{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "name": "model_inversion_attack_with_IBM_ART", "provenance": [], "mount_file_id": "1kFmZ17tpg3tp9op0FQBQQTjvgjJd-3O0", "authorship_tag": "ABX9TyNn+7WiezCV+mKsUexD+YFU", "include_colab_link": true }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "accelerator": "GPU" }, "cells": [ { "cell_type": "markdown", "metadata": { "id": "view-in-github", "colab_type": "text" }, "source": [ "\"Open" ] }, { "cell_type": "code", "metadata": { "id": "MDq0t_BaxkTo" }, "source": [ "import numpy as np\r\n", "import matplotlib.pyplot as plt\r\n", "import tensorflow as tf\r\n", "tf.compat.v1.disable_eager_execution()" ], "execution_count": 4, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "sLdL6QQmxsL9" }, "source": [ "!pip install adversarial-robustness-toolbox" ], "execution_count": null, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "6MYta-5Qx89s" }, "source": [ "from art.attacks.inference import model_inversion \r\n", "from art.estimators.classification import KerasClassifier" ], "execution_count": 6, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "C5CokfVFyQzL" }, "source": [ "def load_mnist():\r\n", " \"\"\"Loads MNIST-Dataset and preprocesses to combine training and test data.\"\"\"\r\n", " \r\n", " # load the existing MNIST digit dataset that comes in form of traing + test data and labels\r\n", " train, test = tf.keras.datasets.mnist.load_data()\r\n", " train_data, train_labels = train\r\n", " test_data, test_labels = test\r\n", "\r\n", " # scale the images from color values 0-255 to numbers from 0-1 to help the training process\r\n", " train_data = np.array(train_data, dtype=np.float32) / 255\r\n", " test_data = np.array(test_data, dtype=np.float32) / 255\r\n", "\r\n", " # convolutional layers expect images to have 3 dimensions (width, height, depth)\r\n", " # in color images the depth is 3 for the RGB channels\r\n", " # MNIST is grayscale and hence originally does not need a third dimension\r\n", " # so we need to artificially add it\r\n", " train_data = train_data.reshape(train_data.shape[0], 28, 28, 1)\r\n", " test_data = test_data.reshape(test_data.shape[0], 28, 28, 1)\r\n", " return train_data, train_labels, test_data, test_labels\r\n", " \r\n", "def make_model():\r\n", " \"\"\" Define a Keras model\"\"\"\r\n", " model = tf.keras.Sequential([\r\n", " tf.keras.layers.Conv2D(16, 8,\r\n", " strides=2,\r\n", " padding='same',\r\n", " activation='relu',\r\n", " input_shape=(28, 28, 1)),\r\n", " tf.keras.layers.MaxPool2D(2, 1),\r\n", " tf.keras.layers.Conv2D(32, 4,\r\n", " strides=2,\r\n", " padding='valid',\r\n", " activation='relu'),\r\n", " tf.keras.layers.MaxPool2D(2, 1),\r\n", " tf.keras.layers.Flatten(),\r\n", " tf.keras.layers.Dense(32, activation='relu'),\r\n", " tf.keras.layers.Dense(10, activation='softmax')\r\n", " ])\r\n", " return model" ], "execution_count": 7, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "oSaW9ZgVyjwj" }, "source": [ "train_data, train_labels, test_data, test_labels = load_mnist()" ], "execution_count": null, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "s_W2k3ARysJD", "outputId": "2ef4751e-0f46-45aa-f50e-cec1c82b6b43" }, "source": [ "# make the neural network model with the function specified above. then look at it\r\n", "model = make_model()\r\n", "model.summary()" ], "execution_count": 9, "outputs": [ { "output_type": "stream", "text": [ "Model: \"sequential\"\n", "_________________________________________________________________\n", "Layer (type) Output Shape Param # \n", "=================================================================\n", "conv2d (Conv2D) (None, 14, 14, 16) 1040 \n", "_________________________________________________________________\n", "max_pooling2d (MaxPooling2D) (None, 13, 13, 16) 0 \n", "_________________________________________________________________\n", "conv2d_1 (Conv2D) (None, 5, 5, 32) 8224 \n", "_________________________________________________________________\n", "max_pooling2d_1 (MaxPooling2 (None, 4, 4, 32) 0 \n", "_________________________________________________________________\n", "flatten (Flatten) (None, 512) 0 \n", "_________________________________________________________________\n", "dense (Dense) (None, 32) 16416 \n", "_________________________________________________________________\n", "dense_1 (Dense) (None, 10) 330 \n", "=================================================================\n", "Total params: 26,010\n", "Trainable params: 26,010\n", "Non-trainable params: 0\n", "_________________________________________________________________\n" ], "name": "stdout" } ] }, { "cell_type": "code", "metadata": { "id": "8TBgM6JnyvIZ" }, "source": [ "# specify hyperparameters\r\n", "optimizer = tf.keras.optimizers.SGD(learning_rate=0.1)\r\n", "loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)\r\n", "\r\n", "# compile the model\r\n", "model.compile(optimizer=optimizer, loss=loss, metrics=['accuracy'])" ], "execution_count": 10, "outputs": [] }, { "cell_type": "markdown", "metadata": { "id": "l3KE3jTS1uQ0" }, "source": [ "# Model Inversion Attack" ] }, { "cell_type": "code", "metadata": { "id": "peuB219bz5Rx" }, "source": [ "# convert keras model to ART model\r\n", "classifier = KerasClassifier(model=model, clip_values=(0, 1), use_logits=False)" ], "execution_count": 11, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "Arhi3ASQyv_r" }, "source": [ "# train the model\r\n", "history = classifier.fit(train_data, train_labels,\r\n", " batch_size=264, nb_epochs=10)" ], "execution_count": null, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "Nu0uaI7dy8jj" }, "source": [ "# create the attack object\r\n", "my_attack = model_inversion.MIFace(classifier)" ], "execution_count": 59, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "wGups9f_ygHT" }, "source": [ "# create an array of the classes to be attacked\r\n", "y_all = np.arange(10)\r\n", "y_all" ], "execution_count": 60, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "w7pcQHMMuEhm" }, "source": [ "inferred_images = my_attack.infer(x=None,y=y_all)\r\n" ], "execution_count": null, "outputs": [] }, { "cell_type": "code", "metadata": { "id": "6Qfh25wFuMuu" }, "source": [ "inferred_images.shape" ], "execution_count": null, "outputs": [] }, { "cell_type": "code", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 273 }, "id": "qLF6rYjPvzoE", "outputId": "64e180b1-0b81-488a-80ca-ef8a5e32cd51" }, "source": [ "# plot the inverted class representations\r\n", "num_row = 2\r\n", "num_col = 5\r\n", "fig, axes = plt.subplots(num_row, num_col, figsize=(1.5*num_col,2*num_row))\r\n", "for i in range(10):\r\n", " ax = axes[i//num_col, i%num_col]\r\n", " ax.set_axis_off()\r\n", " ax.imshow(inferred_images[i,:,:,:].reshape(28,28), cmap='gray')\r\n", " ax.set_title('Label: {}'.format(y_all[i]))\r\n", "\r\n", "plt.tight_layout()\r\n", "plt.show()" ], "execution_count": 64, "outputs": [ { "output_type": "display_data", "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "tags": [], "needs_background": "light" } } ] } ] }