feature_motivation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "ad125052",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import glob\n",
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import json\n",
    "import seaborn as sns\n",
    "import pickle as pkl\n",
    "import copy\n",
    "import matplotlib.patches as mpatches\n",
    "from utils import OPT_PATH as opt_path\n",
    "# refer to https://matplotlib.org/stable/api/markers_api.html\n",
    "markers = [\"o\", \"v\", \"s\", \"*\", \"D\"]\n",
    "# refer to https://matplotlib.org/stable/gallery/color/named_colors.html\n",
    "colors = ['tab:blue', 'tab:orange', 'tab:green', 'tab:red', 'tab:purple', 'tab:brown', 'tab:pink', 'tab:gray']\n",
    "# https://matplotlib.org/3.1.0/gallery/lines_bars_and_markers/linestyles.html\n",
    "linestyles = 'solid'\n",
    "# refer to https://github.com/bvalgard/Boxplots/blob/master/Boxplots.ipynb"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "05e68148",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "glob.glob(os.path.join(opt_path, '*SI_DTMI*'))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c6c7c9f5",
   "metadata": {},
   "source": [
    "### Benign v.s. I-FGSM, DTMI, DTMI-Ours"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "0d549cfa",
   "metadata": {},
   "outputs": [],
   "source": [
    "ifgsm_path = glob.glob(os.path.join(opt_path, '*-DTMI-*motivation_IFGSM/cosine_similarity.json'))\n",
    "with open(ifgsm_path[0], 'rb') as ipt:\n",
    "    ifgsm_info = pkl.load(ipt)\n",
    "    \n",
    "dtmi_path = glob.glob(os.path.join(opt_path, '*-DTMI-*motivation/cosine_similarity.json'))\n",
    "with open(dtmi_path[0], 'rb') as ipt:\n",
    "    dtmi_info = pkl.load(ipt)\n",
    "    \n",
    "our_dtmi_path = glob.glob(os.path.join(opt_path, '*-DTMI_Local_FeatureSimilarityLoss-*motivation/cosine_similarity.json'))\n",
    "with open(our_dtmi_path[0], 'rb') as ipt:\n",
    "    our_dtmi_info = pkl.load(ipt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "69c063ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "names = ['Benign'] * len(ifgsm_info[770][2][0]) + ['I-FGSM'] * len(ifgsm_info[770][2][1]) + \\\n",
    "    ['DTMI'] * len(dtmi_info[770][2][1]) + ['DTMI-LI'] * len(our_dtmi_info[770][2][1])\n",
    "info = pd.DataFrame()\n",
    "info['Calculate Cosine Similarity Between'] = names\n",
    "info['Cosine Similarity Value'] = ifgsm_info[770][2][0] + ifgsm_info[770][2][1] + dtmi_info[770][2][1] + our_dtmi_info[770][2][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "cbac8d75",
   "metadata": {},
   "outputs": [],
   "source": [
    "info.loc[info['Calculate Cosine Similarity Between'] == 'DTMI-LI', 'Calculate Cosine Similarity Between'] = 'DTMI-SU'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "514a9641",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style='whitegrid')\n",
    "sns.set_palette(sns.color_palette(\"Paired\"))\n",
    "fig, ax = plt.subplots(figsize=(4,3))\n",
    "# g = sns.boxplot(data=df, width=0.7)#notch=True, ) # df[['p1_satisfaction','p2_satisfaction','p3_satisfaction']]\n",
    "g = sns.boxplot(x=\"Calculate Cosine Similarity Between\", y=\"Cosine Similarity Value\", data=info, \\\n",
    "                width=0.7, showfliers = False) # , \n",
    "\n",
    "plt.xlabel(\"\", fontsize=14)\n",
    "# with a descriptive title a ylabel might not be necessary\n",
    "plt.ylabel(\"Cosine similarity\", fontsize=14)\n",
    "\n",
    "# plt.xticks([0, 1, 2, 3],['Benign',''])\n",
    "# set y-axis values (I only want whole numbers)\n",
    "# plt.yticks(np.arange(0,14))\n",
    "\n",
    "# remove all borders except bottom\n",
    "sns.despine(top=True,\n",
    "            right=True,\n",
    "            left=True,\n",
    "            bottom=False)\n",
    "plt.savefig('./figures/features_cs.pdf', format='pdf', bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "58e67670",
   "metadata": {},
   "source": [
    "### Rebuttal SI and Admix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "746943ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "ifgsm_path = glob.glob(os.path.join('/share_io02_ssd/zhipeng/targeted_attack_loss/', 'Loss-resnet50-NIPSDataset-Admix_DTMI-Target_CE-baseline/cosine_similarity.json'))\n",
    "with open(ifgsm_path[0], 'rb') as ipt:\n",
    "    ifgsm_info = pkl.load(ipt)\n",
    "    \n",
    "dtmi_path = glob.glob(os.path.join('/share_io02_ssd/zhipeng/targeted_attack_loss/', 'Loss-resnet50-NIPSDataset-SI_DTMI-Target_CE-baseline/cosine_similarity.json'))\n",
    "with open(dtmi_path[0], 'rb') as ipt:\n",
    "    dtmi_info = pkl.load(ipt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "eee46ab9",
   "metadata": {},
   "outputs": [],
   "source": [
    "names = ['Admix'] * len(ifgsm_info[770][2][0]) + ['SI'] * len(ifgsm_info[770][2][1])\n",
    "info = pd.DataFrame()\n",
    "info['Calculate Cosine Similarity Between'] = names\n",
    "info['Cosine Similarity Value'] = ifgsm_info[770][2][1] + dtmi_info[770][2][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "fcbb8aed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.set(style='whitegrid')\n",
    "sns.set_palette(sns.color_palette(\"Paired\"))\n",
    "fig, ax = plt.subplots(figsize=(4,3))\n",
    "# g = sns.boxplot(data=df, width=0.7)#notch=True, ) # df[['p1_satisfaction','p2_satisfaction','p3_satisfaction']]\n",
    "g = sns.boxplot(x=\"Calculate Cosine Similarity Between\", y=\"Cosine Similarity Value\", data=info, \\\n",
    "                width=0.7, showfliers = False) # , \n",
    "\n",
    "plt.xlabel(\"\", fontsize=14)\n",
    "# with a descriptive title a ylabel might not be necessary\n",
    "plt.ylabel(\"Cosine similarity\", fontsize=14)\n",
    "\n",
    "# plt.xticks([0, 1, 2, 3],['Benign',''])\n",
    "# set y-axis values (I only want whole numbers)\n",
    "# plt.yticks(np.arange(0,14))\n",
    "\n",
    "# remove all borders except bottom\n",
    "sns.despine(top=True,\n",
    "            right=True,\n",
    "            left=True,\n",
    "            bottom=False)\n",
    "# plt.savefig('./figures/features_cs.pdf', format='pdf', bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cefb5850",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.7.11"
  },
  "vscode": {
   "interpreter": {
    "hash": "79e10be247e32c74e465721b76249ea10bbcb221b34115737dc8c7eb7f76b5cc"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}