Homework_2_wz2295.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YBipHxOEQyOL"
      },
      "source": [
        "# Designing a backdoor detector for BadNets trained on the YouTube Face dataset using the pruning defense."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "L0N_Ne9jQyOQ"
      },
      "outputs": [],
      "source": [
        "# All necessary imports\n",
        "import os\n",
        "import tarfile\n",
        "import requests\n",
        "import re\n",
        "import sys\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')\n",
        "import h5py\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from keras import backend as K\n",
        "from keras.models import Model\n",
        "import matplotlib.pyplot as plt\n",
        "from mpl_toolkits.axes_grid1.inset_locator import inset_axes\n",
        "import matplotlib.font_manager as font_manager\n",
        "import cv2"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FnY1v5SPLm2u",
        "outputId": "5a80e936-de2a-4231-e766-4f789ae5e231"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pKN7C7QtgK3A"
      },
      "source": [
        "Define function to load the data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "PyyUFi7XgK3A"
      },
      "outputs": [],
      "source": [
        "# Load data\n",
        "def data_loader(filepath):\n",
        "    data = h5py.File(filepath, '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",
        "    return x_data, y_data"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LMAWYQ9rQyOV"
      },
      "source": [
        "Follow instructions under [Data Section](https://github.com/csaw-hackml/CSAW-HackML-2020/tree/master/lab3#ii-data) to download the datasets.  \n",
        "\n",
        "We will be using the clean validation data (valid.h5) from cl folder to design the defense and clean test data (test.h5 from cl folder) and  sunglasses poisoned test data (bd_test.h5 from bd folder) to evaluate the models."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "8xOZ3XdWQyOX"
      },
      "outputs": [],
      "source": [
        "## To-do ##\n",
        "# After downloading the datasets, provide corresponding filepaths below \n",
        "\n",
        "clean_data_valid_filename = \"/content/gdrive/My Drive/valid.h5\"\n",
        "\n",
        "clean_data_test_filename = \"/content/gdrive/My Drive/test.h5\"\n",
        "poisoned_data_test_filename = \"/content/gdrive/My Drive/bd_test.h5\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cffWMErcgK3G"
      },
      "source": [
        "Read the data:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "MvrR2idwgK3G"
      },
      "outputs": [],
      "source": [
        "cl_x_valid, cl_y_valid = data_loader(clean_data_valid_filename)#用来训练B'\n",
        "\n",
        "cl_x_test, cl_y_test = data_loader(clean_data_test_filename)#用来测试G\n",
        "bd_x_test, bd_y_test = data_loader(poisoned_data_test_filename)#用来测试G"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UHFpm5YDgK3G"
      },
      "source": [
        "Visualizing the clean test data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "uYh01TSkgK3G",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "outputId": "95ce8398-79b1-4b48-dd32-1b479bf66658"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x432 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Plot some images from the validation set (see https://mrdatascience.com/how-to-plot-mnist-digits-using-matplotlib/)\n",
        "num = 10\n",
        "np.random.seed(45)\n",
        "randIdx = [np.random.randint(10000) for i in range(num)]\n",
        "num_row = 2\n",
        "num_col = 5# plot images\n",
        "fig, axes = plt.subplots(num_row, num_col, figsize=(3*num_col,3*num_row))\n",
        "for i in range(num):\n",
        "    ax = axes[i//num_col, i%num_col]\n",
        "    ax.imshow(cl_x_test[randIdx[i]].astype('uint8'))\n",
        "    ax.set_title('label: {:.0f}'.format(cl_y_test[randIdx[i]]))\n",
        "    ax.set_xticks([])\n",
        "    ax.set_yticks([])\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ufr1VMDygK3H"
      },
      "source": [
        "Visualizing the sunglasses poisioned test data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "zAJOQYHogK3H",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "outputId": "2315f6a9-4c4c-4422-cdca-d1a67a12ff10"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x432 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# Plot some images from the validation set (see https://mrdatascience.com/how-to-plot-mnist-digits-using-matplotlib/)\n",
        "num = 10\n",
        "np.random.seed(45)\n",
        "randIdx = [np.random.randint(10000) for i in range(num)]\n",
        "num_row = 2\n",
        "num_col = 5# plot images\n",
        "fig, axes = plt.subplots(num_row, num_col, figsize=(3*num_col,3*num_row))\n",
        "for i in range(num):\n",
        "    ax = axes[i//num_col, i%num_col]\n",
        "    ax.imshow(bd_x_test[randIdx[i]].astype('uint8'))\n",
        "    ax.set_title('label: {:.0f}'.format(bd_y_test[randIdx[i]]))\n",
        "    ax.set_xticks([])\n",
        "    ax.set_yticks([])\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WTlCTxc1gK3I"
      },
      "source": [
        "Load the backdoored model.\n",
        "\n",
        "The backdoor model and its weights can be found [here](https://github.com/csaw-hackml/CSAW-HackML-2020/tree/master/lab3/models) "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 48,
      "metadata": {
        "id": "eJj3T9sxgK3L",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c704d774-fec8-4f16-8883-dc4b0ec40154"
      },
      "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: 601,643\n",
            "Trainable params: 601,643\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "## To-do ##\n",
        "\n",
        "# First create clones of the original badnet model (by providing the model filepath below)\n",
        "# The result of repairing B_clone will be B_prime\n",
        "\n",
        "B = keras.models.load_model(\"/content/gdrive/My Drive/bd_net.h5\")\n",
        "#B.load_weights(\"\")\n",
        "\n",
        "B_clone = keras.models.load_model(\"/content/gdrive/My Drive/bd_net.h5\")\n",
        "#B_clone.load_weights(\"\")\n",
        "B.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Output of the original badnet accuracy on the validation data:"
      ],
      "metadata": {
        "id": "tZEH7Ue-3cWk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Get the original badnet model's (B) accuracy on the validation data\n",
        "cl_label_p = np.argmax(B(cl_x_valid), axis=1)\n",
        "clean_accuracy = np.mean(np.equal(cl_label_p, cl_y_valid)) * 100\n",
        "\n",
        "print(\"Clean validation accuracy before pruning {0:3.6f}\".format(clean_accuracy))\n",
        "K.clear_session()"
      ],
      "metadata": {
        "id": "epkj_VOC3bs6",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8dd1bc8f-89e7-44b8-d1d0-fb4172fd7628"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Clean validation accuracy before pruning 98.649000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "## To-do ##\n",
        "## use for generating the plot\n",
        "\n",
        "# intermediate_model = Model(inputs=B.inputs, outputs=B.get_layer('pool_3').output)\n",
        "# feature_maps_cl = intermediate_model.predict(cl_x_valid)\n",
        "# averageActivationsCl = np.mean(np.array(feature_maps_cl), axis= 0)\n",
        "# idxToPrune = np.argsort(np.sum(averageActivationsCl,axis = (0,1)))\n",
        "# conv3_layer = B.get_layer('conv_3')\n",
        "# lastConvLayerWeights, lastConvLayerBiases = conv3_layer.get_weights()\n",
        "# i=0\n",
        "# acc=[]\n",
        "# asr=[]\n",
        "# for chIdx in idxToPrune:\n",
        "#   lastConvLayerWeights[:,:,:,chIdx]=0\n",
        "#   lastConvLayerBiases[chIdx]= 0 \n",
        "#   B_clone.get_layer('conv_3').set_weights([lastConvLayerWeights,lastConvLayerBiases])\n",
        "#   cl_label_p_valid = np.argmax(B_clone(cl_x_valid), axis=1)\n",
        "#   clean_accuracy_valid = np.mean(np.equal(cl_label_p_valid, cl_y_valid)) * 100\n",
        "#   acc.append(clean_accuracy_valid)\n",
        "#   bd_label_p = np.argmax(B_clone(bd_x_test), axis=1)\n",
        "#   asr_bd = np.mean(np.equal(bd_label_p, bd_y_test)) * 100\n",
        "#   asr.append(asr_bd)\n",
        "#   i+=1\n",
        "#   print('epoch:',i)\n",
        "#   print('Clean Classification accuracy',clean_accuracy_valid)\n",
        "#   print('Attack Success Rate',asr_bd)\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XUnIJq_qnzxh",
        "outputId": "42012bc7-da25-48f0-d8d7-502481a692bc"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 10s 28ms/step\n",
            "epoch: 1\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 2\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 3\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 4\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 5\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 6\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 7\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 8\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 9\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 10\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 11\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 12\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 13\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 14\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 15\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 16\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 17\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 18\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 19\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 20\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 21\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 22\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 23\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 24\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 25\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 26\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 27\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 28\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 29\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 30\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 31\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 32\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 33\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 34\n",
            "Clean Classification accuracy 98.64033948211657\n",
            "Attack Success Rate 100.0\n",
            "epoch: 35\n",
            "Clean Classification accuracy 98.64033948211657\n",
            "Attack Success Rate 100.0\n",
            "epoch: 36\n",
            "Clean Classification accuracy 98.63167922404088\n",
            "Attack Success Rate 100.0\n",
            "epoch: 37\n",
            "Clean Classification accuracy 98.65765999826795\n",
            "Attack Success Rate 100.0\n",
            "epoch: 38\n",
            "Clean Classification accuracy 98.64899974019225\n",
            "Attack Success Rate 100.0\n",
            "epoch: 39\n",
            "Clean Classification accuracy 98.6056984498138\n",
            "Attack Success Rate 100.0\n",
            "epoch: 40\n",
            "Clean Classification accuracy 98.57105741751104\n",
            "Attack Success Rate 100.0\n",
            "epoch: 41\n",
            "Clean Classification accuracy 98.53641638520828\n",
            "Attack Success Rate 100.0\n",
            "epoch: 42\n",
            "Clean Classification accuracy 98.19000606218066\n",
            "Attack Success Rate 100.0\n",
            "epoch: 43\n",
            "Clean Classification accuracy 97.65307006148784\n",
            "Attack Success Rate 100.0\n",
            "epoch: 44\n",
            "Clean Classification accuracy 97.50584567420108\n",
            "Attack Success Rate 100.0\n",
            "epoch: 45\n",
            "Clean Classification accuracy 95.75647354291158\n",
            "Attack Success Rate 100.0\n",
            "epoch: 46\n",
            "Clean Classification accuracy 95.20221702606739\n",
            "Attack Success Rate 99.97661730319564\n",
            "epoch: 47\n",
            "Clean Classification accuracy 94.7172425738287\n",
            "Attack Success Rate 99.98441153546376\n",
            "epoch: 48\n",
            "Clean Classification accuracy 92.09318437689443\n",
            "Attack Success Rate 99.98441153546376\n",
            "epoch: 49\n",
            "Clean Classification accuracy 91.49562656967177\n",
            "Attack Success Rate 99.98441153546376\n",
            "epoch: 50\n",
            "Clean Classification accuracy 91.01931237550879\n",
            "Attack Success Rate 99.97661730319564\n",
            "epoch: 51\n",
            "Clean Classification accuracy 89.17467740538669\n",
            "Attack Success Rate 80.6469212782541\n",
            "epoch: 52\n",
            "Clean Classification accuracy 84.43751623798389\n",
            "Attack Success Rate 77.20966484801247\n",
            "epoch: 53\n",
            "Clean Classification accuracy 76.48739932449988\n",
            "Attack Success Rate 36.26656274356976\n",
            "epoch: 54\n",
            "Clean Classification accuracy 54.8627349095003\n",
            "Attack Success Rate 6.96024941543258\n",
            "epoch: 55\n",
            "Clean Classification accuracy 27.08928726076037\n",
            "Attack Success Rate 0.4208885424785659\n",
            "epoch: 56\n",
            "Clean Classification accuracy 13.87373343725643\n",
            "Attack Success Rate 0.0\n",
            "epoch: 57\n",
            "Clean Classification accuracy 7.101411622066338\n",
            "Attack Success Rate 0.0\n",
            "epoch: 58\n",
            "Clean Classification accuracy 1.5501861955486274\n",
            "Attack Success Rate 0.0\n",
            "epoch: 59\n",
            "Clean Classification accuracy 0.7188014202823244\n",
            "Attack Success Rate 0.0\n",
            "epoch: 60\n",
            "Clean Classification accuracy 0.0779423226812159\n",
            "Attack Success Rate 0.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# import matplotlib.pyplot as plt\n",
        "# plt.figure(figsize=(8, 8))\n",
        "# frac = np.array(range(60))/59\n",
        "# plt.plot(frac, acc / acc[0], label='acc')\n",
        "# plt.plot(frac, asr / asr[0], label='asr')\n",
        "# plt.xlabel('fraction of pruned channels')\n",
        "# plt.ylabel('acc & asr rate')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 514
        },
        "id": "3xaPC7ihuFSO",
        "outputId": "0a45b895-22fe-40af-c589-d7de75a25a6d"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'acc & asr rate')"
            ]
          },
          "metadata": {},
          "execution_count": 39
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bZmVoTf9gK3L"
      },
      "source": [
        "Write code to implement pruning defense "
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "## To-do ##\n",
        "\n",
        "# Redefine model to output right after the last pooling layer (\"pool_3\")\n",
        "intermediate_model = Model(inputs=B.inputs, outputs=B.get_layer('pool_3').output)\n",
        "\n",
        "# Get feature map for last pooling layer (\"pool_3\") using the clean validation data and intermediate_model\n",
        "feature_maps_cl = intermediate_model.predict(cl_x_valid)\n",
        "\n",
        "# Get average activation value of each channel in last pooling layer (\"pool_3\")\n",
        "\n",
        "averageActivationsCl = np.mean(np.array(feature_maps_cl), axis= 0)\n",
        "# Store the indices of average activation values (averageActivationsCl) in increasing order\n",
        "idxToPrune = np.argsort(np.sum(averageActivationsCl,axis = (0,1)))\n",
        "\n",
        "# Get the conv_3 layer weights and biases from the original network that will be used for prunning\n",
        "# Hint: Use the get_weights() method (https://stackoverflow.com/questions/43715047/how-do-i-get-the-weights-of-a-layer-in-keras)\n",
        "conv3_layer = B.get_layer('conv_3')\n",
        "lastConvLayerWeights, lastConvLayerBiases = conv3_layer.get_weights()\n",
        "X = 0.10\n",
        "for chIdx in idxToPrune:\n",
        "  # Prune one channel at a time\n",
        "  # Hint: Replace all values in channel 'chIdx' of lastConvLayerWeights and lastConvLayerBiases with 0 \n",
        "  lastConvLayerWeights[:,:,:,chIdx]=0\n",
        "  lastConvLayerBiases[chIdx]= 0 #??\n",
        "\n",
        "  # Update weights and biases of B_clone\n",
        "  # Hint: Use the set_weights() method \n",
        "  B_clone.get_layer('conv_3').set_weights([lastConvLayerWeights,lastConvLayerBiases])\n",
        "\n",
        "  # Evaluate the updated model's (B_clone) clean validation accuracy\n",
        "  cl_label_p_valid = np.argmax(B_clone(cl_x_valid), axis=1)\n",
        "  clean_accuracy_valid = np.mean(np.equal(cl_label_p_valid, cl_y_valid)) * 100\n",
        "\n",
        "  # If drop in clean_accuracy_valid is just greater (or equal to) than the desired threshold compared to clean_accuracy, then save B_clone as B_prime and break\n",
        "  if (clean_accuracy_valid/clean_accuracy)<=(1-X):\n",
        "    # Save B_clone as B_prime and break\n",
        "    B_prime = B_clone\n",
        "    \n",
        "    #print(\"Clean validation accuracy before pruning {0:3.6f}\".format(clean_accuracy_valid))\n",
        "    break\n"
      ],
      "metadata": {
        "id": "cc2Z69AJlvk7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "660967fd-e8f0-4e32-e5b2-a586c9de7eac"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "361/361 [==============================] - 12s 33ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "B_prime.save('/content/bprime_0.10.h5')"
      ],
      "metadata": {
        "id": "lo3gpuKjZ3PW"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "clean_accuracy_valid"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mkpuw92BORgi",
        "outputId": "c9bc1fbe-2b25-4e58-cd6d-763e5e9cb50b"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "84.43751623798389"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5rgfUVofgK3N"
      },
      "source": [
        "Now we need to combine the models into a repaired goodnet G  that outputs the correct class if the test input is clean and class N+1 if the input is backdoored. One way to do it is to \"subclass\" the models in Keras: "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "YHgplkCNgK3O"
      },
      "outputs": [],
      "source": [
        "#https://stackoverflow.com/questions/64983112/keras-vertical-ensemble-model-with-condition-in-between\n",
        "class G(tf.keras.Model):\n",
        "    def __init__(self, B, B_prime):\n",
        "        super(G, self).__init__()\n",
        "        self.B = B\n",
        "        self.B_prime = B_prime\n",
        "\n",
        "    def predict(self,data):\n",
        "        y = np.argmax(self.B(data), axis=1)\n",
        "        y_prime = np.argmax(self.B_prime(data), axis=1)\n",
        "        tmpRes = np.array([y[i] if y[i] == y_prime[i] else 1283 for i in range(y.shape[0])])\n",
        "        res = np.zeros((y.shape[0],1284))\n",
        "        res[np.arange(tmpRes.size),tmpRes] = 1\n",
        "        return res\n",
        "\n",
        "    # For small amount of inputs that fit in one batch, directly using call() is recommended for faster execution, \n",
        "    # e.g., model(x), or model(x, training=False) is faster then model.predict(x) and do not result in\n",
        "    # memory leaks (see for more details https://www.tensorflow.org/api_docs/python/tf/keras/Model#predict)\n",
        "    def call(self,data):\n",
        "        y = np.argmax(self.B(data), axis=1)\n",
        "        y_prime = np.argmax(self.B_prime(data), axis=1)\n",
        "        tmpRes = np.array([y[i] if y[i] == y_prime[i] else 1283 for i in range(y.shape[0])])\n",
        "        res = np.zeros((y.shape[0],1284))\n",
        "        res[np.arange(tmpRes.size),tmpRes] = 1\n",
        "        return res"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vOso1yO3XT4U"
      },
      "source": [
        "However, Keras prevents from saving this kind of subclassed model as HDF5 file since it is not serializable. However, we still can use this architecture for model evaluation."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-FRYBej5gK3O"
      },
      "source": [
        "Load the saved B_prime model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "BpG2-c2cgK3O"
      },
      "outputs": [],
      "source": [
        "## To-do ##\n",
        "# Provide B_prime model filepath below\n",
        "\n",
        "B_prime = keras.models.load_model(\"/content/bprime_0.10.h5\")\n",
        "B_prime.load_weights(\"/content/bprime_0.10.h5\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Los0PV7lh38i"
      },
      "source": [
        "Check performance of the repaired model on the test data:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "5gLtzNBlh2hz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "aaac6b42-15e3-4ecd-bf7d-4123b2961ace"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "401/401 [==============================] - 13s 31ms/step\n",
            "Clean Classification accuracy for B_prime: 84.54403741231489\n",
            "401/401 [==============================] - 12s 31ms/step\n",
            "Attack Success Rate for B_prime: 77.20966484801247\n"
          ]
        }
      ],
      "source": [
        "cl_label_p = np.argmax(B_prime.predict(cl_x_test), axis=1)\n",
        "clean_accuracy_B_prime = np.mean(np.equal(cl_label_p, cl_y_test))*100\n",
        "print('Clean Classification accuracy for B_prime:', clean_accuracy_B_prime)\n",
        "\n",
        "bd_label_p = np.argmax(B_prime.predict(bd_x_test), axis=1)\n",
        "asr_B_prime = np.mean(np.equal(bd_label_p, bd_y_test))*100\n",
        "print('Attack Success Rate for B_prime:', asr_B_prime)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Check performance of the original model on the test data:"
      ],
      "metadata": {
        "id": "fuMWKWLpKAyj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "cl_label_p = np.argmax(B.predict(cl_x_test), axis=1)\n",
        "clean_accuracy_B = np.mean(np.equal(cl_label_p, cl_y_test))*100\n",
        "print('Clean Classification accuracy for B:', clean_accuracy_B)\n",
        "\n",
        "bd_label_p = np.argmax(B.predict(bd_x_test), axis=1)\n",
        "asr_B = np.mean(np.equal(bd_label_p, bd_y_test))*100\n",
        "print('Attack Success Rate for B:', asr_B)"
      ],
      "metadata": {
        "id": "k201G20WKCIt",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "9679e259-1cb8-4b82-c034-df499c393fa2"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "401/401 [==============================] - 12s 31ms/step\n",
            "Clean Classification accuracy for B: 98.62042088854248\n",
            "401/401 [==============================] - 12s 31ms/step\n",
            "Attack Success Rate for B: 100.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KQMLO91SgK3O"
      },
      "source": [
        "Create repaired network"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "eSNrvmc6gK3O"
      },
      "outputs": [],
      "source": [
        "# Repaired network repaired_net\n",
        "repaired_net = G(B, B_prime)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jtViN73fgK3Q"
      },
      "source": [
        "Check the performance of the repaired_net on the test data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "Dn4HUhIcgK3Q",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ed58bc57-4d54-4182-b47c-fa10e2171b1c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Clean Classification accuracy for repaired net: 84.3335931410756\n",
            "Attack Success Rate for repaired net: 77.20966484801247\n"
          ]
        }
      ],
      "source": [
        "cl_label_p = np.argmax(repaired_net(cl_x_test), axis=1)\n",
        "clean_accuracy_repaired_net = np.mean(np.equal(cl_label_p, cl_y_test))*100\n",
        "print('Clean Classification accuracy for repaired net:', clean_accuracy_repaired_net)\n",
        "\n",
        "bd_label_p = np.argmax(repaired_net(bd_x_test), axis=1)\n",
        "asr_repaired_net = np.mean(np.equal(bd_label_p, bd_y_test))*100\n",
        "print('Attack Success Rate for repaired net:', asr_repaired_net)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.4"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}