ActivationDistrubution-Continuity-FC-MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-20T20:13:59.673640Z",
     "start_time": "2019-01-20T20:13:57.997310Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initialized TensorFlow\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline\n",
    "from experiment_train import *\n",
    "from experiment_datasets import *\n",
    "from helpers import *\n",
    "import matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_A(scale = 1.0, epochs = 15, N = [100, 100, 100], take_layer = 1):\n",
    "    # scaling the architecture up/down\n",
    "    N = [int(t * scale) for t in N]\n",
    "\n",
    "    def get_p_arr(p):\n",
    "        \"\"\" p array with failure on the first layer \"\"\"\n",
    "        return [0, p]  + [0] * len(N)\n",
    "\n",
    "    # Lips. coeff\n",
    "    KLips = 1.\n",
    "\n",
    "    # activation function\n",
    "    activation = 'sigmoid'\n",
    "\n",
    "    # training the network\n",
    "    exp = MNISTExperiment(N = N, p_inference = get_p_arr(0), p_train = get_p_arr(0), KLips = KLips, epochs = epochs,\n",
    "                                      activation = activation, reg_type = None, reg_coeff = 0, do_print = 'plot',\n",
    "                                      name = 'experiment_weights', seed = None, batch_size = 100)\n",
    "    \n",
    "    # using outputs of a hidden layer\n",
    "    model_partial = Model(inputs = exp.model_correct.inputs, outputs = exp.model_correct.layers[take_layer].output)\n",
    "    # using a fixed input\n",
    "    res = model_partial.predict(exp.x_train[0:1])\n",
    "    \n",
    "    # returning the activations in the middle\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total runs: 50\n"
     ]
    }
   ],
   "source": [
    "# all scales of the network (up and down)\n",
    "scales = np.linspace(0.5, 5, 5)\n",
    "\n",
    "# repetitions for each scale\n",
    "repetitions = 10\n",
    "\n",
    "print('Total runs: %d' % (len(scales) * repetitions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def process_scale(s, **kwargs):\n",
    "    \"\"\" Run many repetitions for one scale \"\"\"\n",
    "    return [get_A(scale = s, **kwargs) for _ in range(repetitions)]\n",
    "\n",
    "def process_all(**kwargs):\n",
    "    \"\"\" Random variable with all results \"\"\"\n",
    "    return {scale: process_scale(scale, **kwargs) for scale in tqdm(scales)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "  0%|          | 0/5 [00:00<?, ?it/s]\u001b[A\n",
      " 20%|██        | 1/5 [00:34<02:17, 34.25s/it]\u001b[A\n",
      " 40%|████      | 2/5 [01:28<02:00, 40.18s/it]\u001b[A\n",
      " 60%|██████    | 3/5 [02:45<01:42, 51.35s/it]\u001b[A\n",
      " 80%|████████  | 4/5 [04:31<01:07, 67.66s/it]\u001b[A\n",
      "100%|██████████| 5/5 [06:50<00:00, 88.97s/it]\u001b[A"
     ]
    }
   ],
   "source": [
    "# obtaining the results (no training)\n",
    "results = process_all(epochs = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# untrained\n",
    "for scale in scales:\n",
    "    plt.figure()\n",
    "    for i in range(repetitions):\n",
    "        plt.hist(results[scale][i].flatten(), alpha = 0.3)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Activations are ~ same"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# obtaining the results (no training)\n",
    "results = process_all(epochs = 20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# trained 20 epochs\n",
    "for scale in scales:\n",
    "    plt.figure()\n",
    "    for i in range(repetitions):\n",
    "        plt.hist(results[scale][i].flatten(), alpha = 0.3, bins = 100, density = True)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The distribution seems to stabilize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "notify_time": "5",
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "position": {
    "height": "407px",
    "left": "1561px",
    "right": "20px",
    "top": "120px",
    "width": "339px"
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}