependencies ... \u001b[?25l\u001b[?25hdone\n",
            "  Preparing metadata (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: pytz in /usr/local/lib/python3.10/dist-packages (from neo4j) (2023.3.post1)\n",
            "Building wheels for collected packages: neo4j\n",
            "  Building wheel for neo4j (pyproject.toml) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for neo4j: filename=neo4j-5.16.0-py3-none-any.whl size=273811 sha256=c2531cda333ac9d0875fdb60b488d0679bc503139525d5af66e6e487e74ddb59\n",
            "  Stored in directory: /root/.cache/pip/wheels/20/a0/f6/87a1ec9636c915fe2d6c6e859fd55a6231dd9bc95a1d5394b1\n",
            "Successfully built neo4j\n",
            "Installing collected packages: neo4j\n",
            "Successfully installed neo4j-5.16.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Load and Structure Data"
      ],
      "metadata": {
        "id": "7q0PLDr-WN2P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Articles, titles, abstracts and keywords\n",
        "articles = pd.read_csv(data_path+parsed_keys_file)\n",
        "# Keywords, soft labels/clusters, descriptions and labels\n",
        "keywords = pd.read_csv(data_path+descriptions_labels_keys_file)"
      ],
      "metadata": {
        "id": "-ALA4fr_hzig"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import ast\n",
        "# The keywords column entries must be a list\n",
        "articles['titles_keys'] = articles['titles_keys'].apply(ast.literal_eval)"
      ],
      "metadata": {
        "id": "chP2vBaJu1Xy"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the articles dataframe\n",
        "articles.head(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 216
        },
        "id": "nTmQMsAprfVJ",
        "outputId": "18eb5fee-faa0-469c-c09f-c0876eed0032"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "     id                                                              title  \\\n",
              "0  1001  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "1  1002                    Minimum Description Length and Compositionality   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               abstract  \\\n",
              "0                                                                                                                                                               The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n   \n",
              "1    We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n   \n",
              "\n",
              "                                                                                                                          titles_keys  \n",
              "0                                                    [von neumann, quantum logic, classical von neumann architecture, quantum, logic]  \n",
              "1  [minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-1f4b4062-fc15-4227-961f-42daef53a941\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>id</th>\n",
              "      <th>title</th>\n",
              "      <th>abstract</th>\n",
              "      <th>titles_keys</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1001</td>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>The name of John von Neumann is common both in quantum mechanics and computer\\nscience. Are they really two absolutely unconnected areas? Many works devoted\\nto quantum computations and communications are serious argument to suggest\\nabout existence of such a relation, but it is impossible to touch the new and\\nactive theme in a short review. In the paper are described the structures and\\nmodels of linear algebra and just due to their generality it is possible to use\\nuniversal description of very different areas as quantum mechanics and theory\\nof Bayesian image analysis, associative memory, neural networks, fuzzy logic.\\n</td>\n",
              "      <td>[von neumann, quantum logic, classical von neumann architecture, quantum, logic]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1002</td>\n",
              "      <td>Minimum Description Length and Compositionality</td>\n",
              "      <td>We present a non-vacuous definition of compositionality. It is based on the\\nidea of combining the minimum description length principle with the original\\ndefinition of compositionality (that is, that the meaning of the whole is a\\nfunction of the meaning of the parts).\\n  The new definition is intuitive and allows us to distinguish between\\ncompositional and non-compositional semantics, and between idiomatic and\\nnon-idiomatic expressions. It is not ad hoc, since it does not make any\\nreferences to non-intrinsic properties of meaning functions (like being a\\npolynomial). Moreover, it allows us to compare different meaning functions with\\nrespect to how compositional they are. It bridges linguistic and corpus-based,\\nstatistical approaches to natural language understanding.\\n</td>\n",
              "      <td>[minimum description length, compositionality, description length, compositionality, MDL (acronym for Minimum Description Length)]</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-1f4b4062-fc15-4227-961f-42daef53a941')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-1f4b4062-fc15-4227-961f-42daef53a941 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-1f4b4062-fc15-4227-961f-42daef53a941');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-1cfcf307-4f02-4642-9f39-dfbe485982a3\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-1cfcf307-4f02-4642-9f39-dfbe485982a3')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-1cfcf307-4f02-4642-9f39-dfbe485982a3 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 124
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Check the keywords dataframe\n",
        "keywords.head(2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 182
        },
        "id": "dgB8Z-x2re8o",
        "outputId": "4e01d476-8889-4bb5-b623-4e78b9aad5f0"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                        key  cluster  \\\n",
              "0  standard data processing        0   \n",
              "1                      o(n)        0   \n",
              "\n",
              "                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         description  \\\n",
              "0  Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and   \n",
              "1  Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and   \n",
              "\n",
              "                    label  \n",
              "0  Stream Data Processing  \n",
              "1  Stream Data Processing  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c8234bba-1141-49a6-a166-fc1e296a48a5\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>key</th>\n",
              "      <th>cluster</th>\n",
              "      <th>description</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>standard data processing</td>\n",
              "      <td>0</td>\n",
              "      <td>Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and</td>\n",
              "      <td>Stream Data Processing</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>o(n)</td>\n",
              "      <td>0</td>\n",
              "      <td>Focused on developing efficient algorithms and structures for processing large volumes of streaming data, particularly for multidimensional and categorical data, with an emphasis on free and dense-time models, vector clocks, and matrix operations. Emphasizes stream data acquisition, two-stage processing, parity-check matrices, and one-pass algorithms. Also explores non-data-aided and changed data processing, as well as power and multiplicity concepts for structured and individual data streams. Free and categorical structures are explored, including universal arrays, free categories, and</td>\n",
              "      <td>Stream Data Processing</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c8234bba-1141-49a6-a166-fc1e296a48a5')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-c8234bba-1141-49a6-a166-fc1e296a48a5 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-c8234bba-1141-49a6-a166-fc1e296a48a5');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-7095c29a-ff61-4e21-9a14-139c009600c6\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-7095c29a-ff61-4e21-9a14-139c009600c6')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-7095c29a-ff61-4e21-9a14-139c009600c6 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 125
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Establish Neo4j Connection"
      ],
      "metadata": {
        "id": "-W50wb0eFNj9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Import the Neo4j connector module\n",
        "from utils.neo4j_conn import *"
      ],
      "metadata": {
        "id": "LyOz86CDBaCj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create an Neo4j AuraDB free instance and collect the credentials\n",
        "URI = 'neo4j+ssc://xxxxxxxx.databases.neo4j.io'\n",
        "USER = 'neo4j'\n",
        "PWD = 'your_password'\n",
        "\n",
        "# Initialize the Neo4j connector\n",
        "graph=Neo4jGraph(url=URI, username=USER, password=PWD)"
      ],
      "metadata": {
        "id": "kLFvWZ_jp66p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Test the connection for success\n",
        "graph.query(\"MATCH (n) RETURN count(n)\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tmPO_nI3xA0I",
        "outputId": "3b697c87-6d89-4617-9011-e7cb755fa0eb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'count(n)': 0}]"
            ]
          },
          "metadata": {},
          "execution_count": 130
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "%md\n",
        "### Build the Knowledge Graph"
      ],
      "metadata": {
        "id": "LpiXF0ehB4ym"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create uniqueness constraints on each node\n",
        "constraints_articles = \"\"\"\n",
        "CREATE CONSTRAINT article_id IF NOT EXISTS FOR (article:Article) REQUIRE article.id IS UNIQUE;\"\"\"\n",
        "graph.query(constraints_articles)\n",
        "constraints_keywords=\"\"\"\n",
        "CREATE CONSTRAINT keyword_key IF NOT EXISTS FOR (keyword:Keyword) REQUIRE keyword.key IS UNIQUE;\n",
        "\"\"\"\n",
        "graph.query(constraints_keywords)\n",
        "constraints_clusters=\"\"\"\n",
        "CREATE CONSTRAINT topic_cluster IF NOT EXISTS FOR (topic:Topic) REQUIRE topic.cluster IS UNIQUE;\n",
        "\"\"\"\n",
        "graph.query(constraints_clusters)"
      ],
      "metadata": {
        "id": "TUhsvZlqjqMv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e6fc292c-60cf-4b9b-f8ec-f4ab457aa113"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {},
          "execution_count": 131
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load Keyword and Topic nodes, and the relationship HAS_TOPIC\n",
        "query_keywords_topics = \"\"\"\n",
        "    UNWIND $rows AS row\n",
        "    MERGE (k:Keyword {name: row.key})\n",
        "    MERGE (t:Topic {cluster: row.cluster, description: row.description, label: row.label})\n",
        "    MERGE (k)-[:HAS_TOPIC]->(t)\n",
        "    \"\"\"\n",
        "graph.load_data(query_keywords_topics, keywords)"
      ],
      "metadata": {
        "id": "6OUgnm_eqEzf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load Article nodes and the relationship HAS_KEY\n",
        "query_articles = \"\"\"\n",
        "    UNWIND $rows as row\n",
        "    MERGE (a:Article {id: row.id, title: row.title, abstract: row.abstract})\n",
        "    WITH a, row\n",
        "    UNWIND row.titles_keys as key\n",
        "    MATCH (k:Keyword {name: key})\n",
        "    MERGE (a)-[:HAS_KEY]->(k)\n",
        "    \"\"\"\n",
        "graph.load_data(query_articles, articles)"
      ],
      "metadata": {
        "id": "2kYdTFaos7on"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Query the Knowledge Graph"
      ],
      "metadata": {
        "id": "sYevn9wQkwUv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Count the nodes in the graph, grouped by node labels\n",
        "\n",
        "cypher_statement = \"\"\"\n",
        "MATCH (n)\n",
        "RETURN HEAD(LABELS(n)) AS Node, COUNT(n) AS Total\n",
        "ORDER BY Total DESC\n",
        "\"\"\"\n",
        "\n",
        "result = graph.query(cypher_statement)\n",
        "df = pd.DataFrame(result)\n",
        "df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 143
        },
        "id": "uoAcvApkpkl2",
        "outputId": "260c02d8-f85a-4ec7-cd4a-78a4ac953a41"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "      Node  Total\n",
              "0  Keyword   2951\n",
              "1  Article    884\n",
              "2    Topic     59"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-713005d6-8521-45f6-a7c0-9f0ff745e1d7\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Node</th>\n",
              "      <th>Total</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Keyword</td>\n",
              "      <td>2951</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Article</td>\n",
              "      <td>884</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Topic</td>\n",
              "      <td>59</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-713005d6-8521-45f6-a7c0-9f0ff745e1d7')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-713005d6-8521-45f6-a7c0-9f0ff745e1d7 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-713005d6-8521-45f6-a7c0-9f0ff745e1d7');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-34745ce9-5410-465d-b0ee-0d65eaf5651f\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-34745ce9-5410-465d-b0ee-0d65eaf5651f')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-34745ce9-5410-465d-b0ee-0d65eaf5651f button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "  <div id=\"id_94f11554-fe73-41f5-baa7-7ff8e8bfb2b0\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('df')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_94f11554-fe73-41f5-baa7-7ff8e8bfb2b0 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('df');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 134
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Count the relationships in the graph, grouped by type\n",
        "\n",
        "cypher_statement = \"\"\"\n",
        "MATCH ()-[r]-()\n",
        "RETURN TYPE(r) AS Relationship, COUNT(r) AS Total\n",
        "ORDER BY Total DESC\n",
        "\"\"\"\n",
        "\n",
        "result = graph.query(cypher_statement)\n",
        "df = pd.DataFrame(result)\n",
        "df"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 125
        },
        "id": "nk2PQ9BWp93u",
        "outputId": "840211f9-5f1f-4ada-dd8f-0f4af4363445"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  Relationship  Total\n",
              "0      HAS_KEY   8712\n",
              "1    HAS_TOPIC   5902"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-6eb9e9ec-4b31-4e7c-9c49-d3a66940830f\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Relationship</th>\n",
              "      <th>Total</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>HAS_KEY</td>\n",
              "      <td>8712</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>HAS_TOPIC</td>\n",
              "      <td>5902</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-6eb9e9ec-4b31-4e7c-9c49-d3a66940830f')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-6eb9e9ec-4b31-4e7c-9c49-d3a66940830f button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-6eb9e9ec-4b31-4e7c-9c49-d3a66940830f');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-219d367c-1223-4ce1-951e-3287c66ebc91\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-219d367c-1223-4ce1-951e-3287c66ebc91')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-219d367c-1223-4ce1-951e-3287c66ebc91 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "  <div id=\"id_81f195c1-77dc-484f-a6d2-85c228725f71\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('df')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_81f195c1-77dc-484f-a6d2-85c228725f71 button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('df');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 135
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find the counts of articles associated with each topic through the keywords\n",
        "\n",
        "cypher_statement = \"\"\"\n",
        "MATCH (a:Article)-[:HAS_KEY]->(k:Keyword)-[:HAS_TOPIC]->(t:Topic)\n",
        "RETURN t.label AS Label, t.cluster AS Cluster, COUNT(a) AS Articles\n",
        "ORDER BY Articles DESC\n",
        "\"\"\"\n",
        "\n",
        "result = graph.query(cypher_statement)\n",
        "df = pd.DataFrame(result)\n",
        "df.head()"
      ],
      "metadata": {
        "id": "2ey8y1M2i3QR",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "055a783c-36ec-4570-eafc-8ad630584496"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                   Label  Cluster  Articles\n",
              "0                              Semantics       58       136\n",
              "1                           Graph Theory       17       124\n",
              "2                   Software Engineering       46       124\n",
              "3  Game Theory and Strategic Interaction       55       116\n",
              "4        Software Development Principles       50       112"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-debd62c0-3360-4c02-bf86-b96ee531a013\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Label</th>\n",
              "      <th>Cluster</th>\n",
              "      <th>Articles</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Semantics</td>\n",
              "      <td>58</td>\n",
              "      <td>136</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Graph Theory</td>\n",
              "      <td>17</td>\n",
              "      <td>124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Software Engineering</td>\n",
              "      <td>46</td>\n",
              "      <td>124</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Game Theory and Strategic Interaction</td>\n",
              "      <td>55</td>\n",
              "      <td>116</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Software Development Principles</td>\n",
              "      <td>50</td>\n",
              "      <td>112</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-debd62c0-3360-4c02-bf86-b96ee531a013')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-debd62c0-3360-4c02-bf86-b96ee531a013 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-debd62c0-3360-4c02-bf86-b96ee531a013');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-c94683e0-0731-4454-9219-353a8628208c\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-c94683e0-0731-4454-9219-353a8628208c')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-c94683e0-0731-4454-9219-353a8628208c button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 136
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# What are the five most common keywords shared by articles within the same topic?\n",
        "\n",
        "cypher_statement = \"\"\"\n",
        "MATCH (t:Topic)<-[:HAS_TOPIC]-(:Keyword)<-[:HAS_KEY]-(a:Article)\n",
        "WITH t, a\n",
        "MATCH (a)-[:HAS_KEY]->(k:Keyword)\n",
        "RETURN t.label AS Label, k.name AS Keyword, COUNT(*) AS SharedCount\n",
        "ORDER BY SharedCount DESC\n",
        "LIMIT 5\n",
        "\"\"\"\n",
        "\n",
        "result = graph.query(cypher_statement)\n",
        "df = pd.DataFrame(result)\n",
        "df.head()"
      ],
      "metadata": {
        "id": "h1YRdRyei3NI",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "4c431395-6e22-4584-d180-a34db75dd77d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                      Label       Keyword  SharedCount\n",
              "0              Optimization  optimization           26\n",
              "1                 Semantics     reasoning           26\n",
              "2     Logic and Computation     reasoning           26\n",
              "3  Computational Complexity    complexity           22\n",
              "4     Logic and Computation     deduction           21"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-89f4e136-08f2-4442-a000-027133a78ed4\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Label</th>\n",
              "      <th>Keyword</th>\n",
              "      <th>SharedCount</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Optimization</td>\n",
              "      <td>optimization</td>\n",
              "      <td>26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Semantics</td>\n",
              "      <td>reasoning</td>\n",
              "      <td>26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Logic and Computation</td>\n",
              "      <td>reasoning</td>\n",
              "      <td>26</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Computational Complexity</td>\n",
              "      <td>complexity</td>\n",
              "      <td>22</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Logic and Computation</td>\n",
              "      <td>deduction</td>\n",
              "      <td>21</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-89f4e136-08f2-4442-a000-027133a78ed4')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-89f4e136-08f2-4442-a000-027133a78ed4 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-89f4e136-08f2-4442-a000-027133a78ed4');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-17243d5a-2d62-4f8c-bb78-f18970d863bf\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-17243d5a-2d62-4f8c-bb78-f18970d863bf')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-17243d5a-2d62-4f8c-bb78-f18970d863bf button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 137
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Find pairs of articles that are indirectly connected through a chain of length up to 4\n",
        "\n",
        "cypher_statement = \"\"\"\n",
        "MATCH path = (a1:Article)-[:HAS_KEY|HAS_TOPIC*2..4]-(a2:Article)\n",
        "WHERE a1 <> a2\n",
        "RETURN a1.title AS Title1, a2.title AS Title2, LENGTH(path) AS PathLength\n",
        "ORDER BY PathLength DESC\n",
        "LIMIT 5\n",
        "\"\"\"\n",
        "\n",
        "result = graph.query(cypher_statement)\n",
        "df = pd.DataFrame(result)\n",
        "df"
      ],
      "metadata": {
        "id": "_RVzCWJhi3Ek",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "outputId": "639956e8-4d9d-4961-cc77-be1432a4b1c9"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                              Title1  \\\n",
              "0  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "1  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "2  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "3  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "4  Von Neumann Quantum Logic vs. Classical von Neumann Architecture?   \n",
              "\n",
              "                                                                              Title2  \\\n",
              "0  Multiparty computation unconditionally secure against Q^2 adversary\\n  structures   \n",
              "1   Tools and Techniques for Managing Clusters for SciDAC Lattice QCD at\\n  Fermilab   \n",
              "2                                             State complexes for metamorphic robots   \n",
              "3                                             The pre-history of quantum computation   \n",
              "4                    Quantum Computing and Phase Transitions in Combinatorial Search   \n",
              "\n",
              "   PathLength  \n",
              "0           4  \n",
              "1           4  \n",
              "2           4  \n",
              "3           4  \n",
              "4           4  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-f509ee74-cb86-496e-986f-39e1bc168670\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Title1</th>\n",
              "      <th>Title2</th>\n",
              "      <th>PathLength</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>Multiparty computation unconditionally secure against Q^2 adversary\\n  structures</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>Tools and Techniques for Managing Clusters for SciDAC Lattice QCD at\\n  Fermilab</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>State complexes for metamorphic robots</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>The pre-history of quantum computation</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Von Neumann Quantum Logic vs. Classical von Neumann Architecture?</td>\n",
              "      <td>Quantum Computing and Phase Transitions in Combinatorial Search</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-f509ee74-cb86-496e-986f-39e1bc168670')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-f509ee74-cb86-496e-986f-39e1bc168670 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-f509ee74-cb86-496e-986f-39e1bc168670');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-eac72384-c145-476a-a55e-f3bd85042cdb\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-eac72384-c145-476a-a55e-f3bd85042cdb')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-eac72384-c145-476a-a55e-f3bd85042cdb button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "\n",
              "  <div id=\"id_d59fa524-b8ba-43d3-8da7-84f3e711f13c\">\n",
              "    <style>\n",
              "      .colab-df-generate {\n",
              "        background-color: #E8F0FE;\n",
              "        border: none;\n",
              "        border-radius: 50%;\n",
              "        cursor: pointer;\n",
              "        display: none;\n",
              "        fill: #1967D2;\n",
              "        height: 32px;\n",
              "        padding: 0 0 0 0;\n",
              "        width: 32px;\n",
              "      }\n",
              "\n",
              "      .colab-df-generate:hover {\n",
              "        background-color: #E2EBFA;\n",
              "        box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "        fill: #174EA6;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate {\n",
              "        background-color: #3B4455;\n",
              "        fill: #D2E3FC;\n",
              "      }\n",
              "\n",
              "      [theme=dark] .colab-df-generate:hover {\n",
              "        background-color: #434B5C;\n",
              "        box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "        filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "        fill: #FFFFFF;\n",
              "      }\n",
              "    </style>\n",
              "    <button class=\"colab-df-generate\" onclick=\"generateWithVariable('df')\"\n",
              "            title=\"Generate code using this dataframe.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M7,19H8.4L18.45,9,17,7.55,7,17.6ZM5,21V16.75L18.45,3.32a2,2,0,0,1,2.83,0l1.4,1.43a1.91,1.91,0,0,1,.58,1.4,1.91,1.91,0,0,1-.58,1.4L9.25,21ZM18.45,9,17,7.55Zm-12,3A5.31,5.31,0,0,0,4.9,8.1,5.31,5.31,0,0,0,1,6.5,5.31,5.31,0,0,0,4.9,4.9,5.31,5.31,0,0,0,6.5,1,5.31,5.31,0,0,0,8.1,4.9,5.31,5.31,0,0,0,12,6.5,5.46,5.46,0,0,0,6.5,12Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "    <script>\n",
              "      (() => {\n",
              "      const buttonEl =\n",
              "        document.querySelector('#id_d59fa524-b8ba-43d3-8da7-84f3e711f13c button.colab-df-generate');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      buttonEl.onclick = () => {\n",
              "        google.colab.notebook.generateWithVariable('df');\n",
              "      }\n",
              "      })();\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 138
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "gYoUSqNX0BBE"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}