examples/alibi_detect_deploy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Outlier, adversarial and drift detection on CIFAR10\n",
    "\n",
    "* [0. Dataset](#dataset)\n",
    "\n",
    "\n",
    "* [1. Outlier detection with a variational autoencoder (VAE)](#od)\n",
    "    * [Method](#od_method)\n",
    "    * [Load detector or train from scratch](#od_load)\n",
    "    * [Setting the threshold](#od_threshold)\n",
    "    * [Create and detect outliers](#od_detect)\n",
    "    * [Deploy](#od_deploy)\n",
    "\n",
    "\n",
    "* [2. Adversarial detection by matching prediction probabilities](#ad)\n",
    "    * [Method](#ad_method)\n",
    "    * [Utility functions](#ad_utils)\n",
    "    * [Rescale data](#ad_rescale)\n",
    "    * [Load pre-trained classifier](#ad_load_model)\n",
    "    * [Adversarial attack](#ad_attack)\n",
    "    * [Load or train and evaluate the adversarial detector](#ad_load_detector)\n",
    "\n",
    "\n",
    "* [3. Drift detection with Kolmogorov-Smirnov](#cd)\n",
    "    * [Method](#cd_method)\n",
    "    * [Dataset](#cd_dataset)\n",
    "    * [Detect drift](#cd_detect)\n",
    "        * [Untrained AutoEncoder](#cd_uae)\n",
    "        * [BBSDs](#cd_bbsds)\n",
    "    * [Leveraging the adversarial detector for malicious drift detection](#cd_ad)\n",
    "    * [Deploy](#cd_deploy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 0. Dataset <a class=\"anchor\" id=\"dataset\"></a>\n",
    "\n",
    "[CIFAR10](https://www.cs.toronto.edu/~kriz/cifar.html) consists of 60,000 32 by 32 RGB images equally distributed over 10 classes: airplane, automobile, bird, cat, deer, dog, frog, horse, ship and truck."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train:  (50000, 32, 32, 3) (50000,)\n",
      "Test:  (10000, 32, 32, 3) (10000,)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# imports and plot examples\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import tensorflow as tf\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = tf.keras.datasets.cifar10.load_data()\n",
    "X_train = X_train.astype('float32') / 255\n",
    "X_test = X_test.astype('float32') / 255\n",
    "y_train = y_train.astype('int64').reshape(-1,)\n",
    "y_test = y_test.astype('int64').reshape(-1,)\n",
    "print('Train: ', X_train.shape, y_train.shape)\n",
    "print('Test: ', X_test.shape, y_test.shape)\n",
    "\n",
    "plt.figure(figsize=(10, 10))\n",
    "n = 4\n",
    "for i in range(n ** 2):\n",
    "    plt.subplot(n, n, i + 1)\n",
    "    plt.imshow(X_train[i])\n",
    "    plt.axis('off')\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Outlier detection with a variational autoencoder (VAE) <a class=\"anchor\" id=\"od\"></a>\n",
    "\n",
    "### Method <a class=\"anchor\" id=\"od_method\"></a>\n",
    "\n",
    "In a nutshell:\n",
    "\n",
    "- Train a VAE on *normal* data so it can reconstruct *inliers* well\n",
    "- If the VAE cannot reconstruct the incoming requests well? Outlier!\n",
    "\n",
    "More resources on VAE: [paper](https://arxiv.org/abs/1312.6114), [excellent blog post](https://lilianweng.github.io/lil-log/2018/08/12/from-autoencoder-to-beta-vae.html#vae-variational-autoencoder)\n",
    "\n",
    "![vae_lillog](image/vae_lillog.png)\n",
    "\n",
    "*Image source: https://lilianweng.github.io/lil-log/2018/08/12/from-autoencoder-to-beta-vae.html*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [],
   "source": [
    "# more imports\n",
    "import logging\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "from tensorflow.keras.layers import Conv2D, Conv2DTranspose, Dense\n",
    "from tensorflow.keras.layers import Flatten, Layer, Reshape, InputLayer\n",
    "from tensorflow.keras.regularizers import l1\n",
    "\n",
    "from alibi_detect.od import OutlierVAE\n",
    "from alibi_detect.utils.fetching import fetch_detector\n",
    "from alibi_detect.utils.perturbation import apply_mask\n",
    "from alibi_detect.utils.saving import save_detector, load_detector\n",
    "from alibi_detect.utils.visualize import plot_instance_score, plot_feature_outlier_image\n",
    "\n",
    "logger = tf.get_logger()\n",
    "logger.setLevel(logging.ERROR)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load detector or train from scratch <a class=\"anchor\" id=\"od_load\"></a>\n",
    "\n",
    "The pretrained outlier and adversarial detectors used in the notebook can be found [here](https://console.cloud.google.com/storage/browser/seldon-models/alibi-detect). You can use the built-in ```fetch_detector``` function which saves the pre-trained models in a local directory ```filepath``` and loads the detector. Alternatively, you can train a detector from scratch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_pretrained = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "filepath = os.path.join(os.getcwd(), 'outlier')\n",
    "\n",
    "if load_pretrained:  # load pre-trained detector\n",
    "    detector_type = 'outlier'\n",
    "    dataset = 'cifar10'\n",
    "    detector_name = 'OutlierVAE'\n",
    "    od = fetch_detector(filepath, detector_type, dataset, detector_name)\n",
    "    filepath = os.path.join(filepath, detector_name)\n",
    "else:  # define model, initialize, train and save outlier detector\n",
    "    \n",
    "    # define encoder and decoder networks\n",
    "    latent_dim = 1024\n",
    "    encoder_net = tf.keras.Sequential(\n",
    "      [\n",
    "          InputLayer(input_shape=(32, 32, 3)),\n",
    "          Conv2D(64, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "          Conv2D(128, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "          Conv2D(512, 4, strides=2, padding='same', activation=tf.nn.relu)\n",
    "      ]\n",
    "    )\n",
    "\n",
    "    decoder_net = tf.keras.Sequential(\n",
    "      [\n",
    "          InputLayer(input_shape=(latent_dim,)),\n",
    "          Dense(4*4*128),\n",
    "          Reshape(target_shape=(4, 4, 128)),\n",
    "          Conv2DTranspose(256, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "          Conv2DTranspose(64, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "          Conv2DTranspose(3, 4, strides=2, padding='same', activation='sigmoid')\n",
    "      ]\n",
    "    )\n",
    "    \n",
    "    # initialize outlier detector\n",
    "    od = OutlierVAE(\n",
    "        threshold=.015,  # threshold for outlier score\n",
    "        encoder_net=encoder_net,  # can also pass VAE model instead\n",
    "        decoder_net=decoder_net,  # of separate encoder and decoder\n",
    "        latent_dim=latent_dim\n",
    "    )\n",
    "    \n",
    "    # train\n",
    "    od.fit(X_train, epochs=50, verbose=False)\n",
    "    \n",
    "    # save the trained outlier detector\n",
    "    save_detector(od, filepath)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check whether the model manages to reconstruct the in-distribution training data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot original and reconstructed instance\n",
    "idx = 8\n",
    "X = X_train[idx].reshape(1, 32, 32, 3)\n",
    "X_recon = od.vae(X)\n",
    "plt.imshow(X.reshape(32, 32, 3)); plt.axis('off'); plt.show()\n",
    "plt.imshow(X_recon.numpy().reshape(32, 32, 3)); plt.axis('off'); plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Setting the threshold <a class=\"anchor\" id=\"od_threshold\"></a>\n",
    "\n",
    "Finding good threshold values can be tricky since they are typically not easy to interpret. The `infer_threshold` method helps finding a sensible value. We need to pass a batch of instances `X` and specify what percentage of those we consider to be normal via `threshold_perc`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current threshold: 0.015\n",
      "New threshold: 0.00969364859163762\n"
     ]
    }
   ],
   "source": [
    "print('Current threshold: {}'.format(od.threshold))\n",
    "od.infer_threshold(X_train, threshold_perc=99, batch_size=128)  # assume 1% of the training data are outliers\n",
    "print('New threshold: {}'.format(od.threshold))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create and detect outliers <a class=\"anchor\" id=\"od_detect\"></a>\n",
    "\n",
    "We can create some outliers by applying a random noise mask to the original instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "\n",
    "i = 1\n",
    "\n",
    "# create masked instance\n",
    "x = X_test[i].reshape(1, 32, 32, 3)\n",
    "x_mask, mask = apply_mask(\n",
    "    x,\n",
    "    mask_size=(8,8),\n",
    "    n_masks=1,\n",
    "    channels=[0,1,2],\n",
    "    mask_type='normal',\n",
    "    noise_distr=(0,1),\n",
    "    clip_rng=(0,1)\n",
    ")\n",
    "\n",
    "# predict outliers and reconstructions\n",
    "sample = np.concatenate([x_mask, x])\n",
    "preds = od.predict(sample)\n",
    "x_recon = od.vae(sample).numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Is original outlier? No!\n",
      "Is perturbed outlier? Yes!\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# check if outlier and visualize outlier scores\n",
    "labels = ['No!', 'Yes!']\n",
    "print(f\"Is original outlier? {labels[preds['data']['is_outlier'][1]]}\")\n",
    "print(f\"Is perturbed outlier? {labels[preds['data']['is_outlier'][0]]}\")\n",
    "plot_feature_outlier_image(preds, sample, x_recon, max_instances=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Deploy the detector <a class=\"anchor\" id=\"od_deploy\"></a>\n",
    "\n",
    "For this example we use the open source deployment platform [Seldon Core](https://docs.seldon.io/projects/seldon-core/en/v1.1.0/) and eventing based project [Knative](https://github.com/knative) which allows serverless components\n",
    "to be connected to event streams. The Seldon Core payload logger sends events containing model requests to the Knative broker which can farm these out to serverless components such as the outlier, drift or adversarial detection modules. Further eventing components can be added to feed off events produced by these components to send onwards to, for example, alerting or storage modules. This happens asynchronously.\n",
    "\n",
    "![deploy_diagram](image/deploy_diagram.png)\n",
    "\n",
    "We already configured a cluster on DigitalOcean with Seldon Core installed. The configuration steps to set everything up from scratch are detailed in [this example notebook](https://docs.seldon.io/projects/seldon-core/en/latest/examples/outlier_cifar10.html).\n",
    "\n",
    "First we get the IP address of the Istio Ingress Gateway. This assumes Istio is installed with a LoadBalancer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "188.166.139.197\n"
     ]
    }
   ],
   "source": [
    "CLUSTER_IPS=!(kubectl -n istio-system get service istio-ingressgateway -o jsonpath='{.status.loadBalancer.ingress[0].ip}')\n",
    "CLUSTER_IP=CLUSTER_IPS[0]\n",
    "print(CLUSTER_IP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "vae-outlier.default.example.com\n"
     ]
    }
   ],
   "source": [
    "SERVICE_HOSTNAMES=!(kubectl get ksvc vae-outlier -o jsonpath='{.status.url}' | cut -d \"/\" -f 3)\n",
    "SERVICE_HOSTNAME_VAEOD=SERVICE_HOSTNAMES[0]\n",
    "print(SERVICE_HOSTNAME_VAEOD)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We define some utility functions for the prediction of the deployed model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "import json\n",
    "import requests\n",
    "from typing import Union\n",
    "\n",
    "classes = ('plane', 'car', 'bird', 'cat', 'deer', \n",
    "           'dog', 'frog', 'horse', 'ship', 'truck')\n",
    "\n",
    "def predict(x: np.ndarray) -> Union[str, list]:\n",
    "    \"\"\" Model prediction. \"\"\"\n",
    "    formData = {\n",
    "    'instances': x.tolist()\n",
    "    }\n",
    "    headers = {}\n",
    "    res = requests.post(\n",
    "        'http://'+CLUSTER_IP+'/seldon/default/tfserving-cifar10/v1/models/resnet32/:predict', \n",
    "        json=formData, \n",
    "        headers=headers\n",
    "    )\n",
    "    if res.status_code == 200:\n",
    "        return classes[np.array(res.json()[\"predictions\"])[0].argmax()]\n",
    "    else:\n",
    "        print(\"Failed with \",res.status_code)\n",
    "        return []\n",
    "    \n",
    "    \n",
    "def outlier(x: np.ndarray) -> Union[dict, list]:\n",
    "    \"\"\" Outlier prediction. \"\"\"\n",
    "    formData = {\n",
    "    'instances': x.tolist()\n",
    "    }\n",
    "    headers = {\n",
    "        \"Alibi-Detect-Return-Feature-Score\": \"true\",\n",
    "        \"Alibi-Detect-Return-Instance-Score\": \"true\"\n",
    "    }\n",
    "    headers[\"Host\"] = SERVICE_HOSTNAME_VAEOD\n",
    "    res = requests.post('http://'+CLUSTER_IP+'/', json=formData, headers=headers)\n",
    "    if res.status_code == 200:\n",
    "        od = res.json()\n",
    "        od[\"data\"][\"feature_score\"] = np.array(od[\"data\"][\"feature_score\"])\n",
    "        od[\"data\"][\"instance_score\"] = np.array(od[\"data\"][\"instance_score\"])\n",
    "        return od\n",
    "    else:\n",
    "        print(\"Failed with \",res.status_code)\n",
    "        return []\n",
    "    \n",
    "    \n",
    "def show(x: np.ndarray) -> None:\n",
    "    plt.imshow(x.reshape(32, 32, 3))\n",
    "    plt.axis('off')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's make a prediction on the original instance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'ship'"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show(x)\n",
    "predict(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check the message dumper for the output of the outlier detector:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outlier? No!\n"
     ]
    }
   ],
   "source": [
    "res=!kubectl logs $(kubectl get pod -l serving.knative.dev/configuration=message-dumper -o jsonpath='{.items[0].metadata.name}') user-container\n",
    "data = []\n",
    "for i in range(0,len(res)):\n",
    "    if res[i] == 'Data,':\n",
    "        data.append(res[i+1])\n",
    "j = json.loads(json.loads(data[0]))\n",
    "print(\"Outlier?\",labels[j[\"data\"][\"is_outlier\"]==[1]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We then make a prediction on the perturbed instance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'ship'"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "show(x_mask)\n",
    "predict(x_mask)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although the prediction is still correct, the instance is clearly an outlier:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Outlier? Yes!\n"
     ]
    }
   ],
   "source": [
    "res=!kubectl logs $(kubectl get pod -l serving.knative.dev/configuration=message-dumper -o jsonpath='{.items[0].metadata.name}') user-container\n",
    "data= []\n",
    "for i in range(0,len(res)):\n",
    "    if res[i] == 'Data,':\n",
    "        data.append(res[i+1])\n",
    "j = json.loads(json.loads(data[1]))\n",
    "print(\"Outlier?\",labels[j[\"data\"][\"is_outlier\"]==[1]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "preds = outlier(x_mask)\n",
    "plot_feature_outlier_image(preds, x_mask, X_recon=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Adversarial detection by matching prediction probabilities <a class=\"anchor\" id=\"ad\"></a>\n",
    "\n",
    "### Method <a class=\"anchor\" id=\"ad_method\"></a>\n",
    "\n",
    "The adversarial detector is based on [Adversarial Detection and Correction by Matching Prediction Distributions](https://arxiv.org/abs/2002.09364). Usually, autoencoders are trained to find a transformation $T$ that reconstructs the input instance $x$ as accurately as possible with loss functions that are suited to capture the similarities between x and $x'$ such as the mean squared reconstruction error. The novelty of the adversarial autoencoder (AE) detector relies on the use of a classification model-dependent loss function based on a distance metric in the output space of the model to train the autoencoder network. Given a classification model $M$ we optimise the weights of the autoencoder such that the [KL-divergence](https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence) between the model predictions on $x$ and on $x'$ is minimised. Without the presence of a reconstruction loss term $x'$ simply tries to make sure that the prediction probabilities $M(x')$ and $M(x)$ match without caring about the proximity of $x'$ to $x$. As a result, $x'$ is allowed to live in different areas of the input feature space than $x$ with different decision boundary shapes with respect to the model $M$. The carefully crafted adversarial perturbation which is effective around x does not transfer to the new location of $x'$ in the feature space, and the attack is therefore neutralised. Training of the autoencoder is unsupervised since we only need access to the model prediction probabilities and the normal training instances. We do not require any knowledge about the underlying adversarial attack and the classifier weights are frozen during training. \n",
    "\n",
    "The detector can be used as follows:\n",
    "\n",
    "* An adversarial score $S$ is computed. $S$ equals the K-L divergence between the model predictions on $x$ and $x'$.\n",
    "\n",
    "* If $S$ is above a threshold (explicitly defined or inferred from training data), the instance is flagged as adversarial.\n",
    "\n",
    "* For adversarial instances, the model $M$ uses the reconstructed instance $x'$ to make a prediction. If the adversarial score is below the threshold, the model makes a prediction on the original instance $x$.\n",
    "\n",
    "This procedure is illustrated in the diagram below:\n",
    "\n",
    "![adversarialae](image/adversarialae.png)\n",
    "\n",
    "The method is very flexible and can also be used to detect common data corruptions and perturbations which negatively impact the model performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [],
   "source": [
    "# more imports\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "from alibi_detect.ad import AdversarialAE\n",
    "from alibi_detect.datasets import fetch_attack\n",
    "from alibi_detect.utils.fetching import fetch_tf_model\n",
    "from alibi_detect.utils.prediction import predict_batch"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Utility functions <a class=\"anchor\" id=\"ad_utils\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [],
   "source": [
    "# instance scaling and plotting utility functions\n",
    "def scale_by_instance(X: np.ndarray) -> np.ndarray:\n",
    "    mean_ = X.mean(axis=(1, 2, 3)).reshape(-1, 1, 1, 1)\n",
    "    std_ = X.std(axis=(1, 2, 3)).reshape(-1, 1, 1, 1)\n",
    "    return (X - mean_) / std_, mean_, std_\n",
    "\n",
    "\n",
    "def accuracy(y_true: np.ndarray, y_pred: np.ndarray) -> float:\n",
    "    return (y_true == y_pred).astype(int).sum() / y_true.shape[0]\n",
    "\n",
    "\n",
    "def plot_adversarial(idx: list,\n",
    "                     X: np.ndarray,\n",
    "                     y: np.ndarray,\n",
    "                     X_adv: np.ndarray, \n",
    "                     y_adv: np.ndarray,\n",
    "                     mean: np.ndarray, \n",
    "                     std: np.ndarray, \n",
    "                     score_x: np.ndarray = None,\n",
    "                     score_x_adv: np.ndarray = None,\n",
    "                     X_recon: np.ndarray = None,\n",
    "                     y_recon: np.ndarray = None,\n",
    "                     figsize: tuple = (10, 5)) -> None:\n",
    "    \n",
    "    # category map from class numbers to names\n",
    "    cifar10_map = {0: 'airplane', 1: 'automobile', 2: 'bird', 3: 'cat', 4: 'deer', 5: 'dog',\n",
    "                   6: 'frog', 7: 'horse', 8: 'ship', 9: 'truck'}\n",
    "    \n",
    "    nrows = len(idx)\n",
    "    ncols = 3 if isinstance(X_recon, np.ndarray) else 2\n",
    "    fig, ax = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize)\n",
    "    \n",
    "    n_subplot = 1\n",
    "    for i in idx:\n",
    "        \n",
    "        # rescale images in [0, 1]\n",
    "        X_adj = (X[i] * std[i] + mean[i]) / 255\n",
    "        X_adv_adj = (X_adv[i] * std[i] + mean[i]) / 255\n",
    "        if isinstance(X_recon, np.ndarray):\n",
    "            X_recon_adj = (X_recon[i] * std[i] + mean[i]) / 255\n",
    "        \n",
    "        # original image\n",
    "        plt.subplot(nrows, ncols, n_subplot)\n",
    "        plt.axis('off')\n",
    "        if i == idx[0]:\n",
    "            if isinstance(score_x, np.ndarray):\n",
    "                plt.title('CIFAR-10 Image \\n{}: {:.3f}'.format(cifar10_map[y[i]], score_x[i]))\n",
    "            else:\n",
    "                plt.title('CIFAR-10 Image \\n{}'.format(cifar10_map[y[i]]))\n",
    "        else:\n",
    "            if isinstance(score_x, np.ndarray):\n",
    "                plt.title('{}: {:.3f}'.format(cifar10_map[y[i]], score_x[i]))\n",
    "            else:\n",
    "                plt.title('{}'.format(cifar10_map[y[i]]))\n",
    "        plt.imshow(X_adj)\n",
    "        n_subplot += 1\n",
    "        \n",
    "        # adversarial image\n",
    "        plt.subplot(nrows, ncols, n_subplot)\n",
    "        plt.axis('off')\n",
    "        if i == idx[0]:\n",
    "            if isinstance(score_x_adv, np.ndarray):\n",
    "                plt.title('Adversarial \\n{}: {:.3f}'.format(cifar10_map[y_adv[i]], score_x_adv[i]))\n",
    "            else:\n",
    "                plt.title('Adversarial \\n{}'.format(cifar10_map[y_adv[i]]))\n",
    "        else:\n",
    "            if isinstance(score_x_adv, np.ndarray):\n",
    "                plt.title('{}: {:.3f}'.format(cifar10_map[y_adv[i]], score_x_adv[i]))\n",
    "            else:\n",
    "                plt.title('{}'.format(cifar10_map[y_adv[i]]))\n",
    "        plt.imshow(X_adv_adj)\n",
    "        n_subplot += 1\n",
    "     \n",
    "        # reconstructed image\n",
    "        if isinstance(X_recon, np.ndarray):\n",
    "            plt.subplot(nrows, ncols, n_subplot)\n",
    "            plt.axis('off')\n",
    "            if i == idx[0]:\n",
    "                plt.title('AE Reconstruction \\n{}'.format(cifar10_map[y_recon[i]]))\n",
    "            else:\n",
    "                plt.title('{}'.format(cifar10_map[y_recon[i]]))\n",
    "            plt.imshow(X_recon_adj)\n",
    "            n_subplot += 1\n",
    "    \n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "def plot_roc(roc_data: dict, figsize: tuple = (10,5)):\n",
    "    plot_labels = []\n",
    "    scores_attacks = []\n",
    "    labels_attacks = []\n",
    "    for k, v in roc_data.items():\n",
    "        if 'original' in k:\n",
    "            continue\n",
    "        score_x = roc_data[v['normal']]['scores']\n",
    "        y_pred = roc_data[v['normal']]['predictions']\n",
    "        score_v = v['scores']\n",
    "        y_pred_v = v['predictions']\n",
    "        labels_v = np.ones(score_x.shape[0])\n",
    "        idx_remove = np.where(y_pred == y_pred_v)[0]\n",
    "        labels_v = np.delete(labels_v, idx_remove)\n",
    "        score_v = np.delete(score_v, idx_remove)\n",
    "        scores = np.concatenate([score_x, score_v])\n",
    "        labels = np.concatenate([np.zeros(y_pred.shape[0]), labels_v]).astype(int)\n",
    "        scores_attacks.append(scores)\n",
    "        labels_attacks.append(labels)\n",
    "        plot_labels.append(k)\n",
    "    \n",
    "    for sc_att, la_att, plt_la in zip(scores_attacks, labels_attacks, plot_labels):\n",
    "        fpr, tpr, thresholds = roc_curve(la_att, sc_att)\n",
    "        roc_auc = auc(fpr, tpr)\n",
    "        label = str('{}: AUC = {:.2f}'.format(plt_la, roc_auc))\n",
    "        plt.plot(fpr, tpr, lw=1, label='{}: AUC={:.4f}'.format(plt_la, roc_auc))\n",
    "\n",
    "    plt.plot([0, 1], [0, 1], color='black', lw=1, linestyle='--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.05])\n",
    "    plt.xlabel('False Positive Rate')\n",
    "    plt.ylabel('True Positive Rate')\n",
    "    plt.title('{}'.format('ROC curve'))\n",
    "    plt.legend(loc=\"lower right\", ncol=1)\n",
    "    plt.grid()\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Rescale data <a class=\"anchor\" id=\"ad_rescale\"></a>\n",
    "\n",
    "The ResNet classification model is trained on data standardized by instance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [],
   "source": [
    "# rescale data\n",
    "X_train, mean_train, std_train = scale_by_instance(X_train * 255.)\n",
    "X_test, mean_test, std_test = scale_by_instance(X_test * 255.)\n",
    "scale = (mean_train, std_train), (mean_test, std_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load pre-trained classifier <a class=\"anchor\" id=\"ad_load_model\"></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/avl/anaconda3/envs/detect/lib/python3.7/site-packages/tensorflow_core/python/keras/layers/core.py:901: UserWarning:\n",
      "\n",
      "official.vision.image_classification.resnet_cifar_model is not loaded, but a Lambda layer uses it. It may cause errors.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = 'cifar10'\n",
    "model = 'resnet56'\n",
    "clf = fetch_tf_model(dataset, model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Check the predictions on the test:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9315\n"
     ]
    }
   ],
   "source": [
    "y_pred = predict_batch(clf, X_test, batch_size=32, return_class=True)\n",
    "acc_y_pred = accuracy(y_test, y_pred)\n",
    "print('Accuracy: {:.4f}'.format(acc_y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adversarial attack <a class=\"anchor\" id=\"ad_attack\"></a>\n",
    "\n",
    "We investigate both [Carlini-Wagner (C&W)](https://arxiv.org/abs/1608.04644) and [SLIDE](https://arxiv.org/abs/1904.13000) attacks. You can simply load previously found adversarial instances on the pretrained ResNet-56 model. The attacks are generated by using [Foolbox](https://github.com/bethgelab/foolbox):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "# C&W attack\n",
    "data_cw = fetch_attack(dataset, model, 'cw')\n",
    "X_train_cw, X_test_cw = data_cw['data_train'], data_cw['data_test']\n",
    "meta_cw = data_cw['meta'] # metadata with hyperparameters of the attack\n",
    "# SLIDE attack\n",
    "data_slide = fetch_attack(dataset, model, 'slide')\n",
    "X_train_slide, X_test_slide = data_slide['data_train'], data_slide['data_test']\n",
    "meta_slide = data_slide['meta']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can verify that the accuracy of the classifier drops to almost $0$%:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: cw 0.0000 -- SLIDE 0.0001\n"
     ]
    }
   ],
   "source": [
    "y_pred_cw = predict_batch(clf, X_test_cw, batch_size=32, return_class=True)\n",
    "y_pred_slide = predict_batch(clf, X_test_slide, batch_size=32, return_class=True)\n",
    "acc_y_pred_cw = accuracy(y_test, y_pred_cw)\n",
    "acc_y_pred_slide = accuracy(y_test, y_pred_slide)\n",
    "print('Accuracy: cw {:.4f} -- SLIDE {:.4f}'.format(acc_y_pred_cw, acc_y_pred_slide))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's visualise some adversarial instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "C&W attack...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SLIDE attack...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot attacked instances\n",
    "idx = [3, 4]\n",
    "print('C&W attack...')\n",
    "plot_adversarial(idx, X_test, y_pred, X_test_cw, y_pred_cw, \n",
    "                 mean_test, std_test, figsize=(10, 10))\n",
    "print('SLIDE attack...')\n",
    "plot_adversarial(idx, X_test, y_pred, X_test_slide, y_pred_slide, \n",
    "                 mean_test, std_test, figsize=(10, 10))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load or train and evaluate the adversarial detector <a class=\"anchor\" id=\"ad_load_detector\"></a>\n",
    "\n",
    "We can again either fetch the pretrained detector from a [Google Cloud Bucket](https://console.cloud.google.com/storage/browser/seldon-models/alibi-detect/ad/cifar10/resnet56) or train one from scratch:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "load_pretrained = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:alibi_detect.ad.adversarialae:No threshold level set. Need to infer threshold using `infer_threshold`.\n"
     ]
    }
   ],
   "source": [
    "filepath = os.path.join(os.getcwd(), 'adversarial')\n",
    "\n",
    "if load_pretrained:\n",
    "    detector_type = 'adversarial'\n",
    "    detector_name = 'base'\n",
    "    ad = fetch_detector(filepath, detector_type, dataset, detector_name, model=model)\n",
    "    filepath = os.path.join(filepath, detector_name)\n",
    "else:  # train detector from scratch\n",
    "    \n",
    "    # define encoder and decoder networks\n",
    "    encoder_net = tf.keras.Sequential(\n",
    "            [\n",
    "                InputLayer(input_shape=(32, 32, 3)),\n",
    "                Conv2D(32, 4, strides=2, padding='same', \n",
    "                       activation=tf.nn.relu, kernel_regularizer=l1(1e-5)),\n",
    "                Conv2D(64, 4, strides=2, padding='same', \n",
    "                       activation=tf.nn.relu, kernel_regularizer=l1(1e-5)),\n",
    "                Conv2D(256, 4, strides=2, padding='same', \n",
    "                       activation=tf.nn.relu, kernel_regularizer=l1(1e-5)),\n",
    "                Flatten(),\n",
    "                Dense(40)\n",
    "            ]\n",
    "        )\n",
    "    \n",
    "    decoder_net = tf.keras.Sequential(\n",
    "        [\n",
    "                InputLayer(input_shape=(40,)),\n",
    "                Dense(4 * 4 * 128, activation=tf.nn.relu),\n",
    "                Reshape(target_shape=(4, 4, 128)),\n",
    "                Conv2DTranspose(256, 4, strides=2, padding='same', \n",
    "                                activation=tf.nn.relu, kernel_regularizer=l1(1e-5)),\n",
    "                Conv2DTranspose(64, 4, strides=2, padding='same', \n",
    "                                activation=tf.nn.relu, kernel_regularizer=l1(1e-5)),\n",
    "                Conv2DTranspose(3, 4, strides=2, padding='same', \n",
    "                                activation=None, kernel_regularizer=l1(1e-5))\n",
    "            ]\n",
    "        )\n",
    "    \n",
    "    # initialise and train detector\n",
    "    ad = AdversarialAE(\n",
    "        encoder_net=encoder_net, \n",
    "        decoder_net=decoder_net, \n",
    "        model=clf\n",
    "    )\n",
    "    ad.fit(X_train, epochs=40, batch_size=64, verbose=True)\n",
    "    \n",
    "    # save the trained adversarial detector\n",
    "    save_detector(ad, filepath)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The detector first reconstructs the input instances which can be adversarial. The reconstructed input is then fed to the classifier to compute the adversarial score. If the score is above a threshold, the instance is classified as adversarial and the detector tries to correct the attack. Let's investigate what happens when we reconstruct attacked instances and make predictions on them:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_recon_cw = predict_batch(ad.ae, X_test_cw, batch_size=32)\n",
    "X_recon_slide = predict_batch(ad.ae, X_test_slide, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_recon_cw = predict_batch(clf, X_recon_cw, batch_size=32, return_class=True)\n",
    "y_recon_slide = predict_batch(clf, X_recon_slide, batch_size=32, return_class=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Accuracy on attacked vs. reconstructed instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy after C&W attack 0.0000 -- reconstruction 0.8048\n",
      "Accuracy after SLIDE attack 0.0001 -- reconstruction 0.8159\n"
     ]
    }
   ],
   "source": [
    "acc_y_recon_cw = accuracy(y_test, y_recon_cw)\n",
    "acc_y_recon_slide = accuracy(y_test, y_recon_slide)\n",
    "print('Accuracy after C&W attack {:.4f} -- reconstruction {:.4f}'.format(acc_y_pred_cw, acc_y_recon_cw))\n",
    "print('Accuracy after SLIDE attack {:.4f} -- reconstruction {:.4f}'.format(acc_y_pred_slide, acc_y_recon_slide))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The detector restores the accuracy after the attacks from almost $0$% to well over $80$%! We can compute the adversarial scores and inspect some of the reconstructed instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "score_x = ad.score(X_test, batch_size=32)\n",
    "score_cw = ad.score(X_test_cw, batch_size=32)\n",
    "score_slide = ad.score(X_test_slide, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "code_folding": [
     0
    ],
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "C&W attack...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1080 with 15 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "WARNING:matplotlib.image:Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SLIDE attack...\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x1080 with 15 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualize original, attacked and reconstructed instances with adversarial scores\n",
    "print('C&W attack...')\n",
    "idx = [10, 13, 14, 16, 17]\n",
    "plot_adversarial(idx, X_test, y_pred, X_test_cw, y_pred_cw, mean_test, std_test, \n",
    "                 score_x=score_x, score_x_adv=score_cw, X_recon=X_recon_cw, \n",
    "                 y_recon=y_recon_cw, figsize=(10, 15))\n",
    "print('SLIDE attack...')\n",
    "idx = [23, 25, 27, 29, 34]\n",
    "plot_adversarial(idx, X_test, y_pred, X_test_slide, y_pred_slide, mean_test, std_test, \n",
    "                 score_x=score_x, score_x_adv=score_slide, X_recon=X_recon_slide, \n",
    "                 y_recon=y_recon_slide, figsize=(10, 15))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The ROC curves and AUC values show the effectiveness of the adversarial score to detect adversarial instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot roc curve\n",
    "roc_data = {\n",
    "    'original': {'scores': score_x, 'predictions': y_pred},\n",
    "    'C&W': {'scores': score_cw, 'predictions': y_pred_cw, 'normal': 'original'},\n",
    "    'SLIDE': {'scores': score_slide, 'predictions': y_pred_slide, 'normal': 'original'}\n",
    "}\n",
    "\n",
    "plot_roc(roc_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The threshold for the adversarial score can be set via ```infer_threshold```. We need to pass a batch of instances $X$ and specify what percentage of those we consider to be normal via `threshold_perc`. Assume we have only normal instances some of which the model has misclassified leading to a higher score if the reconstruction picked up features from the correct class or some might look adversarial in the first place. As a result, we set our threshold at $95$%:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adversarial threshold: 2.6722\n"
     ]
    }
   ],
   "source": [
    "ad.infer_threshold(X_test, threshold_perc=95, margin=0., batch_size=32)\n",
    "print('Adversarial threshold: {:.4f}'.format(ad.threshold))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `correct` method of the detector executes the diagram in Figure 1. First the adversarial scores is computed. For instances where the score is above the threshold, the classifier prediction on the reconstructed instance is returned. Otherwise the original prediction is kept. The method returns a dictionary containing the metadata of the detector, whether the instances in the batch are adversarial (above the threshold) or not, the classifier predictions using the correction mechanism and both the original and reconstructed predictions. Let's illustrate this on a batch containing some adversarial (C&W) and original test set instances:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2000, 32, 32, 3) (2000,)\n"
     ]
    }
   ],
   "source": [
    "n_test = X_test.shape[0]\n",
    "np.random.seed(0)\n",
    "idx_normal = np.random.choice(n_test, size=1600, replace=False)\n",
    "idx_cw = np.random.choice(n_test, size=400, replace=False)\n",
    "\n",
    "X_mix = np.concatenate([X_test[idx_normal], X_test_cw[idx_cw]])\n",
    "y_mix = np.concatenate([y_test[idx_normal], y_test[idx_cw]])\n",
    "print(X_mix.shape, y_mix.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check the model performance:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy 0.7380\n"
     ]
    }
   ],
   "source": [
    "y_pred_mix = predict_batch(clf, X_mix, batch_size=32, return_class=True)\n",
    "acc_y_pred_mix = accuracy(y_mix, y_pred_mix)\n",
    "print('Accuracy {:.4f}'.format(acc_y_pred_mix))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This can be improved with the correction mechanism:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy 0.8205\n"
     ]
    }
   ],
   "source": [
    "preds = ad.correct(X_mix, batch_size=32)\n",
    "acc_y_corr_mix = accuracy(y_mix, preds['data']['corrected'])\n",
    "print('Accuracy {:.4f}'.format(acc_y_corr_mix))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are a few other tricks highlighted in the [paper](https://arxiv.org/abs/2002.09364) (**temperature scaling** and **hidden layer K-L divergence**) and implemented in Alibi Detect which can further boost the adversarial detector's performance. Check [this example notebook](https://docs.seldon.io/projects/alibi-detect/en/latest/examples/ad_ae_cifar10.html) for more details."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Drift detection with Kolmogorov-Smirnov <a class=\"anchor\" id=\"cd\"></a>\n",
    "\n",
    "### Method <a class=\"anchor\" id=\"cd_method\"></a>\n",
    "\n",
    "The drift detector applies feature-wise two-sample [Kolmogorov-Smirnov](https://en.wikipedia.org/wiki/Kolmogorov%E2%80%93Smirnov_test) (K-S) tests. For multivariate data, the obtained p-values for each feature are aggregated either via the [Bonferroni](https://mathworld.wolfram.com/BonferroniCorrection.html) or the [False Discovery Rate](http://www.math.tau.ac.il/~ybenja/MyPapers/benjamini_hochberg1995.pdf) (FDR) correction. The Bonferroni correction is more conservative and controls for the probability of at least one false positive. The FDR correction on the other hand allows for an expected fraction of false positives to occur.\n",
    "\n",
    "For high-dimensional data, we typically want to reduce the dimensionality before computing the feature-wise univariate K-S tests and aggregating those via the chosen correction method. Following suggestions in [Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift](https://arxiv.org/abs/1810.11953), we incorporate Untrained AutoEncoders (UAE), black-box shift detection using the classifier's softmax outputs ([BBSDs](https://arxiv.org/abs/1802.03916)) and [PCA](https://en.wikipedia.org/wiki/Principal_component_analysis) as out-of-the box preprocessing methods. Preprocessing methods which do not rely on the classifier will usually pick up drift in the input data, while BBSDs focuses on label shift. The [adversarial detector](https://arxiv.org/abs/2002.09364) which is part of the library can also be transformed into a drift detector picking up drift that reduces the performance of the classification model. We can therefore combine different preprocessing techniques to figure out if there is drift which hurts the model performance, and whether this drift can be classified as input drift or label shift.\n",
    "\n",
    "Note that the library also has a drift detector based on the [Maximum Mean Discrepancy](https://docs.seldon.io/projects/alibi-detect/en/latest/examples/cd_mmd_cifar10.html#) and contains [drift on text functionality](https://docs.seldon.io/projects/alibi-detect/en/latest/examples/cd_text_imdb.html) as well.\n",
    "\n",
    "\n",
    "### Dataset <a class=\"anchor\" id=\"cd_dataset\"></a>\n",
    "\n",
    "We will use the CIFAR-10-C dataset ([Hendrycks & Dietterich, 2019](https://arxiv.org/abs/1903.12261)) to evaluate the drift detector. The instances in\n",
    "CIFAR-10-C come from the test set in CIFAR-10 but have been corrupted and perturbed by various types of noise, blur, brightness etc. at different levels of severity, leading to a gradual decline in the classification model performance. We also check for drift against the original test set with class imbalances."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [],
   "source": [
    "# yet again import stuff\n",
    "from alibi_detect.cd import KSDrift\n",
    "from alibi_detect.cd.preprocess import UAE, HiddenOutput\n",
    "from alibi_detect.datasets import fetch_cifar10c, corruption_types_cifar10c"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can select from the following corruption types at 5 severity levels:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['brightness', 'contrast', 'defocus_blur', 'elastic_transform', 'fog', 'frost', 'gaussian_blur', 'gaussian_noise', 'glass_blur', 'impulse_noise', 'jpeg_compression', 'motion_blur', 'pixelate', 'saturate', 'shot_noise', 'snow', 'spatter', 'speckle_noise', 'zoom_blur']\n"
     ]
    }
   ],
   "source": [
    "corruptions = corruption_types_cifar10c()\n",
    "print(corruptions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's pick a subset of the corruptions at corruption level 5. Each corruption type consists of perturbations on all of the original test set images."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "corruption = ['gaussian_noise', 'motion_blur', 'brightness', 'pixelate']\n",
    "X_corr, y_corr = fetch_cifar10c(corruption=corruption, severity=5, return_X_y=True)\n",
    "X_corr = X_corr.astype('float32') / 255"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We split the original test set in a reference dataset and a dataset which should not be rejected under the *H<sub>0</sub>* of the K-S test. We also split the corrupted data by corruption type:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5000, 32, 32, 3) (5000, 32, 32, 3)\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(0)\n",
    "n_test = X_test.shape[0]\n",
    "idx = np.random.choice(n_test, size=n_test // 2, replace=False)\n",
    "idx_h0 = np.delete(np.arange(n_test), idx, axis=0)\n",
    "X_ref,y_ref = X_test[idx], y_test[idx]\n",
    "X_h0, y_h0 = X_test[idx_h0], y_test[idx_h0]\n",
    "print(X_ref.shape, X_h0.shape)\n",
    "\n",
    "X_c = []\n",
    "n_corr = len(corruption)\n",
    "for i in range(n_corr):\n",
    "    X_c.append(scale_by_instance(X_corr[i * n_test:(i + 1) * n_test])[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can visualise the same instance for each corruption type:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAAD3CAYAAADmIkO7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAcU0lEQVR4nO2deZCdV3nmn7f3263e91W9alerJWux5X3BdsAgO3iMGYc4KYaBDMlko1JJpTIJVWFSkwpVEyp/jIcAYzBgx+NAbGJkAV5l7fvaUksttdS7el9vryd/9HWNopzni0mF8RE8vyqK9nn6fPf77v2e/q7Oe973NecchBDhkfRBn4AQwo/MKUSgyJxCBIrMKUSgyJxCBIrMKUSgyJw3CWb2lJnt+qDPg2FmE2ZW/0Gfx88TpjinEGGiJ6cQgSJz/iuY2SYzO2pm42b2opm9YGZ/bmb5ZvYDM7tmZsOJn6uum3fZzB647r//zMyeS/ycYWbPmdmgmY2Y2UEzK01ov2Zm7YnXu2RmT103vvu64/21mV01szEzO2xmd97wWn9nZt9MHOe0mW1+H9d62cy+YGYnzGw0ca0Z1+mfMbMLZjZkZi+bWcV1mjOzxsTPHzazM4nX7jKzL1z3e4+Y2bHEde8xs+af/lP5xUDmjMDM0gB8D8D/AVAA4LsAHkvISQC+AWA5gBoA0wD+5n0e+mkAuQCqARQC+ByAaTPLAvAVAL/knMsGsB3AMXKMgwBaEuf1HQAvXm8kAB8D8DyAPAAv/xTn9gSAhwHUAWgG8GsAYGb3AfiLhF4OoCNxfB9fA/DZxDWsA/B64hgbAXwdwGcT1/0MgJfNLP19ntsvFDJnNLcCSAHwFefcnHPu7wEcAADn3KBz7iXn3JRzbhzAlwDc/T6PO4elm7PRObfgnDvsnBtLaIsA1plZzDnX45w77TuAc+65xDnMO+e+DCAdwMrrfmW3c+5V59wCgG8B2PA+z+0rzrlu59wQgFew9AcAAJ4C8HXn3BHn3AyAPwJwm5nVkutbY2Y5zrlh59yRxPh/BvCMc25/4rqfBTCDpfdZ3IDMGU0FgC73z1fNrgKAmWWa2TNm1mFmYwDeBpBnZsnv47jfAvAagOfNrNvM/tLMUp1zkwA+gaUnaY+Z/aOZrfIdIPH182zi6+cIlp7ERdf9Su91P08ByDCzlPdxbjfOW5b4uQJLT0sAgHNuAsAggErPMT4O4MMAOszsLTO7LTG+HMDvJ77SjiTOuzpxbHEDMmc0PQAqzcyuG6tO/P/vY+lJtc05lwPgrsT4e787CSDzunll7/2QeAp/0Tm3BktfXR8B8KsJ7TXn3Iew9NWxFcBXbzypxL8v/wBLXzHznXN5AEave+2fBd1YMtd755CFpad/142/6Jw76JzbAaAEwPcB/F1CugrgS865vOv+l+mc++7P8LxvWmTOaPYCWADwm2aWYmY7AGxNaNlY+nfmiJkVAPjTG+YeA/CkmaUmFmMef08ws3vNbH3iKTuGpa+Bi2ZWamY7Ejf+DIAJLH3NvZFsAPMArgFIMbP/BiDn3+maGd8F8Otm1pL4N+J/B7DfOXf5+l8ys7RETDbXOTeHpet77xq+CuBzZrbNlsgys4+YWfbP+NxvSmTOCJxzswB+GcCnAYwA+BUAP8CScf4ngBiAAQD7AOy8YfqfAGgAMAzgi1hatHmPMgD/F0s37lkAb2Hpq24SgN/D0lNqCEv/hv0Nz6m9lni981j6qhlH4uv2zwrn3I+xdE0vYekbRQOAJ8mvfwrA5cTX/c9h6d+rcM4dAvAZLC1ODQO4gMSCk/iXaBPCT4mZ7Qfwv5xz3/igz0X8fKMn57+Cmd1tZmWJr7VPYym8cONTUoh/d97P6t0vOiuxtKCRBaAdwOPOuZ4P9pT+bZhZDYAzRF7jnLvy//N8RDT6WitEoOhrrRCBEvm11uECfawO/b+w3b+gsG2Zd7x93wydU79+N9UW5vhrvbOR/32pGvHvCmvs4jHv4ck9VGtvvI9qt5RMUW3u7DDVjqf3ecc3t5XSOUNlY1TLzT1HteTaR6l27pJ/fLKqwy8AWNlXTDWX5t3YtKQtbqFa9/D3veOVx9fROZ1PNlJt1b52qmEFz3A7NsWXFVaUlXjH2xc30TnDaUNUuxMF3vi0npxCBIrMKUSgyJxCBIrMKUSgyJxCBIrMKUSgRIZSrJ8vUVvqBNXamvwhk6amUf5iczyp4kzKWqrFhvlmnZy3BvwCjyhgaHY71W4ZIMcD4GZ56ODiwdv5vDv872O8zpcmmdDSeMpoRim/OLdAJfSSSFDsYj+dcw7+kAIAJBXMcY1HxlCX9rB3fFkFD0WsusQKMgBdZfzznHy1k2r5t/vPAwDm5970jlfF6BRMXizgYoN/WE9OIQJF5hQiUGROIQJF5hQiUGROIQJF5hQiUCJDKfGJa1SbGeHL+fV5qd7xa+fm6Zyrq3imwqY4X0YfnuFL9rGP+5f6r+IknTM1uoJqu6qLqFY5xENB02X8ba7tX+kd713DsylK3SzVdnXy0FJz0iDVMmuXe8ez4/xzSYp474vzeAZSPlWAswv+8MbY+iw6Z9vMbVQbmK+h2obNrVQb57cqunryvOOj9ZN0TlEDP3+GnpxCBIrMKUSgyJxCBIrMKUSgyJxCBEp09b2h/Vwc2EaltvFe73hWMz9cheN1cS5eyKXa3Gpe42ba/HVnWg7zVbqBZP9KMwDEW/hm9Fg7X7k8sMD/Bq5o8K8o53bxFerzmf66QwCwopfvvi4e4Kvvp+72r8q+++M4nVOQwovM31nDEwHGjBQsAnB16j95x+tX8PPIW+TXnPwjfz0rANi1lm98LyviK+x3jPhXoqfzeCLA9ISvq8YSBeVZqiEkxM2EzClEoMicQgSKzClEoMicQgSKzClEoERufB8p4OGSvDY+r67Rv1R+EbwmUX8a1xpifMl7t/Fz3PCGf97QvXzjeBKaqFYNvpm7c/plqj2Yx8+xtcu//F5XzUMpvAkCUHymlmoztbweUM3b/g3zKTO8lcTLg7wm1Lp6/h7HFv0bxwGgtszf5Lp8kIe4ZsbTqNb5S1TC451VXMzmN/jRS/57ZH1uRPjoCr+/Ue4f1pNTiECROYUIFJlTiECROYUIFJlTiECROYUIlOislOFOKg5f4cvQsbEu73jGnVGRG177BvtbqNS/jdfauXDI36W6MfUBOsd1ZvDzqJmm0sL6iNjSGA9+zKT4WzzkJd/J56S/SbXYla1UO9d1kWpTqePe8dxF/2cJALvnNlAtu4P3fni4jod0Sm7zv1cvtfLKQ3fH+GtlRXxmmdd43aeZMd4+YaHR/5m9cbGQzikp4uGerblQVooQNxMypxCBInMKESgypxCBInMKESgypxCBEpmVMj/Kl7zTC7qp1pHhb9XQOLyezjkQUaP/tohOyCU9I1RLbvZ3eR7q5q0kWkt5saiPFvJsimsjJLUAQPEgb4OwtyjTO143RlpNA+jP5m/Wtwf2Ua16koccOrL9oay1s7wFReoBHpp5ufIE1c450soZwO/M+jt9N0zy9z6tihfPOn8hoohXE890KS/hocI98BcNK6o8S+fUz0c1ofC3jNCTU4hAkTmFCBSZU4hAkTmFCBSZU4hAkTmFCJTIUMrLC3xnflZtBdVW4KB3vGOIv1wOeNgGKKVKVwnPOqg87++jUmhr6ZympHNU64z7u1ADQFHyKaq1ZfLCYHNtt/iFzTwEcG0PD/dsBA+XzDbw93FlKenz8TIPiQwU8xDGb2z6FNVOvr6LartW+kMp227h91sb/Bk1AJAxX0e1qbO1VBtN4/10aor92U5Jx3bQOUV3qbO1ED83yJxCBIrMKUSgyJxCBIrMKUSgRNYQGnU/pGJ8ite5nyELUwMneWfl5PV8o3flGF+pG+ZlYBCb8HfLrsrk9Vww4W8HAABTOWeoljKzhmqnB3iZ/sUk/5u1Kp+vyLal8qSDmv3+bt4A0JrHV4Ab0ye94xeLeLfm9km+Ur59nnccn5gnK9QActr970fRyBU6J/Y4jyocG+Pv46Yc/rlciGgdcsCl+493jdezqkrmiRHLCpephpAQNxMypxCBInMKESgypxCBInMKESgypxCBErnx/S8+769tAgBXd3yBai2nb/WOT3fyUET9p5dTLXfucaoNZ8ap1pTkr9tyqHCeztk8xjs5n89ZRbW6oeeoFk/fRLUC598Un3WaJwLM1vM2CKjh3aYz83irifFl/g3iqeM8/PVYLv/MMvv5e3xwjIdgjkz6wxsLA7xTdsXXeH2eVxt5QsLRNfz2z188RrUtpf77IK+EhyXTpv11h6LQk1OIQJE5hQgUmVOIQJE5hQgUmVOIQJE5hQiUyFDK1c1HqZbfyjMBOoa+5x3PXnYHnTO7n/dcKHvCnzEBAAV7edbExIYh7/immX+gc07O/DrVKv0NjQEAVs47Ua8GDznMLpBWDam8XtFUZx/Vpkt5naCWaxFdzPv853G+gWfpnPwr/j4e2uavBQQA8+ZNwgAAJPX6W2UMXuMhrouOh+hm4zwUVNTJwxsDpb9NtdYe//24tYuHHmdaeCZRWqU/60pPTiECReYUIlBkTiECReYUIlBkTiECReYUIlAiC3x949I+Ks7v50WyBu7379pvOc47SleVnaRa0TwP21wa4WGWzEZ/t+aMPt5GYL6Yv9a6GZ4ZcTWDHzN9ineHvrCiyzve18PbCNwR0YJiICI4tvoc7yx+rPa4d7xlkLeuaBvkGR/7uni4pLKSn/+BOX9WzYbUSjrnUB/vbn7/Ph4uuXZPNdXyHe96XRf3X3d1Pi/yhs1RUUsV+BLipkLmFCJQZE4hAkXmFCJQZE4hAkXmFCJQIkMpe9FNxU39PEVjosTf72JiD+/wfGWKL0Nv2+zPLgGAc2+dpdqRk/7l/NKtc3TO2ioeIvp2Ie+gvOeZH/BjTmRSrT3Hn42Ts8B7fCwrphIeKOJhm9JT9VSb/i1/KOjWXN6n5swE7/8x5o9iAQC2ZD9PtZnxbd7xUVzl5zF8F9XWTfJsEPcQL0KWNcrDfgW5/jDL0BUeaiuoqaIaAIVShLiZkDmFCBSZU4hAkTmFCBSZU4hAkTmFCJTIAl97vvQC1Q5f4oWfrpX5+10MxjbSOaP3nqNaZysvkLVxnPfJyF3+une86MGH6Jx97bzdeNL3ebhk4Chv6f5M631U25r9He/4oYm76ZzihRNUS9/+UaolbeBF1P4k6XbveDwiJDIf82eyAMC6at7P5dWRX6Za82F/aOztB/j9tmX5Eapd6eH31cxBf9gGAFIrePjrwuUx7/gDG3gGTOcl/xwAaK7L8Y7rySlEoMicQgSKzClEoMicQgSKzClEoESu1j67z99WAQDSX99KtcxHL3vHWzL4ZuL4UAPVOpt5Sf2eGv9KFwDcUuZfgUxGE51TN9FOteztq6l2cprXo6lYw+vRbO70HzP783wl8enJ/0K19NV8M/emab4x+3ivv3bPkbO5dM7Di7wNwt9m8oSE6gle9+n5cn9H79m/5V25bZTfxmVz/i7rAPDcqv1UGxvjr/exQxe94we/zO+B2DBvr9Fc1+Id15NTiECROYUIFJlTiECROYUIFJlTiECROYUIlMgaQvV5D1BxxWbeLqCs1L80vGbLAp0zeIDXvhks4Ev2Bcn+EAAA3F+2wzt+LK+VzrkvmS+Hx1N5SGSkgF9byT15VCs/7J+XkcOTBAYjSvv3/cM3qJbTxzej93cc9o6/WMpDAD1X9lKtdIK3tZjIOka1nKv+OlNzGz/MX6vlENUm9/BQ0P1zvVRbvvV/U23nxD7veMu6TXTOaGwX1Z5a/YhqCAlxMyFzChEoMqcQgSJzChEoMqcQgSJzChEokVkpl0qHqTbR9ROq/eEf/I13/PfW8horf13h3+kPACWneWuC2j5+CfM5/ho3t2fx68KyZipVVPOl9/XLec2fyXTe9bq/xZ8psraNn+OeC7uptrOX92pYF9FiYC6zxjveXcZrKu1o4eGjpAzevfr8X/HWG/U7/Mc8P9VH52w/x0NEM7f5u6wDAOL+awaAscxnqfZY2dPe8YMVPEQUP8fvfZBkJz05hQgUmVOIQJE5hQgUmVOIQJE5hQgUmVOIQIkMpeByJZWS73+Yak+UP+kdn5oopHMeBC+sNfoQL/t/ppyXzc9c5u/kPJvOz+OONp5NMXwLLzQ23TpBtdEM3i27YdSfNbF/G89k6UAX1TZl+IuaAUDhwAjV5mL+LtV/2sW7eR/MLaHa+n4eknI7+PlfafTfkvVFvDhZ22GewZOeysNfg7M8O2nxFZ6dNJn5Ze94WR0PLfXW888TuMc7qienEIEicwoRKDKnEIEicwoRKDKnEIEicwoRKJGhlJoqvvT+uxkf4wdd7w8P7H7hJJ3TvYN3Ls7v4pkR2wr94RIAyOn3hz4mGnk/kSPzPFzSsetdqr1Zwnu23HV4LdVe7flH7/jmZ3mxqPwGfv4Ny5+n2ivDvJhbQ5q/wNqp5/0FtwCgF/zzHMsboBpqecZNw3n/MX8Ykdkzlcz7smx7ib9XFzP5vTO64SrVqhYf8AtJ/j4vAJCyM041kKiknpxCBIrMKUSgyJxCBIrMKUSgyJxCBErkau1TH/8k1bIf56uTr377W97x/hneGfrSN3lbhXhE5+L1lby2zNCsf9P21Ot8VW0kzjdYb6geo9rlN2qpdsvMO1QrLxz0jr/S763QDwBYnOB1gk7NvkC1LfO85cU7l/wb9y3iM7uYVUa17OP8PZ4b5hvEL63xb9w/NnKBzrm3la+Ejt/KN+DnDbdRrbaglGppzf5zmd55hM5ZsWoz1Rh6cgoRKDKnEIEicwoRKDKnEIEicwoRKDKnEIES2dn6oU/+MRVzunlYYWT8Tr+wvpvOmVyVQbVVb56i2tEFHmaZLvmId7wk/wCdkzGVT7Xm1bxOUI7bQrWBBr6ZvuhHy73jZ36Fh4jyXjlPtQy+Tx1TK85RrS/5Pu94/RwPiSzbzTt9927hm+Kr4rwtxFuv+Y955+p0OufMGa7FinlIZCTGW4BU3s3n1Yz6axZt/zDvwH5wmN+nX7jz8+psLcTNhMwpRKDInEIEiswpRKDInEIEiswpRKBEhlJszb1c7OZZJMj0L9mv7eGHm6zjS9eXp/ytAgAgf2COallNdd7x8aGI+jAxng1SVriCan2zvNx+chFvPzCV6593axZvdTAxxjsoF8zyTs4Tufza2sk5Nl74KJ3z7vkXqdZSyrtXt1ZRCSW9/tpD+e28rUJ383+lWu7qo1RrTL9EtYV1POtq9G1/PaPyNTyjqWktP95vPfhFhVKEuJmQOYUIFJlTiECROYUIFJlTiECROYUIlMgCXw8W8xL4u4p5J2q8PeUdHiu9i06ZXLGPH6+Hd5ue6+OhlM5WstTPoxSY7+AZEwUdV6h2qsR/zUsiD8Gsj/lDUi9s2U/nuF6ewVNfzLWZizwLoybdn2nRuvgTOqc0HtEZ+oHDXGuvpdqVyVu847FNLXTOhsXLVKsc4EXNWup4llF8kN/7x/L8LR4Wq/m9uK7tQaqBSHpyChEoMqcQgSJzChEoMqcQgSJzChEoMqcQgRIZSjk+zbM3/sP8Vqq9mPEj73jdWl6Yajh3PdVW7eN/Q5r+nBeZ6v6S//JOVPJux339PHvgZIxnl6SN8fMov72QasuOZ3rHF7p4vGdlKe8N0jfOK3wNgocVLqf6P7PaeV7wrDaLp5ecH+Uhnea5W6m26iF/5kxKLs+CmpvnRciSeiupdqIzi2rdjfyYTSP+bKfKi2l0zlQzvz8YenIKESgypxCBInMKESgypxCBInMKESiRNYQqsrdQsafqMj/qObLy6o7TKU23+VfAAKCyqIJqw8P8NOov+zd6nwbf8Hz+Gt/cDv/CKgCghpdAwnQn30z/2Sd+xzteVbiJzpmfGqJa2xyv63PlMGmTAWDmmn+jd2kh/8ySH6+n2t3jvD1Fcg9fic4r9a+Wf7vgJTonafer/HiVv0q12NBpqm1vXEe1q2v97R82nOH3afejvBv5J7bcpxpCQtxMyJxCBIrMKUSgyJxCBIrMKUSgyJxCBErkxveUFbzmDCZ4DKPZveEdH/uP/HBz3+HhgbObeEfmyiO8/cD3ivydhlcN8vNASg2Vamp5mGXsaIxqM+Ahh9Gz/uSC5k/xjeNXU3jrh7/s/QTVdq7hLQ02TPg3xb9xgodEeip5Halrr/CaUPsneWfx8VP+zfQPN/H6PPE8fg8ML+O3+JacR6j25jwPfTzYe8Q7nvEo77BdsMDbUzD05BQiUGROIQJF5hQiUGROIQJF5hQiUGROIQIlMpSydfU9VKvu4Evle877l+yLvsNDCgPgYZvKI7w0fhd4t2YM+Je2u1bxOjvo4eGSAd4kGVOIqhHDsx++utdfVym19WE65+ktPIy1M2uaahsiwiK7c77uHU8f5K0kxo3XmPrxu01Uq8jl4arNy/3doftS/W0aAMA+eYFq517jIZjhGA/f1abxeJs1+ENZrSf4vbh5C8+EYujJKUSgyJxCBIrMKUSgyJxCBIrMKUSgyJxCBEpkKGVL0TWq/fBAD9UM/mX0AfCsiMxyfh5dPRFdowt5d+WNg7ne8aNtEcva5TwkMlXeyOct+jMVAAApPIS06cwy73h6zQk6Z3RmM9VqBvj78c7F16n29iZ/24W9mR10TtkUb9UwvuwU1WbLP8PnbfPHq0oKeZG0zGO8utpUwQaqFX7In7UEANVZPNx2sMMfCnqq+jY652hJH9VWk3E9OYUIFJlTiECROYUIFJlTiECROYUIFJlTiECJDKWcSGugWm0WD6W0Z/lDH92T4/x4PbwRCS8jBcQHeVjkaK7/9WKNvHu1u8qzGGbT+DkunuA9Z2JZ/ArYqw328nDJ8dP8fbw4zwuNvRKRHTOW7A873ZXvD/UAQH6cv1eNj36UasX1vEv17AV/j5Kd2fy11i7w92rNjL9jNwDEdvKiW/3reAgpZ7W/W/bRDt5xPGVvRNbSY/5hPTmFCBSZU4hAkTmFCBSZU4hAkTmFCBSZU4hAiQylJM3wrJRn03nr8Jxy/1L/whVeUGmomRemil/gYYqMtDSqZWX5i0WtOjxJ5+xPX6DaYv87VMvnEQwMR7Sr75v2hzAO2iE6p32eF92aK+RhovrsD1EtKdOfNdGdzzMtHpvbSLVX5nh/m7p7+JsVg//8PzTIM0iOt/G+MiXrPk+18TPPUW3+Hh5GXN7pv3/yKo7z483yzBmGnpxCBIrMKUSgyJxCBIrMKUSgyJxCBErkau1kG6/bcs/Gaqr17veXx69Z5Cuyyb18g3XdWv+qKwDsfde/CRkAMuouecffTeG1gDbn85XhlHl/6wQAyOALwBhOv51qDVmXveP77+BdkpPj2VR7PLmWakc38s3jdV/zrwDHt/PP5eitfKP3k8v55vz+b56lWk+pf0V5cno9nZP9ad46ofl/8LpJA4/yY8a7+TELmlZ6x6+08+hG82In1bBtq3dYT04hAkXmFCJQZE4hAkXmFCJQZE4hAkXmFCJQIkMpmSV8Ofzqu7zN88oU/0bkCyv4y3UX8hBG0xjfYP2RR/ZT7czocu/46NZiOif70jaqTS18mWo5Mzy8kbajn2rJp/x/H4vW8ffq4cMfp1qs6CdUyx/lna3zvu5vlZF9lNcd+nEb7+ScMsdDEfGV/lAEAMRH/RvmK1J4N+yUv2+h2sBv7qWavVlFtfuf4GGWU8v89YWKF/jxxuL8mhl6cgoRKDKnEIEicwoRKDKnEIEicwoRKDKnEIFizvEQhhDig0NPTiECReYUIlBkTiECReYUIlBkTiECReYUIlD+CVVUgS/9MitaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot original and corrupted images\n",
    "i = 1\n",
    "\n",
    "n_test = X_test.shape[0]\n",
    "plt.title('Original')\n",
    "plt.axis('off')\n",
    "plt.imshow((X_test[i] * std_test[i] + mean_test[i]) / 255.)\n",
    "plt.show()\n",
    "for _ in range(len(corruption)):\n",
    "    plt.title(corruption[_])\n",
    "    plt.axis('off')\n",
    "    plt.imshow(X_corr[n_test * _+ i])\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also verify that the performance of a ResNet-32 classification model on CIFAR-10 drops significantly on this perturbed dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set accuracy:\n",
      "Original 0.9278\n",
      "gaussian_noise 0.2208\n",
      "motion_blur 0.6339\n",
      "brightness 0.8913\n",
      "pixelate 0.3666\n"
     ]
    }
   ],
   "source": [
    "dataset = 'cifar10'\n",
    "model = 'resnet32'\n",
    "clf = fetch_tf_model(dataset, model)\n",
    "acc = clf.evaluate(X_test, y_test, batch_size=128, verbose=0)[1]\n",
    "print('Test set accuracy:')\n",
    "print('Original {:.4f}'.format(acc))\n",
    "clf_accuracy = {'original': acc}\n",
    "for _ in range(len(corruption)):\n",
    "    acc = clf.evaluate(X_c[_], y_test, batch_size=128, verbose=0)[1]\n",
    "    clf_accuracy[corruption[_]] = acc\n",
    "    print('{} {:.4f}'.format(corruption[_], acc))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Given the drop in performance, it is important that we detect the harmful data drift!\n",
    "\n",
    "### Detect drift <a class=\"anchor\" id=\"cd_detect\"></a>\n",
    "\n",
    "We are trying to detect data drift on high-dimensional (*32x32x3*) data using an aggregation of univariate K-S tests. It therefore makes sense to apply dimensionality reduction first. Some dimensionality reduction methods also used in [Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shift](https://arxiv.org/pdf/1810.11953.pdf) are readily available: **UAE** (Untrained AutoEncoder), **BBSDs** (black-box shift detection using the classifier's softmax outputs) and **PCA**. \n",
    "\n",
    "#### Untrained AutoEncoder <a class=\"anchor\" id=\"cd_uae\"></a>\n",
    "\n",
    "First we try UAE:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "tf.random.set_seed(0)\n",
    "\n",
    "# define encoder\n",
    "encoding_dim = 32\n",
    "encoder_net = tf.keras.Sequential(\n",
    "  [\n",
    "      InputLayer(input_shape=(32, 32, 3)),\n",
    "      Conv2D(64, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "      Conv2D(128, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "      Conv2D(512, 4, strides=2, padding='same', activation=tf.nn.relu),\n",
    "      Flatten(),\n",
    "      Dense(encoding_dim,)\n",
    "  ]\n",
    ")\n",
    "uae = UAE(encoder_net=encoder_net)\n",
    "preprocess_kwargs = {'model': uae, 'batch_size': 128}\n",
    "\n",
    "# initialise drift detector\n",
    "p_val = .05\n",
    "cd = KSDrift(\n",
    "    p_val=p_val,        # p-value for K-S test \n",
    "    X_ref=X_ref,       # test against original test set\n",
    "    preprocess_kwargs=preprocess_kwargs\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's check whether the detector thinks drift occurred within the original test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drift? No!\n"
     ]
    }
   ],
   "source": [
    "preds_h0 = cd.predict(X_h0, return_p_val=True)\n",
    "print('Drift? {}'.format(labels[preds_h0['data']['is_drift']]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected, no drift occurred. We can also inspect the feature-wise K-S statistics, threshold value and p-values for each univariate K-S test by (encoded) feature before the multivariate correction. Most of them are well above the $0.05$ threshold:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "K-S statistics:\n",
      "[0.017  0.013  0.0124 0.0234 0.0156 0.0144 0.025  0.0144 0.0216 0.0112\n",
      " 0.0122 0.0146 0.012  0.0154 0.015  0.0148 0.012  0.0158 0.0104 0.0144\n",
      " 0.0116 0.0148 0.0194 0.0112 0.0176 0.0154 0.0152 0.0158 0.0166 0.0202\n",
      " 0.0208 0.0196]\n",
      "\n",
      "K-S statistic threshold: 0.0015625\n",
      "\n",
      "p-values:\n",
      "[0.46531922 0.79201305 0.8367454  0.12939005 0.5769981  0.677735\n",
      " 0.08786641 0.677735   0.19387017 0.912423   0.850771   0.66088647\n",
      " 0.8642828  0.5936282  0.62716705 0.6440195  0.8642828  0.5604951\n",
      " 0.94969434 0.677735   0.88960564 0.6440195  0.30355498 0.912423\n",
      " 0.420929   0.5936282  0.61036026 0.5604951  0.496191   0.25943416\n",
      " 0.22956407 0.2920585 ]\n"
     ]
    }
   ],
   "source": [
    "# print stats for H0\n",
    "print('K-S statistics:')\n",
    "print(preds_h0['data']['distance'])\n",
    "print(f\"\\nK-S statistic threshold: {preds_h0['data']['threshold']}\")\n",
    "print('\\np-values:')\n",
    "print(preds_h0['data']['p_val'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's now check the predictions on the perturbed data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corruption type: gaussian_noise\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[1.3873853e-04 2.6511249e-01 9.4417885e-02 3.8390055e-01 7.1598392e-04\n",
      " 5.9501763e-04 1.7882723e-02 6.8460542e-01 2.9989053e-02 3.7781857e-02\n",
      " 6.6516680e-01 3.5206401e-07 4.4194090e-05 7.9351515e-03 6.5543061e-01\n",
      " 1.0220607e-04 1.9904150e-01 8.7863362e-01 1.1572546e-02 8.7863362e-01\n",
      " 2.7950103e-03 1.1150700e-02 3.6837953e-01 4.4130555e-01 7.8937606e-05\n",
      " 5.1594800e-03 6.2871156e-03 1.3455943e-03 9.7143359e-02 1.4105450e-07\n",
      " 1.7258449e-02 1.4648294e-01]\n",
      "\n",
      "Corruption type: motion_blur\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[3.75149103e-13 1.34077845e-02 2.96938539e-01 1.38920277e-01\n",
      " 5.03090501e-01 1.71172902e-01 6.25065863e-02 5.15948003e-03\n",
      " 4.85032678e-01 2.16893386e-03 7.48323351e-02 1.28741434e-03\n",
      " 1.24597345e-02 2.57197134e-05 1.32995670e-09 7.83811294e-10\n",
      " 7.93515146e-03 3.10395747e-01 2.09074125e-01 5.97174764e-01\n",
      " 5.95017627e-04 9.71433595e-02 3.38559955e-01 3.31362933e-01\n",
      " 7.15983915e-04 3.45860541e-01 5.68405211e-01 2.43121485e-06\n",
      " 5.30714750e-01 2.57197134e-05 1.80098459e-01 8.85808766e-01]\n",
      "\n",
      "Corruption type: brightness\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[4.8499879e-07 1.3873853e-04 1.1572546e-02 1.6691783e-10 2.7127105e-01\n",
      " 2.5326056e-02 1.6036824e-03 4.6618203e-05 2.1745481e-04 1.2474350e-01\n",
      " 6.2343346e-14 3.1785589e-08 8.9277834e-01 3.8390055e-01 8.2942984e-12\n",
      " 8.9514272e-09 2.4149875e-01 3.5845602e-03 2.2287663e-29 5.2143931e-01\n",
      " 2.2484422e-01 2.7950277e-16 1.9522957e-05 1.0870906e-01 7.6357084e-03\n",
      " 3.2130507e-04 3.4248088e-02 4.7411124e-14 2.5942018e-18 1.5351619e-03\n",
      " 4.6728885e-01 2.2888689e-06]\n",
      "\n",
      "Corruption type: pixelate\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[3.0361506e-01 8.5970649e-04 1.6684313e-01 7.0512637e-02 5.0309050e-01\n",
      " 7.0512637e-02 2.5905645e-01 2.2484422e-01 8.9953583e-01 3.5326433e-01\n",
      " 5.4005289e-01 3.3136293e-01 2.1316923e-02 3.9982069e-01 1.4057782e-06\n",
      " 7.9374194e-02 3.7114436e-04 7.7923602e-01 2.7127105e-01 2.5905645e-01\n",
      " 8.9953583e-01 1.6260068e-01 2.8040027e-02 9.1756582e-02 1.9192154e-02\n",
      " 1.7117290e-01 7.1357483e-01 2.9989053e-02 3.6077085e-01 1.2459734e-02\n",
      " 4.2442977e-01 2.6461019e-04]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# print stats for corrupted data\n",
    "for x, c in zip(X_c, corruption):\n",
    "    preds = cd.predict(x, return_p_val=True)\n",
    "    print(f'Corruption type: {c}')\n",
    "    print('Drift? {}'.format(labels[preds['data']['is_drift']]))\n",
    "    print('Feature-wise p-values:')\n",
    "    print(preds['data']['p_val'])\n",
    "    print('')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### BBSDs <a class=\"anchor\" id=\"cd_bbsds\"></a>\n",
    "\n",
    "For **BBSDs**, we use the classifier's softmax outputs for black-box shift detection. This method is based on [Detecting and Correcting for Label Shift with Black Box Predictors](https://arxiv.org/abs/1802.03916).\n",
    "\n",
    "Here we use the output of the softmax layer to detect the drift, but other hidden layers can be extracted as well by setting *'layer'* to the index of the desired hidden layer in the model:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# use output softmax layer\n",
    "preprocess_kwargs = {'model': HiddenOutput(model=clf, layer=-1), 'batch_size': 128}\n",
    "\n",
    "cd = KSDrift(\n",
    "    p_val=p_val,\n",
    "    X_ref=X_ref,\n",
    "    preprocess_kwargs=preprocess_kwargs\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is again no drift on the original held out test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "code_folding": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drift? No!\n",
      "\n",
      "p-values:\n",
      "[0.11774229 0.52796143 0.19387017 0.20236294 0.496191   0.72781175\n",
      " 0.12345381 0.420929   0.8367454  0.7604178 ]\n"
     ]
    }
   ],
   "source": [
    "preds_h0 = cd.predict(X_h0)\n",
    "print('Drift? {}'.format(labels[preds_h0['data']['is_drift']]))\n",
    "print('\\np-values:')\n",
    "print(preds_h0['data']['p_val'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We compare this with the perturbed data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "code_folding": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corruption type: gaussian_noise\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "\n",
      "Corruption type: motion_blur\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "\n",
      "Corruption type: brightness\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[0.0000000e+00 4.2024049e-15 2.8963613e-33 4.8499879e-07 2.3718185e-15\n",
      " 1.2473309e-05 2.9714003e-30 1.0611427e-09 4.6048109e-12 4.1857830e-17]\n",
      "\n",
      "Corruption type: pixelate\n",
      "Drift? Yes!\n",
      "Feature-wise p-values:\n",
      "[0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for x, c in zip(X_c, corruption):\n",
    "    preds = cd.predict(x)\n",
    "    print(f'Corruption type: {c}')\n",
    "    print('Drift? {}'.format(labels[preds['data']['is_drift']]))\n",
    "    print('Feature-wise p-values:')\n",
    "    print(preds['data']['p_val'])\n",
    "    print('')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For more functionality and examples, such as updating the reference data with reservoir sampling or picking another multivariate correction mechanism, check out [this example notebook](https://docs.seldon.io/projects/alibi-detect/en/latest/examples/cd_ks_cifar10.html).\n",
    "\n",
    "### Leveraging the adversarial detector for malicious drift detection <a class=\"anchor\" id=\"cd_ad\"></a>\n",
    "\n",
    "While monitoring covariate and predicted label shift is all very interesting and exciting, at the end of the day we are mainly interested in whether the drift actually hurt the model performance significantly. To this end, we can leverage the adversarial detector and measure univiariate drift on the adversarial scores! "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.random.seed(0)\n",
    "idx = np.random.choice(n_test, size=n_test // 2, replace=False)\n",
    "X_ref = scale_by_instance(X_test[idx])[0]\n",
    "\n",
    "cd = KSDrift(\n",
    "    p_val=.05,\n",
    "    X_ref=X_ref,\n",
    "    preprocess_fn=ad.score,  # adversarial score fn = preprocess step\n",
    "    preprocess_kwargs={'batch_size': 128}\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Make drift predictions on the original test set and corrupted data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "code_folding": [
     0
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "H0: Accuracy 0.9286 -- Drift? No!\n",
      "gaussian_noise: Accuracy 0.2208 -- Drift? Yes!\n",
      "motion_blur: Accuracy 0.6339 -- Drift? Yes!\n",
      "brightness: Accuracy 0.8913 -- Drift? Yes!\n",
      "pixelate: Accuracy 0.3666 -- Drift? Yes!\n"
     ]
    }
   ],
   "source": [
    "# evaluate classifier on different datasets\n",
    "clf_accuracy['h0'] = clf.evaluate(X_h0, y_h0, batch_size=128, verbose=0)[1]\n",
    "preds_h0 = cd.predict(X_h0)\n",
    "print('H0: Accuracy {:.4f} -- Drift? {}'.format(\n",
    "    clf_accuracy['h0'], labels[preds_h0['data']['is_drift']]))\n",
    "for x, c in zip(X_c, corruption):\n",
    "    preds = cd.predict(x)\n",
    "    print('{}: Accuracy {:.4f} -- Drift? {}'.format(\n",
    "        c, clf_accuracy[c],labels[preds['data']['is_drift']]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can therefore **use the scores of the detector itself to quantify the harmfulness of the drift**! We can generalise this to all the corruptions at each severity level in CIFAR-10-C.\n",
    "\n",
    "On the plot below we show the mean values and standard deviations of the adversarial scores per severity level. The plot shows the mean adversarial scores (lhs) and ResNet-32 accuracies (rhs) for increasing data corruption severity levels. Level 0 corresponds to the original test set. Harmful scores  are scores from instances which have been flipped from the correct to an incorrect prediction because of the corruption. Not harmful means that the prediction was unchanged after the corruption. The chart can be reproduced in [this notebook](https://docs.seldon.io/projects/alibi-detect/en/latest/examples/cd_ks_cifar10.html).\n",
    "\n",
    "![adversarialscores](image/adversarialscores.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Deploy <a class=\"anchor\" id=\"cd_deploy\"></a>\n",
    "\n",
    "We can deploy the drift detector in a similar fashion as the [outlier detector](#od_deploy). For a more detailed step-by-step overview of the deployment process, check [this notebook](https://docs.seldon.io/projects/seldon-core/en/latest/examples/drift_cifar10.html)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "drift-detector.default.example.com\n"
     ]
    }
   ],
   "source": [
    "SERVICE_HOSTNAMES=!(kubectl get ksvc drift-detector -o jsonpath='{.status.url}' | cut -d \"/\" -f 3)\n",
    "SERVICE_HOSTNAME_CD=SERVICE_HOSTNAMES[0]\n",
    "print(SERVICE_HOSTNAME_CD)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The deployed drift detector accumulates requests until a predefined `drift_batch_size` is reached, in our case $5000$ which is defined in the [*yaml* for the deployment](https://github.com/SeldonIO/seldon-core/blob/master/components/drift-detection/cifar10/cifar10.yaml) and set in the [drift detector wrapper](https://github.com/SeldonIO/seldon-core/blob/master/components/alibi-detect-server/adserver/cd_model.py). After $5000$ instances, the batch is cleared and fills up again."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a0a2b42ff1184e51b084087e891bb086",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=50.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Drift? No!\n"
     ]
    }
   ],
   "source": [
    "from tqdm.notebook import tqdm\n",
    "\n",
    "drift_batch_size = 5000\n",
    "\n",
    "# accumulate batches\n",
    "for i in tqdm(range(0, drift_batch_size, 100)):\n",
    "    x = X_h0[i:i+100]\n",
    "    predict(x)\n",
    "\n",
    "# check message dumper\n",
    "res=!kubectl logs $(kubectl get pod -l serving.knative.dev/configuration=message-dumper-drift -o jsonpath='{.items[0].metadata.name}') user-container\n",
    "data= []\n",
    "for i in range(0,len(res)):\n",
    "    if res[i] == 'Data,':\n",
    "        data.append(res[i+1])\n",
    "j = json.loads(json.loads(data[0]))\n",
    "print(\"Drift?\", labels[j[\"data\"][\"is_drift\"]==1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We now run the same test on some corrupted data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corruption: gaussian_noise\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "53cb25da3ffa4a7c87eb6312c1347567",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=50.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Drift? Yes!\n"
     ]
    }
   ],
   "source": [
    "c = 0\n",
    "\n",
    "print(f'Corruption: {corruption[c]}')\n",
    "\n",
    "# accumulate batches\n",
    "for i in tqdm(range(0, drift_batch_size, 100)):\n",
    "    x = X_c[c][i:i+100]\n",
    "    predict(x)\n",
    "\n",
    "# check message dumper\n",
    "res=!kubectl logs $(kubectl get pod -l serving.knative.dev/configuration=message-dumper-drift -o jsonpath='{.items[0].metadata.name}') user-container\n",
    "data= []\n",
    "for i in range(0,len(res)):\n",
    "    if res[i] == 'Data,':\n",
    "        data.append(res[i+1])\n",
    "j = json.loads(json.loads(data[1]))\n",
    "print(\"Drift?\", labels[j[\"data\"][\"is_drift\"]==1])"
   ]
  }
 ],
 "metadata": {
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}