backdoor_detector.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "FgHJspO5Af_z"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import seaborn as sns\n",
        "import keras\n",
        "import sys\n",
        "import h5py\n",
        "import warnings\n",
        "from tqdm import tqdm"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## OG BadNet"
      ],
      "metadata": {
        "id": "7Xi9Z9TTArUu"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CYWhNc3VA6AS",
        "outputId": "b58df9ca-39fd-4ddf-c362-2da94ccdfcbe"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# clean_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/cl/clean_validation_data.h5'\n",
        "# bad_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/bd/sunglasses_poisoned_data.h5'\n",
        "# model_name = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/sunglasses_bd_net.h5'\n",
        "\n",
        "valid_clean_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/cl/valid.h5'\n",
        "test_clean_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/cl/test.h5'\n",
        "valid_bad_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/bd/bd_valid.h5'\n",
        "test_bad_data = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/data/bd/bd_test.h5'\n",
        "model_name = '/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/bd_net.h5'"
      ],
      "metadata": {
        "id": "D6nN9bmBAo6Q"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "wAEflpCD7lR0"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Sz4x9gpFJFAq"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Test the given backdoored model -- code from eval.py"
      ],
      "metadata": {
        "id": "e4ANNCc9JFYO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def load_data(path):\n",
        "    data = h5py.File(path, 'r')\n",
        "    x_data = np.array(data['data'])\n",
        "    y_data = np.array(data['label'])\n",
        "    x_data = x_data.transpose((0,2,3,1))\n",
        "\n",
        "    return x_data, y_data"
      ],
      "metadata": {
        "id": "cyl--_iPHU_K"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Check input images -- 5 sample images"
      ],
      "metadata": {
        "id": "cwsSGG847zx1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_data, y_data = load_data(valid_clean_data)\n",
        "\n",
        "plt.figure(figsize=(15, 3))  # Adjust the figure size as needed\n",
        "\n",
        "for i in range(5):\n",
        "    plt.subplot(1, 5, i + 1)\n",
        "    plt.imshow(x_data[np.random.randint(x_data.shape[0], size=1)][0]/255)\n",
        "    plt.title(f'Label: {y_data[i]}')\n",
        "    plt.axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        },
        "id": "VuU26pkt7yvn",
        "outputId": "312af162-165e-4dc0-a278-169f80ff3b95"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x300 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### check poisoned images -- 5 sample images"
      ],
      "metadata": {
        "id": "TuYlis-DAx3U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_data, y_data = load_data(valid_bad_data)\n",
        "\n",
        "plt.figure(figsize=(15, 3))  # Adjust the figure size as needed\n",
        "\n",
        "for i in range(5):\n",
        "    plt.subplot(1, 5, i + 1)\n",
        "    plt.imshow(x_data[np.random.randint(x_data.shape[0], size=1)][0]/255)\n",
        "    plt.title(f'Label: {y_data[i]}')\n",
        "    plt.axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 290
        },
        "id": "Xc96vkt0AxSa",
        "outputId": "b6c484db-e58a-474e-ff22-ed3f147dc8bc"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1500x300 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Verify performance of the given Badnet ~ Clean Classification accuracy: 98.64%"
      ],
      "metadata": {
        "id": "eCdAsPqgCd2Y"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def main():\n",
        "  cl_x_test, cl_y_test = load_data(valid_clean_data)\n",
        "  bd_x_test, bd_y_test = load_data(valid_bad_data)\n",
        "\n",
        "  bd_model = keras.models.load_model(model_name)\n",
        "\n",
        "  cl_label_p = np.argmax(bd_model.predict(cl_x_test), axis=1)\n",
        "  clean_accuracy = np.mean(np.equal(cl_label_p, cl_y_test))*100\n",
        "  print('Clean Classification accuracy:', clean_accuracy)\n",
        "\n",
        "  bd_label_p = np.argmax(bd_model.predict(bd_x_test), axis=1)\n",
        "  asr = np.mean(np.equal(bd_label_p, bd_y_test))*100\n",
        "  print('Attack Success Rate:', asr)"
      ],
      "metadata": {
        "id": "HJ0YZA7HHtwr"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "main()"
      ],
      "metadata": {
        "id": "FK_wOGNBIMUi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e55bd49c-9adc-461c-809d-55e3f2b08486"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 8s 3ms/step\n",
            "Clean Classification accuracy: 98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Attack Success Rate: 100.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Verified!!"
      ],
      "metadata": {
        "id": "ZNe-AW62DMmb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Start Pruning defense"
      ],
      "metadata": {
        "id": "cCbtdkySJYsG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### check model architecture to see which layer to prune"
      ],
      "metadata": {
        "id": "CZgUkXFnJe45"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model = keras.models.load_model(model_name)"
      ],
      "metadata": {
        "id": "muf4oJVtJiPG"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(model.summary())"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Aj1TgTO2Npih",
        "outputId": "acd3e2b1-1c17-4faf-f77f-6696efece880"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                Output Shape                 Param #   Connected to                  \n",
            "==================================================================================================\n",
            " input (InputLayer)          [(None, 55, 47, 3)]          0         []                            \n",
            "                                                                                                  \n",
            " conv_1 (Conv2D)             (None, 52, 44, 20)           980       ['input[0][0]']               \n",
            "                                                                                                  \n",
            " pool_1 (MaxPooling2D)       (None, 26, 22, 20)           0         ['conv_1[0][0]']              \n",
            "                                                                                                  \n",
            " conv_2 (Conv2D)             (None, 24, 20, 40)           7240      ['pool_1[0][0]']              \n",
            "                                                                                                  \n",
            " pool_2 (MaxPooling2D)       (None, 12, 10, 40)           0         ['conv_2[0][0]']              \n",
            "                                                                                                  \n",
            " conv_3 (Conv2D)             (None, 10, 8, 60)            21660     ['pool_2[0][0]']              \n",
            "                                                                                                  \n",
            " pool_3 (MaxPooling2D)       (None, 5, 4, 60)             0         ['conv_3[0][0]']              \n",
            "                                                                                                  \n",
            " conv_4 (Conv2D)             (None, 4, 3, 80)             19280     ['pool_3[0][0]']              \n",
            "                                                                                                  \n",
            " flatten_1 (Flatten)         (None, 1200)                 0         ['pool_3[0][0]']              \n",
            "                                                                                                  \n",
            " flatten_2 (Flatten)         (None, 960)                  0         ['conv_4[0][0]']              \n",
            "                                                                                                  \n",
            " fc_1 (Dense)                (None, 160)                  192160    ['flatten_1[0][0]']           \n",
            "                                                                                                  \n",
            " fc_2 (Dense)                (None, 160)                  153760    ['flatten_2[0][0]']           \n",
            "                                                                                                  \n",
            " add_1 (Add)                 (None, 160)                  0         ['fc_1[0][0]',                \n",
            "                                                                     'fc_2[0][0]']                \n",
            "                                                                                                  \n",
            " activation_1 (Activation)   (None, 160)                  0         ['add_1[0][0]']               \n",
            "                                                                                                  \n",
            " output (Dense)              (None, 1283)                 206563    ['activation_1[0][0]']        \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 601643 (2.30 MB)\n",
            "Trainable params: 601643 (2.30 MB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "__________________________________________________________________________________________________\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "clean_x, clean_y = load_data(valid_clean_data)\n",
        "bad_x, bad_y = load_data(valid_bad_data)"
      ],
      "metadata": {
        "id": "NBf-4H5jVQaL"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model_copy = keras.models.clone_model(model)\n",
        "model_copy.set_weights(model.get_weights())\n",
        "\n",
        "# Get the output of the 'pool_3' layer from the model\n",
        "last_pool_output = model_copy.get_layer('pool_3').output\n",
        "\n",
        "# model creation with i/p as the given model and produces output of the pooling layer\n",
        "temp_model = keras.models.Model(inputs=model_copy.input,outputs=last_pool_output)\n",
        "\n",
        "# predict output of temp_model on clean_x data\n",
        "temp_pred = temp_model.predict(clean_x)\n",
        "\n",
        "# activations calculated along all dimensions\n",
        "avg_act = np.mean(temp_pred, axis=(0,1,2))\n",
        "\n",
        "# sort the above indices in descending order\n",
        "sorted_act_ind = np.argsort(avg_act)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "62t9HSLcZZXD",
        "outputId": "bcbc0198-44d0-4f8f-caa1-a043b441263d"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 1s 2ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sorted_act_ind"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Wi3WKRJUV3l5",
        "outputId": "f28f38d8-fbb0-4235-f543-21cd1c523aec"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0, 26, 27, 30, 31, 33, 34, 36, 37, 38, 25, 39, 41, 44, 45, 47, 48,\n",
              "       49, 50, 53, 55, 40, 24, 59,  9,  2, 12, 13, 17, 14, 15, 23,  6, 51,\n",
              "       32, 22, 21, 20, 19, 43, 58,  3, 42,  1, 29, 16, 56, 46,  5,  8, 11,\n",
              "       54, 10, 28, 35, 18,  4,  7, 52, 57])"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### we get 60 channels in sorted_act_ind sorted in descending order of their avg activations"
      ],
      "metadata": {
        "id": "JvxKcnShgxMV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "temp_pred.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UcIBfNYVV-bb",
        "outputId": "8d023fc0-cf88-4548-dfa7-9bd97aa51fc7"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(11547, 5, 4, 60)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# get original model's clean accuracy to compare\n",
        "cl_label_p = np.argmax(model.predict(clean_x), axis=1)\n",
        "og_clean_accuracy = np.mean(np.equal(cl_label_p, clean_y))*100\n",
        "print('Clean Classification accuracy:', og_clean_accuracy)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q3BOSkgJhDAO",
        "outputId": "fb1eb245-c6e0-4fbc-9834-cbf12b2fddec"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 1s 4ms/step\n",
            "Clean Classification accuracy: 98.64899974019225\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_clean_acc = []\n",
        "attack_success_rate = []\n",
        "channel_index = 0"
      ],
      "metadata": {
        "id": "yQC89CamaO_7"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# get initial weights\n",
        "weights = model_copy.layers[5].get_weights()"
      ],
      "metadata": {
        "id": "8ggJmCkzaTPv"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Prune until accuracy dropped >= 2%"
      ],
      "metadata": {
        "id": "XY77VSKsm6ce"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "while channel_index < len(sorted_act_ind):\n",
        "\n",
        "  weights[0][:,:,:,sorted_act_ind[channel_index]] = 0\n",
        "  weights[1][sorted_act_ind[channel_index]] = 0\n",
        "\n",
        "  model_copy.layers[5].set_weights(weights)\n",
        "\n",
        "  cl_label_p = np.argmax(model_copy.predict(clean_x), axis=1)\n",
        "  clean_accuracy = np.mean(np.equal(cl_label_p, clean_y))*100\n",
        "\n",
        "  print(\"Accuracy after dropping channel index: \", sorted_act_ind[channel_index], \" is: \", clean_accuracy)\n",
        "\n",
        "  pruned_clean_acc.append(clean_accuracy)\n",
        "\n",
        "\n",
        "\n",
        "  if (og_clean_accuracy-clean_accuracy >= 2):\n",
        "    print(\"The accuracy is dropped at least 2%, the model is saved!\")\n",
        "    model_copy.save('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_2.h5')\n",
        "\n",
        "    bd_label_p = np.argmax(model_copy.predict(bad_x), axis=1)\n",
        "    asr = np.mean(np.equal(bd_label_p, bad_y))*100\n",
        "    attack_success_rate.append(asr)\n",
        "\n",
        "    print()\n",
        "    print(\"The clean accuracy is: \",clean_accuracy)\n",
        "    print(\"The attack success rate is: \",asr)\n",
        "    print(\"The pruned channel index is: \",sorted_act_ind[channel_index])\n",
        "    break\n",
        "  channel_index += 1\n",
        "\n",
        "  # keras.backend.clear_session()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8NrMLfwEaUfS",
        "outputId": "27ffbad0-bd91-4d41-e50f-6919ac6a39d2"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  0  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  26  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  27  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  30  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  31  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  33  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  34  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  36  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  37  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  38  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  25  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  39  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  41  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  44  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  45  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  47  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  48  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  49  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  50  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  53  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  55  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  40  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  24  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  59  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  9  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  2  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  12  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  13  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  17  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  14  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  15  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  23  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  6  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  51  is:  98.64033948211657\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  32  is:  98.64033948211657\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  22  is:  98.63167922404088\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  21  is:  98.65765999826795\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  20  is:  98.64899974019225\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  19  is:  98.6056984498138\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  43  is:  98.57105741751104\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  58  is:  98.53641638520828\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  3  is:  98.19000606218066\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  42  is:  97.65307006148784\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  1  is:  97.50584567420108\n",
            "361/361 [==============================] - 1s 3ms/step\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3079: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n",
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy after dropping channel index:  29  is:  95.75647354291158\n",
            "The accuracy is dropped at least 2%, the model is saved!\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "\n",
            "The clean accuracy is:  95.75647354291158\n",
            "The attack success rate is:  100.0\n",
            "The pruned channel index is:  29\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Prune until accuracy dropped >= 4%"
      ],
      "metadata": {
        "id": "HCxVu_sSnNhU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "while channel_index < len(sorted_act_ind):\n",
        "\n",
        "  weights[0][:,:,:,sorted_act_ind[channel_index]] = 0\n",
        "  weights[1][sorted_act_ind[channel_index]] = 0\n",
        "\n",
        "  model_copy.layers[5].set_weights(weights)\n",
        "\n",
        "  cl_label_p = np.argmax(model_copy.predict(clean_x), axis=1)\n",
        "  clean_accuracy = np.mean(np.equal(cl_label_p, clean_y))*100\n",
        "  print(\"Accuracy after dropping channel index: \", sorted_act_ind[channel_index], \" is: \", clean_accuracy)\n",
        "\n",
        "  pruned_clean_acc.append(clean_accuracy)\n",
        "\n",
        "  if (og_clean_accuracy-clean_accuracy >= 4):\n",
        "    print(\"The accuracy is dropped at least 4%, the model is saved!\")\n",
        "    model_copy.save('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_4.h5')\n",
        "\n",
        "    # pruned_clean_acc.append(clean_accuracy)\n",
        "\n",
        "    bd_label_p = np.argmax(model_copy.predict(bad_x), axis=1)\n",
        "    asr = np.mean(np.equal(bd_label_p, bad_y))*100\n",
        "    attack_success_rate.append(asr)\n",
        "    print()\n",
        "    print(\"The clean accuracy is: \",clean_accuracy)\n",
        "    print(\"The attack success rate is: \",asr)\n",
        "    print(\"The pruned channel index is: \",sorted_act_ind[channel_index])\n",
        "    break\n",
        "  channel_index += 1\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qQSuOLpdmtIz",
        "outputId": "f1aa2522-0aa9-4662-9d57-a33db10bf5a3"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 2s 4ms/step\n",
            "Accuracy after dropping channel index:  29  is:  95.75647354291158\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  16  is:  95.20221702606739\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  56  is:  94.7172425738287\n",
            "361/361 [==============================] - 1s 3ms/step\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy after dropping channel index:  46  is:  92.09318437689443\n",
            "The accuracy is dropped at least 4%, the model is saved!\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "\n",
            "The clean accuracy is:  92.09318437689443\n",
            "The attack success rate is:  99.9913397419243\n",
            "The pruned channel index is:  46\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Prune until accuracy dropped >= 10%"
      ],
      "metadata": {
        "id": "PPsEOBmVoGsN"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "while channel_index < len(sorted_act_ind):\n",
        "\n",
        "  weights[0][:,:,:,sorted_act_ind[channel_index]] = 0\n",
        "  weights[1][sorted_act_ind[channel_index]] = 0\n",
        "\n",
        "  model_copy.layers[5].set_weights(weights)\n",
        "\n",
        "  cl_label_p = np.argmax(model_copy.predict(clean_x), axis=1)\n",
        "  clean_accuracy = np.mean(np.equal(cl_label_p, clean_y))*100\n",
        "  print(\"Accuracy after dropping channel index: \", sorted_act_ind[channel_index], \" is: \", clean_accuracy)\n",
        "\n",
        "  pruned_clean_acc.append(clean_accuracy)\n",
        "\n",
        "  if (og_clean_accuracy-clean_accuracy >= 10):\n",
        "    print(\"The accuracy is dropped at least 10%, the model is saved!\")\n",
        "    model_copy.save('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_10.h5')\n",
        "\n",
        "    bd_label_p = np.argmax(model_copy.predict(bad_x), axis=1)\n",
        "    asr = np.mean(np.equal(bd_label_p, bad_y))*100\n",
        "    attack_success_rate.append(asr)\n",
        "    print()\n",
        "    print(\"The clean accuracy is: \",clean_accuracy)\n",
        "    print(\"The attack success rate is: \",asr)\n",
        "    print(\"The pruned channel index is: \",sorted_act_ind[channel_index])\n",
        "    break\n",
        "  channel_index += 1"
      ],
      "metadata": {
        "id": "R-TUOaSAhDi2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c187bbb2-3664-43f0-c6d8-34ca5de1d22a"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  46  is:  92.09318437689443\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  5  is:  91.49562656967177\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  8  is:  91.01931237550879\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  11  is:  89.17467740538669\n",
            "361/361 [==============================] - 1s 3ms/step\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy after dropping channel index:  54  is:  84.43751623798389\n",
            "The accuracy is dropped at least 10%, the model is saved!\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "\n",
            "The clean accuracy is:  84.43751623798389\n",
            "The attack success rate is:  77.015675067117\n",
            "The pruned channel index is:  54\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "keras.backend.clear_session()"
      ],
      "metadata": {
        "id": "cmJXPbVG0wuD"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Remaining channel pruning"
      ],
      "metadata": {
        "id": "qyQu8_mJ0xim"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "while channel_index < len(sorted_act_ind)-2:\n",
        "\n",
        "  weights[0][:,:,:,sorted_act_ind[channel_index]] = 0\n",
        "  weights[1][sorted_act_ind[channel_index]] = 0\n",
        "\n",
        "  model_copy.layers[5].set_weights(weights)\n",
        "\n",
        "  cl_label_p = np.argmax(model_copy.predict(clean_x), axis=1)\n",
        "  clean_accuracy = np.mean(np.equal(cl_label_p, clean_y))*100\n",
        "  print(\"Accuracy after dropping channel index: \", sorted_act_ind[channel_index], \" is: \", clean_accuracy)\n",
        "\n",
        "  pruned_clean_acc.append(clean_accuracy)\n",
        "\n",
        "  if channel_index==57:\n",
        "    bd_label_p = np.argmax(model_copy.predict(bad_x), axis=1)\n",
        "    asr = np.mean(np.equal(bd_label_p, bad_y))*100\n",
        "    attack_success_rate.append(asr)\n",
        "    print()\n",
        "    print(\"The clean accuracy is: \",clean_accuracy)\n",
        "    print(\"The attack success rate is: \",asr)\n",
        "    print(\"The pruned channel index is: \",sorted_act_ind[channel_index])\n",
        "  channel_index += 1\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1U-7DL9d0tqB",
        "outputId": "070aee6b-414e-4529-da25-8f5cf4c6cbbf"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  54  is:  84.43751623798389\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  10  is:  76.48739932449988\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  28  is:  54.8627349095003\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  35  is:  27.08928726076037\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  18  is:  13.87373343725643\n",
            "361/361 [==============================] - 1s 3ms/step\n",
            "Accuracy after dropping channel index:  4  is:  7.101411622066338\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "Accuracy after dropping channel index:  7  is:  1.5501861955486274\n",
            "361/361 [==============================] - 1s 2ms/step\n",
            "\n",
            "The clean accuracy is:  1.5501861955486274\n",
            "The attack success rate is:  0.0\n",
            "The pruned channel index is:  7\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "8BD9mVe6hDqo"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Could only prune till 58 our of 60 channels with the available RAM -- otherwise it was getting timed out -- but it is sufficient for the plot below"
      ],
      "metadata": {
        "id": "29KxzniqTCgr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "len(pruned_clean_acc)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ivFTeOPtu8FU",
        "outputId": "ba47f7b4-7ac2-40b7-d454-6bdcd15db4df"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "61"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## this is because the accuracy is appended twice at channel_index = 29, 46 and 54 -- let's drop it for the plot"
      ],
      "metadata": {
        "id": "wzX4naysyfkE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_clean_acc.pop(np.where(sorted_act_ind == 29)[0][0])\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eDarCamRye5A",
        "outputId": "d23ed3c0-8e93-4f76-ea94-3b62ed4015a7"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "95.75647354291158"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_clean_acc.pop(np.where(sorted_act_ind == 46)[0][0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UHZejk1nzuAb",
        "outputId": "2d6c1c82-b66e-4f56-8211-ba46e1f21870"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "92.09318437689443"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_clean_acc.pop(np.where(sorted_act_ind == 54)[0][0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MA7jLIPE1i4Q",
        "outputId": "c30b6992-a7ca-4b8d-9b10-1191382bb2dd"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "84.43751623798389"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(pruned_clean_acc)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5EPXsNnBzYyM",
        "outputId": "978ba9d2-7e6f-4275-86c1-5f2987c7343e"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "58"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### How does channel-by-channel pruning affects clean_accuracy and attack_success_rate"
      ],
      "metadata": {
        "id": "iYsyaB4JvDMT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# make attack success rate in proportion to channels ~ approximation\n",
        "i=0\n",
        "asr_plot = []\n",
        "for ci in sorted_act_ind:\n",
        "  print(i)\n",
        "  asr_plot.append(attack_success_rate[i])\n",
        "  if ci==29 or ci==46 or ci==54:\n",
        "    i+=1\n"
      ],
      "metadata": {
        "id": "bTgnOmZIvKlj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "asr_plot.pop()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QbAo1t3Gx0Vw",
        "outputId": "57cb8247-18be-4ffc-d5b6-db80654a4c56"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.0"
            ]
          },
          "metadata": {},
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "asr_plot.pop()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tpj2OfKrNG0G",
        "outputId": "fe870a9c-32ed-4d45-b0c1-bbee220a4f82"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.0"
            ]
          },
          "metadata": {},
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "len(asr_plot)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lQqp6WheNIsF",
        "outputId": "72d2f271-aa1c-402b-ca55-410387c2a314"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "58"
            ]
          },
          "metadata": {},
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### NOTE: I have done all these adjustments because:\n",
        "- the model could only be pruned till 58 channels\n",
        "- attack success rate was only calculated and stored at pruning thresholds i.e. 2,4,10 hence I extrapolated them to make it of length - 58\n",
        "- the plot below is an approximated plot-- but it gives the gist of how ASR and accuracy varies with the fraction of pruned channels"
      ],
      "metadata": {
        "id": "WDa020nlS47h"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "x_axis = np.arange(1,59)/59\n",
        "plt.plot(x_axis,pruned_clean_acc)\n",
        "plt.plot(x_axis,asr_plot)\n",
        "plt.legend(['Clean_accuracy','Attack Success Rate'])\n",
        "plt.xlabel(\"Fraction of Channels Pruned\")\n",
        "plt.ylabel(\"Rate\")\n",
        "plt.title(\"Accuracy and Attack Success Rate as a function of the Fraction of Channels Pruned\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 489
        },
        "id": "rowlVhyb0RB6",
        "outputId": "1d36be4b-d71a-4fd4-d2f0-27d54f11f575"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Accuracy and Attack Success Rate as a function of the Fraction of Channels Pruned')"
            ]
          },
          "metadata": {},
          "execution_count": 33
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "keras.backend.clear_session()"
      ],
      "metadata": {
        "id": "FS1sfrKpNyqi"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# pruning completed, now we'll make our goodnet"
      ],
      "metadata": {
        "id": "600HSCr9hFFQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "1. Output the correct class if the test input is clean. The correct class will be in [1,N].\n",
        "2. Output class N+1 if the input is backdoored\n",
        "\n",
        "Here, N = 1282"
      ],
      "metadata": {
        "id": "HLqMNtvVik8v"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "len(np.unique(clean_y))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tmE3RydbhDsx",
        "outputId": "c81b87c8-cd20-411f-b1a4-49a8eec7a0de"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1283"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "class GoodNet(keras.Model):\n",
        "\n",
        "  def __init__(self, BadNet, B_p):\n",
        "      super(GoodNet, self).__init__()\n",
        "      self.BadNet = BadNet\n",
        "      self.B_p = B_p\n",
        "\n",
        "  def predict(self,data):\n",
        "      y = np.argmax(self.BadNet(data), axis=1)\n",
        "      y_p = np.argmax(self.B_p(data), axis=1)\n",
        "      pred = np.zeros(data.shape[0])\n",
        "      for i in range(data.shape[0]):\n",
        "        if y[i]==y_p[i]:\n",
        "          pred[i] = y[i]\n",
        "        else:\n",
        "          pred[i] = 1283 # N+1\n",
        "      return pred"
      ],
      "metadata": {
        "id": "mZFHl1yxhDvd"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "48iTmrllhDx9"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Evaluate the goodnet"
      ],
      "metadata": {
        "id": "721kV_3-lYMV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_model_2 = keras.models.load_model('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_2.h5')\n",
        "pruned_model_4 = keras.models.load_model('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_4.h5')\n",
        "pruned_model_10 = keras.models.load_model('/content/drive/MyDrive/Sem 3/MLinCyberSec/Labs/Lab4/model/model_10.h5')"
      ],
      "metadata": {
        "id": "vcvmzALKlb2U",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "46d773ae-c59a-4cc1-a37c-fb67374d4bd8"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n",
            "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n",
            "WARNING:tensorflow:No training configuration found in the save file, so the model was *not* compiled. Compile it manually.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "good_model_2 = GoodNet(model, pruned_model_2)\n",
        "good_model_4 = GoodNet(model, pruned_model_4)\n",
        "good_model_10 = GoodNet(model, pruned_model_10)\n"
      ],
      "metadata": {
        "id": "2g41iyOM5L18"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# load test data\n",
        "clean_test_x, clean_test_y = load_data(test_clean_data)\n",
        "bad_test_x, bad_test_y = load_data(test_bad_data)"
      ],
      "metadata": {
        "id": "GzCfDFrtlcAO"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pruned_thresholds = [2,4,10]\n",
        "clean_good_model_acc = []\n",
        "clean_pruned_model_acc = []\n",
        "asr_good_model = []\n",
        "asr_pruned_model = []\n"
      ],
      "metadata": {
        "id": "uxUxk1P37e5C"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# Clean accuracy\n",
        "\n",
        "# Pruned\n",
        "print(\"Pruned model:\")\n",
        "cl_test_2_label_p = np.argmax(pruned_model_2.predict(clean_test_x), axis=1)\n",
        "clean_test_2_accuracy = np.mean(np.equal(cl_test_2_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 2% drop model:', clean_test_2_accuracy)\n",
        "clean_pruned_model_acc.append(clean_test_2_accuracy)\n",
        "\n",
        "cl_test_4_label_p = np.argmax(pruned_model_4.predict(clean_test_x), axis=1)\n",
        "clean_test_4_accuracy = np.mean(np.equal(cl_test_4_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 4% drop model:', clean_test_4_accuracy)\n",
        "clean_pruned_model_acc.append(clean_test_4_accuracy)\n",
        "\n",
        "cl_test_10_label_p = np.argmax(pruned_model_10.predict(clean_test_x), axis=1)\n",
        "clean_test_10_accuracy = np.mean(np.equal(cl_test_10_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 10% drop model:', clean_test_10_accuracy)\n",
        "clean_pruned_model_acc.append(clean_test_10_accuracy)\n",
        "\n",
        "# Good\n",
        "print(\"Good model:\")\n",
        "cl_test_2_label_p = good_model_2.predict(clean_test_x)\n",
        "clean_test_2_accuracy = np.mean(np.equal(cl_test_2_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 2% drop model:', clean_test_2_accuracy)\n",
        "clean_good_model_acc.append(clean_test_2_accuracy)\n",
        "\n",
        "cl_test_4_label_p = good_model_4.predict(clean_test_x)\n",
        "clean_test_4_accuracy = np.mean(np.equal(cl_test_4_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 4% drop model:', clean_test_4_accuracy)\n",
        "clean_good_model_acc.append(clean_test_4_accuracy)\n",
        "\n",
        "cl_test_10_label_p = good_model_10.predict(clean_test_x)\n",
        "clean_test_10_accuracy = np.mean(np.equal(cl_test_10_label_p, clean_test_y))*100\n",
        "print('Clean test data Classification accuracy for 10% drop model:', clean_test_10_accuracy)\n",
        "clean_good_model_acc.append(clean_test_10_accuracy)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U1hxg6wj_Y1d",
        "outputId": "7ac22627-d816-437c-9132-b6f6cd8976c5"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Pruned model:\n",
            "401/401 [==============================] - 1s 2ms/step\n",
            "Clean test data Classification accuracy for 2% drop model: 95.90023382696803\n",
            "401/401 [==============================] - 1s 2ms/step\n",
            "Clean test data Classification accuracy for 4% drop model: 92.29150428682775\n",
            "401/401 [==============================] - 1s 2ms/step\n",
            "Clean test data Classification accuracy for 10% drop model: 84.54403741231489\n",
            "Good model:\n",
            "Clean test data Classification accuracy for 2% drop model: 95.74434918160561\n",
            "Clean test data Classification accuracy for 4% drop model: 92.1278254091972\n",
            "Clean test data Classification accuracy for 10% drop model: 84.3335931410756\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Attack success rate\n",
        "\n",
        "# Pruned\n",
        "print(\"Pruned Model:\")\n",
        "bd_test_2_label_p = np.argmax(pruned_model_2.predict(bad_test_x), axis=1)\n",
        "asr_2 = np.mean(np.equal(bd_test_2_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 2% drop model:', asr_2)\n",
        "asr_pruned_model.append(asr_2)\n",
        "\n",
        "bd_test_4_label_p = np.argmax(pruned_model_4.predict(bad_test_x), axis=1)\n",
        "asr_4 = np.mean(np.equal(bd_test_4_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 4% drop model:', asr_4)\n",
        "asr_pruned_model.append(asr_4)\n",
        "\n",
        "bd_test_10_label_p = np.argmax(pruned_model_10.predict(bad_test_x), axis=1)\n",
        "asr_10 = np.mean(np.equal(bd_test_10_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 10% drop model:', asr_10)\n",
        "asr_pruned_model.append(asr_10)\n",
        "\n",
        "#Good\n",
        "print(\"Good model:\")\n",
        "bd_test_2_label_p = good_model_2.predict(bad_test_x)\n",
        "asr_2 = np.mean(np.equal(bd_test_2_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 2% drop model:', asr_2)\n",
        "asr_good_model.append(asr_2)\n",
        "\n",
        "bd_test_4_label_p = good_model_4.predict(bad_test_x)\n",
        "asr_4 = np.mean(np.equal(bd_test_4_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 4% drop model:', asr_4)\n",
        "asr_good_model.append(asr_4)\n",
        "\n",
        "bd_test_10_label_p = good_model_10.predict(bad_test_x)\n",
        "asr_10 = np.mean(np.equal(bd_test_10_label_p, bad_test_y))*100\n",
        "print('Attack success rate for 10% drop model:', asr_10)\n",
        "asr_good_model.append(asr_10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FXpqxYuX_rgQ",
        "outputId": "c347a06c-e1b8-488f-b493-0b0a6fdcbc85"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Pruned Model:\n",
            "401/401 [==============================] - 1s 3ms/step\n",
            "Attack success rate for 2% drop model: 100.0\n",
            "401/401 [==============================] - 1s 3ms/step\n",
            "Attack success rate for 4% drop model: 99.98441153546376\n",
            "401/401 [==============================] - 1s 2ms/step\n",
            "Attack success rate for 10% drop model: 77.20966484801247\n",
            "Good model:\n",
            "Attack success rate for 2% drop model: 100.0\n",
            "Attack success rate for 4% drop model: 99.98441153546376\n",
            "Attack success rate for 10% drop model: 77.20966484801247\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Summarizing models"
      ],
      "metadata": {
        "id": "1z4zFKniBkdt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Pruned model:\")\n",
        "test_acc = clean_pruned_model_acc\n",
        "attack_rate = asr_pruned_model\n",
        "data = {\n",
        "    \"test_acc\": test_acc,\n",
        "    \"attack_success_rate\": attack_rate,\n",
        "    \"model\": [\"pruned_model_2%\", \"pruned_model_4%\", \"pruned_model_10%\"]\n",
        "}\n",
        "df = pd.DataFrame(data)\n",
        "df.set_index('model')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "nDV8yN9el_4B",
        "outputId": "f4813c18-b203-4356-dc82-5696e7e81a3d"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Pruned model:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                   test_acc  attack_success_rate\n",
              "model                                           \n",
              "pruned_model_2%   95.900234           100.000000\n",
              "pruned_model_4%   92.291504            99.984412\n",
              "pruned_model_10%  84.544037            77.209665"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a0cd68ef-f800-4636-b165-b8ff4f657c12\" 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>test_acc</th>\n",
              "      <th>attack_success_rate</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>pruned_model_2%</th>\n",
              "      <td>95.900234</td>\n",
              "      <td>100.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pruned_model_4%</th>\n",
              "      <td>92.291504</td>\n",
              "      <td>99.984412</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pruned_model_10%</th>\n",
              "      <td>84.544037</td>\n",
              "      <td>77.209665</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-a0cd68ef-f800-4636-b165-b8ff4f657c12')\"\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-a0cd68ef-f800-4636-b165-b8ff4f657c12 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-a0cd68ef-f800-4636-b165-b8ff4f657c12');\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-e685819a-c25d-4fa8-a009-e1b770d9c091\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-e685819a-c25d-4fa8-a009-e1b770d9c091')\"\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-e685819a-c25d-4fa8-a009-e1b770d9c091 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Good model:\")\n",
        "test_acc = clean_good_model_acc\n",
        "attack_rate = asr_good_model\n",
        "data = {\n",
        "    \"test_acc\": test_acc,\n",
        "    \"attack_success_rate\": attack_rate,\n",
        "    \"model\": [\"good_model_2%\", \"good_model_4%\", \"good_model_10%\"]\n",
        "}\n",
        "df = pd.DataFrame(data)\n",
        "df.set_index('model')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 192
        },
        "id": "M-HI70P8B8rQ",
        "outputId": "7bd322e8-8ba4-42a7-bd80-8b2e3d1ba926"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Good model:\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                 test_acc  attack_success_rate\n",
              "model                                         \n",
              "good_model_2%   95.744349           100.000000\n",
              "good_model_4%   92.127825            99.984412\n",
              "good_model_10%  84.333593            77.209665"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-b43f7aa7-684c-40fd-b94b-401c8d02a105\" 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>test_acc</th>\n",
              "      <th>attack_success_rate</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>model</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>good_model_2%</th>\n",
              "      <td>95.744349</td>\n",
              "      <td>100.000000</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>good_model_4%</th>\n",
              "      <td>92.127825</td>\n",
              "      <td>99.984412</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>good_model_10%</th>\n",
              "      <td>84.333593</td>\n",
              "      <td>77.209665</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-b43f7aa7-684c-40fd-b94b-401c8d02a105')\"\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-b43f7aa7-684c-40fd-b94b-401c8d02a105 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-b43f7aa7-684c-40fd-b94b-401c8d02a105');\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-a56f1741-3a51-457a-a891-767c9b83bbf2\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a56f1741-3a51-457a-a891-767c9b83bbf2')\"\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-a56f1741-3a51-457a-a891-767c9b83bbf2 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Plotting\n",
        "bar_width = 0.2\n",
        "index = np.arange(len(pruned_thresholds))\n",
        "\n",
        "fig, ax = plt.subplots(figsize=(10, 6))\n",
        "rects1 = ax.bar(index - bar_width, clean_good_model_acc, bar_width, label='Clean Good Model')\n",
        "rects2 = ax.bar(index, clean_pruned_model_acc, bar_width, label='Clean Pruned Model')\n",
        "rects3 = ax.bar(index + bar_width, asr_good_model, bar_width, label='ASR Good Model')\n",
        "rects4 = ax.bar(index + 2 * bar_width, asr_pruned_model, bar_width, label='ASR Pruned Model')\n",
        "\n",
        "# Add values at the top of the bars\n",
        "def add_values(rects):\n",
        "    for rect in rects:\n",
        "        height = rect.get_height()\n",
        "        ax.annotate('{}'.format(round(height, 2)),\n",
        "                    xy=(rect.get_x() + rect.get_width() / 2, height),\n",
        "                    xytext=(0, 3),  # 3 points vertical offset\n",
        "                    textcoords=\"offset points\",\n",
        "                    ha='center', va='bottom')\n",
        "\n",
        "add_values(rects1)\n",
        "add_values(rects2)\n",
        "add_values(rects3)\n",
        "add_values(rects4)\n",
        "\n",
        "ax.set_xlabel('X Values')\n",
        "ax.set_ylabel('Accuracy')\n",
        "ax.set_title('Comparison plot with Values at the Top')\n",
        "ax.set_xticks(index)\n",
        "ax.set_xticklabels(pruned_thresholds)\n",
        "ax.legend(loc='lower center', bbox_to_anchor=(0.5, -0.25), ncol=4)\n",
        "# ax.legend()\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "6khngziZXB5p",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 629
        },
        "outputId": "dd0dd259-31f1-49eb-ec45-0990528de764"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Thanks!"
      ],
      "metadata": {
        "id": "Du3OB3g9UDDw"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "EhFIupj5RKmd"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}