generating_adversarial_patches.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from torchvision import datasets, transforms\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import  cv2\n",
    "\n",
    "animals = [\"chicken\", \"pig\", \"panda\", \"terrain\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Objective - To generate adversarial examples with a dark grey image as a starting point\n",
    "* This is done with the *negative epsilon* approach where we backpropagate on the image in order to decrease ts loss on a certain class\n",
    "\n",
    "## The ideal value of epsilon varies with the image\n",
    "And guessng the right value of epsilon is a time consuming process\n",
    "The right value of epsilon can be found by just looking for the probabilty maxima on a probability v/s epsilon graph."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Net(\n",
       "  (conv1): Conv2d(3, 7, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (pool): MaxPool2d(kernel_size=3, stride=3, padding=0, dilation=1, ceil_mode=False)\n",
       "  (dropout): Dropout(p=0.5, inplace=False)\n",
       "  (conv2): Conv2d(7, 10, kernel_size=(5, 5), stride=(1, 1))\n",
       "  (fc1): Linear(in_features=10, out_features=10, bias=True)\n",
       "  (fc2): Linear(in_features=10, out_features=8, bias=True)\n",
       "  (fc3): Linear(in_features=8, out_features=4, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, 7, 5)\n",
    "        self.pool = nn.MaxPool2d(3, 3)\n",
    "        \n",
    "        self.dropout = nn.Dropout(p=0.5)\n",
    "        self.conv2 = nn.Conv2d(7, 10, 5)\n",
    "        \n",
    "        self.fc1 = nn.Linear(10, 10)\n",
    "        self.fc2 = nn.Linear(10, 8)\n",
    "        self.fc3 = nn.Linear(8, 4)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.pool(F.relu(self.conv1(x)))\n",
    "        x = self.pool(F.relu(self.conv2(x)))\n",
    "        x = x.view(x.size(0), -1)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = self.fc3(x)\n",
    "        return x\n",
    "    \n",
    "hunter = Net()\n",
    "hunter.zero_grad()\n",
    "hunter.load_state_dict(torch.load(\"hunter.pt\"))\n",
    "hunter.eval()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Defining some utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def result_to_prob_dist_percentage(output_tensor):\n",
    "    prob_dist = torch.softmax(output_tensor.flatten(), dim = 0)\n",
    "    index = torch.argmax(prob_dist).item()\n",
    "    percentage = prob_dist[index]*100\n",
    "    return index, percentage.item()\n",
    "\n",
    "def tensor_to_image(tensor):\n",
    "    np_arr = tensor.squeeze(0).detach().numpy()\n",
    "    np_arr = np.moveaxis(np_arr, 0,2)\n",
    "    return np_arr.astype(np.uint8)\n",
    "\n",
    "def preprocess_image_for_model(img_path):\n",
    "    im = cv2.imread(img_path)\n",
    "    im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)  # convert to RGB color space \n",
    "    im_moveaxis =  np.moveaxis(im, 2, 0) \n",
    "    input_tensor =  torch.from_numpy(im_moveaxis).unsqueeze(0).float()\n",
    "    return input_tensor\n",
    "\n",
    "def result_to_bar_graph(output_tensor):\n",
    "    prob_dist = torch.softmax(output_tensor.flatten(), dim = 0)\n",
    "   \n",
    "    prob_dist = prob_dist.tolist()\n",
    "    \n",
    "    plt.bar(animals, prob_dist)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Starting out with a grey image and viewing the model's predictions on it\n",
    "* As we can see, it says that this dark grey patch is a pig, indirectly proving that we're pretty far from machines taking over the world"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = (5,5)\n",
    "\n",
    "noise = torch.ones([1,3,25,25])\n",
    "noise_np = tensor_to_image(noise)\n",
    "\n",
    "res = hunter(noise)\n",
    "\n",
    "plt.imshow(noise_np)\n",
    "plt.show()\n",
    "\n",
    "result_to_bar_graph(res)\n",
    "cv2.imwrite(\"images/noise.png\", noise_np)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## More utils for generating the patch\n",
    "* `spawn_ghost_animal_with_visualisation()` gives us the perturbed image in the form of a numpy array, and also gives the result in the form of `[<animal_name>, <percentage>]` along with `delta = epsilon*sign(data_gradient)`\n",
    "\n",
    "* `fgsm_attack_return_result_list()` is very similar t o the function above, but returns a probability distribution and the argmax of the distrbution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def spawn_ghost_animal_with_visualisation(image_path, model, epsilon, target_label):\n",
    "    \n",
    "    loss = nn.CrossEntropyLoss()\n",
    "    \n",
    "    input_tensor = preprocess_image_for_model(image_path)\n",
    "    input_tensor.requires_grad = True  # gradients \n",
    "    label = torch.tensor([target_label], dtype=torch.long)  \n",
    "    pred = model(input_tensor)\n",
    "    loss_val = loss(pred,label)\n",
    "    loss_val.backward() #backprop\n",
    "    data_grad = input_tensor.grad.data\n",
    "    bamboozle_tensor , addition= fgsm_attack(input_tensor, epsilon, data_grad)\n",
    "    bamboozle_image = tensor_to_image(bamboozle_tensor)\n",
    "    addition = tensor_to_image(addition)\n",
    "    predict = model(bamboozle_tensor)\n",
    "    index, perc = result_to_prob_dist_percentage(predict)\n",
    "#     print (\"Prediction = \", animals[index], perc, \"%\" )\n",
    "    \n",
    "    result = [animals[index], perc]\n",
    "\n",
    "    return bamboozle_image, result , addition\n",
    "\n",
    "def fgsm_attack_return_result_list(image_path, model, epsilon, target_label):\n",
    "    \n",
    "    loss = nn.CrossEntropyLoss()\n",
    "    \n",
    "    input_tensor = preprocess_image_for_model(image_path)\n",
    "    input_tensor.requires_grad = True  # gradients \n",
    "    label = torch.tensor([target_label], dtype=torch.long)  \n",
    "    pred = model(input_tensor)\n",
    "    loss_val = loss(pred,label)\n",
    "    loss_val.backward() #backprop\n",
    "    data_grad = input_tensor.grad.data\n",
    "    bamboozle_tensor , addition= fgsm_attack(input_tensor, epsilon, data_grad)\n",
    "    bamboozle_image = tensor_to_image(bamboozle_tensor)\n",
    "    addition = tensor_to_image(addition)\n",
    "    predict = model(bamboozle_tensor)\n",
    "    prob_dist = torch.softmax(predict.flatten(), dim = 0).tolist()\n",
    "    index = prob_dist.index(max(prob_dist))\n",
    "\n",
    "    return prob_dist, index\n",
    "\n",
    "def fgsm_attack(image, epsilon, data_grad):\n",
    "    # Collect the element-wise sign of the data gradient\n",
    "    sign_data_grad = data_grad.sign()\n",
    "        # Create the perturbed image by adjusting each pixel of the input image\n",
    "    perturbed_image = image + epsilon*sign_data_grad\n",
    "        \n",
    "    delta = epsilon * sign_data_grad\n",
    "    perturbed_image = torch.clamp(perturbed_image, 0, 255)\n",
    "    \n",
    "    return perturbed_image, delta\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plotting a probability distribution of various classes with varying epsilon\n",
    "\n",
    "This was one of the ways to verify the piecewise linear relationship  between the probability function and epsilon\n",
    "\n",
    "It also proves that **using too negative of an epsilon would not help** when we're trying to \"optimize\" an image to minimize the loss for a certain class. Have a look at the maxima that we got for \"chicken\" at around epsilon = -15\n",
    "\n",
    "This approach can be used to produced adversarial examples with *optimized epsilons*.\n",
    "\n",
    "In short, **the best possible value of epsilon is when the probabilty of the target class hits a maxima**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "preds = []\n",
    "x_axis = [m for m in range(-25,25)]\n",
    "for i in range(4):\n",
    "    temp = []\n",
    "    for eps in range(-25, 25):\n",
    "        prob, index = fgsm_attack_return_result_list(image_path = \"images/noise.png\", model = hunter, epsilon = eps, target_label = 0)\n",
    "        temp.append(prob[i])\n",
    "    preds.append(temp)\n",
    "    plt.plot(x_axis ,preds[i], label = animals[i])\n",
    "\n",
    "plt.rcParams[\"figure.figsize\"] = (20,10)\n",
    "plt.ylabel(\"class probability with target label = 0 (chicken) \", fontsize = 20)\n",
    "plt.xlabel(\"epsilon\", fontsize = 30)\n",
    "plt.legend(prop={\"size\":15})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating adversarial examples with optimized epsilons within a given domain\n",
    "* `generate_optimized_adversarial_examples()` first produces all the adversarial examples for a given range of epsilons\n",
    "* Then it looks for the value of epsilon which gave the highest probability on the target class. Thus giving us the ideal value of epsilon\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_optimized_adversarial_examples(image_path , model , lower_eps, higher_eps, target_label):\n",
    "    \n",
    "    x_axis = [m for m in range(lower_eps,higher_eps+1)]\n",
    "\n",
    "    temp = []\n",
    "    for eps in range(lower_eps, higher_eps +1):\n",
    "        prob, index = fgsm_attack_return_result_list(image_path = image_path, model = model, epsilon = eps, target_label = target_label)\n",
    "        temp.append(prob[target_label])\n",
    "    \n",
    "    max_eps = x_axis[temp.index(max(temp))]\n",
    "    image, result, delta = spawn_ghost_animal_with_visualisation(image_path, model, max_eps, target_label)\n",
    "    \n",
    "    return image, result   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = \"images/noise.png\"\n",
    "\n",
    "images = []\n",
    "results = []\n",
    "\n",
    "for d in range(4):\n",
    "    im, res = generate_optimized_adversarial_examples(image_path = path , model = hunter , lower_eps = -120, higher_eps = 1, target_label = d)\n",
    "    \n",
    "    images.append(im)\n",
    "    results.append(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plt.rcParams[\"figure.figsize\"] = (10,10)\n",
    "fig, axs = plt.subplots(2,2)\n",
    "\n",
    "for a in range(2):\n",
    "    for b in range(2):\n",
    "        axs[a][b].imshow(images[a+b])\n",
    "        axs[a][b].set_xlabel(results[a+b][0] + \" \"+ str(round(results[a+b][1],2)) + \"%\", fontsize = 15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = (5,5)\n",
    "\n",
    "plt.imshow(images[2])\n",
    "plt.axis(\"off\")\n",
    "plt.show()\n",
    "cv2.imwrite(\"images/bricks_org.png\", images[2])\n"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}