{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table style=\"border: none\" align=\"center\">\n",
    "   <tr style=\"border: none\">\n",
    "      <th style=\"border: none\"><font face=\"verdana\" size=\"4\" color=\"black\"><b>  Demonstrate detection of adversarial samples using ART  </b></font></font></th>\n",
    "   </tr> \n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook we demonstrate the detection of adversarial samples using ART. Our classifier will be a **ResNet** architecture for the [CIFAR-10](https://www.cs.toronto.edu/~kriz/cifar.html) image data set.\n",
    "\n",
    "\n",
    "## Contents\n",
    "\n",
    "1.\t[Loading prereqs and data](#prereqs)\n",
    "2.  [Evaluating the classifier](#classifier)\n",
    "3.  [Training the detector](#train_detector)\n",
    "4.  [Evaluating the detector](#detector)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"prereqs\"></a>\n",
    "## 1. Loading prereqs and data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import load_model\n",
    "\n",
    "if tf.__version__[0] == \"2\":\n",
    "    tf.compat.v1.disable_eager_execution()\n",
    "\n",
    "from art import config\n",
    "from art.utils import load_dataset, get_file\n",
    "from art.estimators.classification import KerasClassifier\n",
    "from art.attacks.evasion import FastGradientMethod\n",
    "from art.defences.detector.evasion import BinaryInputDetector\n",
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the CIFAR10 data set and class descriptions:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "(x_train, y_train), (x_test, y_test), min_, max_ = load_dataset('cifar10')\n",
    "\n",
    "num_samples_train = 100\n",
    "num_samples_test = 100\n",
    "x_train = x_train[0:num_samples_train]\n",
    "y_train = y_train[0:num_samples_train]\n",
    "x_test = x_test[0:num_samples_test]\n",
    "y_test = y_test[0:num_samples_test]\n",
    "\n",
    "class_descr = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"classifier\"></a>\n",
    "## 2. Evaluating the classifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the pre-trained classifier (a ResNet architecture):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /usr/local/anaconda3/envs/art/lib/python3.8/site-packages/keras/layers/normalization/batch_normalization.py:532: _colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-17 15:00:34.867373: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "path = get_file('cifar_resnet.h5',extract=False, path=config.ART_DATA_PATH,\n",
    "                url='https://www.dropbox.com/s/ta75pl4krya5djj/cifar_resnet.h5?dl=1')\n",
    "classifier_model = load_model(path)\n",
    "classifier = KerasClassifier(clip_values=(min_, max_), model=classifier_model, use_logits=False, \n",
    "                             preprocessing=(0.5, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_1 (InputLayer)           [(None, 32, 32, 3)]  0           []                               \n",
      "                                                                                                  \n",
      " conv2d_1 (Conv2D)              (None, 32, 32, 16)   448         ['input_1[0][0]']                \n",
      "                                                                                                  \n",
      " batch_normalization_1 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_1[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_1 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_1[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_2 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_1[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_2 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_2[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_2 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_2[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_3 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_2[0][0]']           \n",
      "                                                                                                  \n",
      " add_1 (Add)                    (None, 32, 32, 16)   0           ['activation_1[0][0]',           \n",
      "                                                                  'conv2d_3[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_3 (BatchNo  (None, 32, 32, 16)  64          ['add_1[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_3 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_3[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_4 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_3[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_4 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_4[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_4 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_4[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_5 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_4[0][0]']           \n",
      "                                                                                                  \n",
      " add_2 (Add)                    (None, 32, 32, 16)   0           ['add_1[0][0]',                  \n",
      "                                                                  'conv2d_5[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_5 (BatchNo  (None, 32, 32, 16)  64          ['add_2[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_5 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_5[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_6 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_5[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_6 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_6[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_6 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_6[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_7 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_6[0][0]']           \n",
      "                                                                                                  \n",
      " add_3 (Add)                    (None, 32, 32, 16)   0           ['add_2[0][0]',                  \n",
      "                                                                  'conv2d_7[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_7 (BatchNo  (None, 32, 32, 16)  64          ['add_3[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_7 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_7[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_8 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_7[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_8 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_8[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_8 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_8[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_9 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_8[0][0]']           \n",
      "                                                                                                  \n",
      " add_4 (Add)                    (None, 32, 32, 16)   0           ['add_3[0][0]',                  \n",
      "                                                                  'conv2d_9[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_9 (BatchNo  (None, 32, 32, 16)  64          ['add_4[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_9 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_9[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_10 (Conv2D)             (None, 32, 32, 16)   2320        ['activation_9[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_10 (BatchN  (None, 32, 32, 16)  64          ['conv2d_10[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_10 (Activation)     (None, 32, 32, 16)   0           ['batch_normalization_10[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_11 (Conv2D)             (None, 32, 32, 16)   2320        ['activation_10[0][0]']          \n",
      "                                                                                                  \n",
      " add_5 (Add)                    (None, 32, 32, 16)   0           ['add_4[0][0]',                  \n",
      "                                                                  'conv2d_11[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_11 (BatchN  (None, 32, 32, 16)  64          ['add_5[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_11 (Activation)     (None, 32, 32, 16)   0           ['batch_normalization_11[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_12 (Conv2D)             (None, 16, 16, 32)   4640        ['activation_11[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_12 (BatchN  (None, 16, 16, 32)  128         ['conv2d_12[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_12 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_12[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_14 (Conv2D)             (None, 16, 16, 32)   544         ['add_5[0][0]']                  \n",
      "                                                                                                  \n",
      " conv2d_13 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_12[0][0]']          \n",
      "                                                                                                  \n",
      " add_6 (Add)                    (None, 16, 16, 32)   0           ['conv2d_14[0][0]',              \n",
      "                                                                  'conv2d_13[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_13 (BatchN  (None, 16, 16, 32)  128         ['add_6[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_13 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_13[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_15 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_13[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_14 (BatchN  (None, 16, 16, 32)  128         ['conv2d_15[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_14 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_14[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_16 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_14[0][0]']          \n",
      "                                                                                                  \n",
      " add_7 (Add)                    (None, 16, 16, 32)   0           ['add_6[0][0]',                  \n",
      "                                                                  'conv2d_16[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_15 (BatchN  (None, 16, 16, 32)  128         ['add_7[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_15 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_15[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_17 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_15[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_16 (BatchN  (None, 16, 16, 32)  128         ['conv2d_17[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_16 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_16[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_18 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_16[0][0]']          \n",
      "                                                                                                  \n",
      " add_8 (Add)                    (None, 16, 16, 32)   0           ['add_7[0][0]',                  \n",
      "                                                                  'conv2d_18[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_17 (BatchN  (None, 16, 16, 32)  128         ['add_8[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_17 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_17[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_19 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_17[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_18 (BatchN  (None, 16, 16, 32)  128         ['conv2d_19[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_18 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_18[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_20 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_18[0][0]']          \n",
      "                                                                                                  \n",
      " add_9 (Add)                    (None, 16, 16, 32)   0           ['add_8[0][0]',                  \n",
      "                                                                  'conv2d_20[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_19 (BatchN  (None, 16, 16, 32)  128         ['add_9[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_19 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_19[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_21 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_19[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_20 (BatchN  (None, 16, 16, 32)  128         ['conv2d_21[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_20 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_20[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_22 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_20[0][0]']          \n",
      "                                                                                                  \n",
      " add_10 (Add)                   (None, 16, 16, 32)   0           ['add_9[0][0]',                  \n",
      "                                                                  'conv2d_22[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_21 (BatchN  (None, 16, 16, 32)  128         ['add_10[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_21 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_21[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_23 (Conv2D)             (None, 8, 8, 64)     18496       ['activation_21[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_22 (BatchN  (None, 8, 8, 64)    256         ['conv2d_23[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_22 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_22[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_25 (Conv2D)             (None, 8, 8, 64)     2112        ['add_10[0][0]']                 \n",
      "                                                                                                  \n",
      " conv2d_24 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_22[0][0]']          \n",
      "                                                                                                  \n",
      " add_11 (Add)                   (None, 8, 8, 64)     0           ['conv2d_25[0][0]',              \n",
      "                                                                  'conv2d_24[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_23 (BatchN  (None, 8, 8, 64)    256         ['add_11[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_23 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_23[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_26 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_23[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_24 (BatchN  (None, 8, 8, 64)    256         ['conv2d_26[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_24 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_24[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_27 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_24[0][0]']          \n",
      "                                                                                                  \n",
      " add_12 (Add)                   (None, 8, 8, 64)     0           ['add_11[0][0]',                 \n",
      "                                                                  'conv2d_27[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_25 (BatchN  (None, 8, 8, 64)    256         ['add_12[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_25 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_25[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_28 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_25[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_26 (BatchN  (None, 8, 8, 64)    256         ['conv2d_28[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_26 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_26[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_29 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_26[0][0]']          \n",
      "                                                                                                  \n",
      " add_13 (Add)                   (None, 8, 8, 64)     0           ['add_12[0][0]',                 \n",
      "                                                                  'conv2d_29[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_27 (BatchN  (None, 8, 8, 64)    256         ['add_13[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_27 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_27[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_30 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_27[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_28 (BatchN  (None, 8, 8, 64)    256         ['conv2d_30[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_28 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_28[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_31 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_28[0][0]']          \n",
      "                                                                                                  \n",
      " add_14 (Add)                   (None, 8, 8, 64)     0           ['add_13[0][0]',                 \n",
      "                                                                  'conv2d_31[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_29 (BatchN  (None, 8, 8, 64)    256         ['add_14[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_29 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_29[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_32 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_29[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_30 (BatchN  (None, 8, 8, 64)    256         ['conv2d_32[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_30 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_30[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_33 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_30[0][0]']          \n",
      "                                                                                                  \n",
      " add_15 (Add)                   (None, 8, 8, 64)     0           ['add_14[0][0]',                 \n",
      "                                                                  'conv2d_33[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_31 (BatchN  (None, 8, 8, 64)    256         ['add_15[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_31 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_31[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_1 (Dropout)            (None, 8, 8, 64)     0           ['activation_31[0][0]']          \n",
      "                                                                                                  \n",
      " average_pooling2d_1 (AveragePo  (None, 1, 1, 64)    0           ['dropout_1[0][0]']              \n",
      " oling2D)                                                                                         \n",
      "                                                                                                  \n",
      " flatten_1 (Flatten)            (None, 64)           0           ['average_pooling2d_1[0][0]']    \n",
      "                                                                                                  \n",
      " classifier (Dense)             (None, 10)           650         ['flatten_1[0][0]']              \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 470,218\n",
      "Trainable params: 467,946\n",
      "Non-trainable params: 2,272\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "classifier_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the classifier on the first 100 test images:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original test data (first 100 images):\n",
      "Correctly classified: 98\n",
      "Incorrectly classified: 2\n"
     ]
    }
   ],
   "source": [
    "x_test_pred = np.argmax(classifier.predict(x_test[:100]), axis=1)\n",
    "nb_correct_pred = np.sum(x_test_pred == np.argmax(y_test[:100], axis=1))\n",
    "\n",
    "print(\"Original test data (first 100 images):\")\n",
    "print(\"Correctly classified: {}\".format(nb_correct_pred))\n",
    "print(\"Incorrectly classified: {}\".format(100-nb_correct_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For illustration purposes, look at the first 9 images. (In brackets: true labels.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 9 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "for i in range(0, 9):\n",
    "    pred_label, true_label = class_descr[x_test_pred[i]], class_descr[np.argmax(y_test[i])]\n",
    "    plt.subplot(330 + 1 + i)\n",
    "    fig=plt.imshow(x_test[i])\n",
    "    fig.axes.get_xaxis().set_visible(False)\n",
    "    fig.axes.get_yaxis().set_visible(False)\n",
    "    fig.axes.text(0.5, -0.1, pred_label + \" (\" + true_label + \")\", fontsize=12, transform=fig.axes.transAxes, \n",
    "                  horizontalalignment='center')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate some adversarial samples:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "attacker = FastGradientMethod(classifier, eps=0.05)\n",
    "x_test_adv = attacker.generate(x_test[:100]) # this takes about two minutes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the classifier on 100 adversarial samples:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adversarial test data (first 100 images):\n",
      "Correctly classified: 20\n",
      "Incorrectly classified: 80\n"
     ]
    }
   ],
   "source": [
    "x_test_adv_pred = np.argmax(classifier.predict(x_test_adv), axis=1)\n",
    "nb_correct_adv_pred = np.sum(x_test_adv_pred == np.argmax(y_test[:100], axis=1))\n",
    "\n",
    "print(\"Adversarial test data (first 100 images):\")\n",
    "print(\"Correctly classified: {}\".format(nb_correct_adv_pred))\n",
    "print(\"Incorrectly classified: {}\".format(100-nb_correct_adv_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now plot the adversarial images and their predicted labels (in brackets: true labels)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 9 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "for i in range(0, 9):\n",
    "    pred_label, true_label = class_descr[x_test_adv_pred[i]], class_descr[np.argmax(y_test[i])]\n",
    "    plt.subplot(330 + 1 + i)\n",
    "    fig=plt.imshow(x_test_adv[i])\n",
    "    fig.axes.get_xaxis().set_visible(False)\n",
    "    fig.axes.get_yaxis().set_visible(False)\n",
    "    fig.axes.text(0.5, -0.1, pred_label + \" (\" + true_label + \")\", fontsize=12, transform=fig.axes.transAxes, \n",
    "                  horizontalalignment='center')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"train_detector\"></a>\n",
    "## 3. Training the detector"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the detector model (which also uses a ResNet architecture):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = get_file('BID_eps=0.05.h5',extract=False, path=config.ART_DATA_PATH,\n",
    "                url='https://www.dropbox.com/s/cbyfk65497wwbtn/BID_eps%3D0.05.h5?dl=1')\n",
    "detector_model = load_model(path)\n",
    "detector_classifier = KerasClassifier(clip_values=(-0.5, 0.5), model=detector_model, use_logits=False)\n",
    "detector = BinaryInputDetector(detector_classifier)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "__________________________________________________________________________________________________\n",
      " Layer (type)                   Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      " input_1 (InputLayer)           [(None, 32, 32, 3)]  0           []                               \n",
      "                                                                                                  \n",
      " conv2d_1 (Conv2D)              (None, 32, 32, 16)   448         ['input_1[0][0]']                \n",
      "                                                                                                  \n",
      " batch_normalization_1 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_1[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_1 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_1[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_2 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_1[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_2 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_2[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_2 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_2[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_3 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_2[0][0]']           \n",
      "                                                                                                  \n",
      " add_1 (Add)                    (None, 32, 32, 16)   0           ['activation_1[0][0]',           \n",
      "                                                                  'conv2d_3[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_3 (BatchNo  (None, 32, 32, 16)  64          ['add_1[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_3 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_3[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_4 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_3[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_4 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_4[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_4 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_4[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_5 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_4[0][0]']           \n",
      "                                                                                                  \n",
      " add_2 (Add)                    (None, 32, 32, 16)   0           ['add_1[0][0]',                  \n",
      "                                                                  'conv2d_5[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_5 (BatchNo  (None, 32, 32, 16)  64          ['add_2[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_5 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_5[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_6 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_5[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_6 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_6[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_6 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_6[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_7 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_6[0][0]']           \n",
      "                                                                                                  \n",
      " add_3 (Add)                    (None, 32, 32, 16)   0           ['add_2[0][0]',                  \n",
      "                                                                  'conv2d_7[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_7 (BatchNo  (None, 32, 32, 16)  64          ['add_3[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_7 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_7[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_8 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_7[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_8 (BatchNo  (None, 32, 32, 16)  64          ['conv2d_8[0][0]']               \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_8 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_8[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_9 (Conv2D)              (None, 32, 32, 16)   2320        ['activation_8[0][0]']           \n",
      "                                                                                                  \n",
      " add_4 (Add)                    (None, 32, 32, 16)   0           ['add_3[0][0]',                  \n",
      "                                                                  'conv2d_9[0][0]']               \n",
      "                                                                                                  \n",
      " batch_normalization_9 (BatchNo  (None, 32, 32, 16)  64          ['add_4[0][0]']                  \n",
      " rmalization)                                                                                     \n",
      "                                                                                                  \n",
      " activation_9 (Activation)      (None, 32, 32, 16)   0           ['batch_normalization_9[0][0]']  \n",
      "                                                                                                  \n",
      " conv2d_10 (Conv2D)             (None, 32, 32, 16)   2320        ['activation_9[0][0]']           \n",
      "                                                                                                  \n",
      " batch_normalization_10 (BatchN  (None, 32, 32, 16)  64          ['conv2d_10[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_10 (Activation)     (None, 32, 32, 16)   0           ['batch_normalization_10[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_11 (Conv2D)             (None, 32, 32, 16)   2320        ['activation_10[0][0]']          \n",
      "                                                                                                  \n",
      " add_5 (Add)                    (None, 32, 32, 16)   0           ['add_4[0][0]',                  \n",
      "                                                                  'conv2d_11[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_11 (BatchN  (None, 32, 32, 16)  64          ['add_5[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_11 (Activation)     (None, 32, 32, 16)   0           ['batch_normalization_11[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_12 (Conv2D)             (None, 16, 16, 32)   4640        ['activation_11[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_12 (BatchN  (None, 16, 16, 32)  128         ['conv2d_12[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_12 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_12[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_14 (Conv2D)             (None, 16, 16, 32)   544         ['add_5[0][0]']                  \n",
      "                                                                                                  \n",
      " conv2d_13 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_12[0][0]']          \n",
      "                                                                                                  \n",
      " add_6 (Add)                    (None, 16, 16, 32)   0           ['conv2d_14[0][0]',              \n",
      "                                                                  'conv2d_13[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_13 (BatchN  (None, 16, 16, 32)  128         ['add_6[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_13 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_13[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_15 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_13[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_14 (BatchN  (None, 16, 16, 32)  128         ['conv2d_15[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_14 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_14[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_16 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_14[0][0]']          \n",
      "                                                                                                  \n",
      " add_7 (Add)                    (None, 16, 16, 32)   0           ['add_6[0][0]',                  \n",
      "                                                                  'conv2d_16[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_15 (BatchN  (None, 16, 16, 32)  128         ['add_7[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_15 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_15[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_17 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_15[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_16 (BatchN  (None, 16, 16, 32)  128         ['conv2d_17[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_16 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_16[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_18 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_16[0][0]']          \n",
      "                                                                                                  \n",
      " add_8 (Add)                    (None, 16, 16, 32)   0           ['add_7[0][0]',                  \n",
      "                                                                  'conv2d_18[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_17 (BatchN  (None, 16, 16, 32)  128         ['add_8[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_17 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_17[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_19 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_17[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_18 (BatchN  (None, 16, 16, 32)  128         ['conv2d_19[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_18 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_18[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_20 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_18[0][0]']          \n",
      "                                                                                                  \n",
      " add_9 (Add)                    (None, 16, 16, 32)   0           ['add_8[0][0]',                  \n",
      "                                                                  'conv2d_20[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_19 (BatchN  (None, 16, 16, 32)  128         ['add_9[0][0]']                  \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_19 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_19[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_21 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_19[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_20 (BatchN  (None, 16, 16, 32)  128         ['conv2d_21[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_20 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_20[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_22 (Conv2D)             (None, 16, 16, 32)   9248        ['activation_20[0][0]']          \n",
      "                                                                                                  \n",
      " add_10 (Add)                   (None, 16, 16, 32)   0           ['add_9[0][0]',                  \n",
      "                                                                  'conv2d_22[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_21 (BatchN  (None, 16, 16, 32)  128         ['add_10[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_21 (Activation)     (None, 16, 16, 32)   0           ['batch_normalization_21[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_23 (Conv2D)             (None, 8, 8, 64)     18496       ['activation_21[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_22 (BatchN  (None, 8, 8, 64)    256         ['conv2d_23[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_22 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_22[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_25 (Conv2D)             (None, 8, 8, 64)     2112        ['add_10[0][0]']                 \n",
      "                                                                                                  \n",
      " conv2d_24 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_22[0][0]']          \n",
      "                                                                                                  \n",
      " add_11 (Add)                   (None, 8, 8, 64)     0           ['conv2d_25[0][0]',              \n",
      "                                                                  'conv2d_24[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_23 (BatchN  (None, 8, 8, 64)    256         ['add_11[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_23 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_23[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_26 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_23[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_24 (BatchN  (None, 8, 8, 64)    256         ['conv2d_26[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_24 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_24[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_27 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_24[0][0]']          \n",
      "                                                                                                  \n",
      " add_12 (Add)                   (None, 8, 8, 64)     0           ['add_11[0][0]',                 \n",
      "                                                                  'conv2d_27[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_25 (BatchN  (None, 8, 8, 64)    256         ['add_12[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_25 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_25[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_28 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_25[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_26 (BatchN  (None, 8, 8, 64)    256         ['conv2d_28[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_26 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_26[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_29 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_26[0][0]']          \n",
      "                                                                                                  \n",
      " add_13 (Add)                   (None, 8, 8, 64)     0           ['add_12[0][0]',                 \n",
      "                                                                  'conv2d_29[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_27 (BatchN  (None, 8, 8, 64)    256         ['add_13[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_27 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_27[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_30 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_27[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_28 (BatchN  (None, 8, 8, 64)    256         ['conv2d_30[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_28 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_28[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_31 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_28[0][0]']          \n",
      "                                                                                                  \n",
      " add_14 (Add)                   (None, 8, 8, 64)     0           ['add_13[0][0]',                 \n",
      "                                                                  'conv2d_31[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_29 (BatchN  (None, 8, 8, 64)    256         ['add_14[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_29 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_29[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_32 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_29[0][0]']          \n",
      "                                                                                                  \n",
      " batch_normalization_30 (BatchN  (None, 8, 8, 64)    256         ['conv2d_32[0][0]']              \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_30 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_30[0][0]'] \n",
      "                                                                                                  \n",
      " conv2d_33 (Conv2D)             (None, 8, 8, 64)     36928       ['activation_30[0][0]']          \n",
      "                                                                                                  \n",
      " add_15 (Add)                   (None, 8, 8, 64)     0           ['add_14[0][0]',                 \n",
      "                                                                  'conv2d_33[0][0]']              \n",
      "                                                                                                  \n",
      " batch_normalization_31 (BatchN  (None, 8, 8, 64)    256         ['add_15[0][0]']                 \n",
      " ormalization)                                                                                    \n",
      "                                                                                                  \n",
      " activation_31 (Activation)     (None, 8, 8, 64)     0           ['batch_normalization_31[0][0]'] \n",
      "                                                                                                  \n",
      " dropout_1 (Dropout)            (None, 8, 8, 64)     0           ['activation_31[0][0]']          \n",
      "                                                                                                  \n",
      " average_pooling2d_1 (AveragePo  (None, 1, 1, 64)    0           ['dropout_1[0][0]']              \n",
      " oling2D)                                                                                         \n",
      "                                                                                                  \n",
      " flatten_1 (Flatten)            (None, 64)           0           ['average_pooling2d_1[0][0]']    \n",
      "                                                                                                  \n",
      " classifier (Dense)             (None, 2)            130         ['flatten_1[0][0]']              \n",
      "                                                                                                  \n",
      "==================================================================================================\n",
      "Total params: 469,698\n",
      "Trainable params: 467,426\n",
      "Non-trainable params: 2,272\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "detector_model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To train the detector:\n",
    "- we expand our training set with adversarial samples\n",
    "- we label the data with 0 (original) and 1 (adversarial)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train_adv = attacker.generate(x_train)\n",
    "nb_train = x_train.shape[0]\n",
    "\n",
    "x_train_detector = np.concatenate((x_train, x_train_adv), axis=0)\n",
    "y_train_detector = np.concatenate((np.array([[1,0]]*nb_train), np.array([[0,1]]*nb_train)), axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Perform the training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 200 samples\n",
      "Epoch 1/20\n",
      "200/200 [==============================] - 4s 22ms/sample - loss: 0.0062 - accuracy: 1.0000\n",
      "Epoch 2/20\n",
      "200/200 [==============================] - 1s 7ms/sample - loss: 0.0057 - accuracy: 1.0000\n",
      "Epoch 3/20\n",
      "200/200 [==============================] - 2s 8ms/sample - loss: 0.0056 - accuracy: 1.0000\n",
      "Epoch 4/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0076 - accuracy: 1.0000\n",
      "Epoch 5/20\n",
      "200/200 [==============================] - 2s 11ms/sample - loss: 0.0062 - accuracy: 1.0000\n",
      "Epoch 6/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0052 - accuracy: 1.0000\n",
      "Epoch 7/20\n",
      "200/200 [==============================] - 2s 8ms/sample - loss: 0.0051 - accuracy: 1.0000\n",
      "Epoch 8/20\n",
      "200/200 [==============================] - 2s 8ms/sample - loss: 0.0145 - accuracy: 0.9950\n",
      "Epoch 9/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0334 - accuracy: 0.9950\n",
      "Epoch 10/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0053 - accuracy: 1.0000\n",
      "Epoch 11/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0053 - accuracy: 1.0000\n",
      "Epoch 12/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0049 - accuracy: 1.0000\n",
      "Epoch 13/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0048 - accuracy: 1.0000\n",
      "Epoch 14/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0049 - accuracy: 1.0000\n",
      "Epoch 15/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0051 - accuracy: 1.0000\n",
      "Epoch 16/20\n",
      "200/200 [==============================] - 2s 10ms/sample - loss: 0.0047 - accuracy: 1.0000\n",
      "Epoch 17/20\n",
      "200/200 [==============================] - 2s 10ms/sample - loss: 0.0045 - accuracy: 1.0000\n",
      "Epoch 18/20\n",
      "200/200 [==============================] - 2s 9ms/sample - loss: 0.0043 - accuracy: 1.0000\n",
      "Epoch 19/20\n",
      "200/200 [==============================] - 2s 10ms/sample - loss: 0.0058 - accuracy: 1.0000\n",
      "Epoch 20/20\n",
      "200/200 [==============================] - 2s 10ms/sample - loss: 0.0043 - accuracy: 1.0000\n"
     ]
    }
   ],
   "source": [
    "detector.fit(x_train_detector, y_train_detector, nb_epochs=20, batch_size=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a id=\"detector\"></a>\n",
    "## 4. Evaluating the detector"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apply the detector to the adversarial test data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adversarial test data (first 100 images):\n",
      "Flagged: 100\n",
      "Not flagged: 0\n"
     ]
    }
   ],
   "source": [
    "_, is_adversarial = detector.detect(x_test_adv)\n",
    "flag_adv = np.sum(is_adversarial)\n",
    "\n",
    "print(\"Adversarial test data (first 100 images):\")\n",
    "print(\"Flagged: {}\".format(flag_adv))\n",
    "print(\"Not flagged: {}\".format(100 - flag_adv))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Apply the detector to the first 100 original test images:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original test data (first 100 images):\n",
      "Flagged: 100\n",
      "Not flagged: 0\n"
     ]
    }
   ],
   "source": [
    "_, is_adversarial = detector.detect(x_test[:100])\n",
    "flag_original = np.sum(is_adversarial)\n",
    "\n",
    "print(\"Original test data (first 100 images):\")\n",
    "print(\"Flagged: {}\".format(flag_original))\n",
    "print(\"Not flagged: {}\".format(100 - flag_original))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Evaluate the detector for different attack strengths `eps`\n",
    "(**Note**: for the training of detector, `eps=0.05` was used)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "eps_range = [0.01, 0.02, 0.03, 0.04, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9]\n",
    "nb_flag_adv = []\n",
    "nb_missclass = []\n",
    "\n",
    "for eps in eps_range:\n",
    "    attacker.set_params(**{'eps': eps})\n",
    "    x_test_adv = attacker.generate(x_test[:100])\n",
    "    nb_flag_adv += [np.sum(detector.detect(x_test_adv)[1])]\n",
    "    nb_missclass += [np.sum(np.argmax(classifier.predict(x_test_adv), axis=1) != np.argmax(y_test[:100], axis=1))]\n",
    "    \n",
    "eps_range = [0] + eps_range\n",
    "nb_flag_adv = [flag_original] + nb_flag_adv\n",
    "nb_missclass = [2] + nb_missclass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "ax.plot(np.array(eps_range)[:8], np.array(nb_flag_adv)[:8], 'b--', label='Detector flags')\n",
    "ax.plot(np.array(eps_range)[:8], np.array(nb_missclass)[:8], 'r--', label='Classifier errors')\n",
    "\n",
    "legend = ax.legend(loc='center right', shadow=True, fontsize='large')\n",
    "legend.get_frame().set_facecolor('#00FFCC')\n",
    "\n",
    "plt.xlabel('Attack strength (eps)')\n",
    "plt.ylabel('Per 100 adversarial samples')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "art",
   "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": "3.8.13"
  },
  "vscode": {
   "interpreter": {
    "hash": "30b15819a73bf738d2c051012d518af2175fe5da693b3ec4b95bab668851eb25"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}