tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "from sklearn.metrics import accuracy_score, roc_curve, recall_score, RocCurveDisplay, PrecisionRecallDisplay, precision_score, f1_score\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from exe_scanner import EXEScanner, MalwareDetectorWrapper\n",
    "from malware_classifier import GBDTMalwareClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Setup EXE-scanner"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load pretrained EXE-scanner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "exe_scanner_model = GBDTMalwareClassifier(\"EXE_scanner_GBDT-dataset=benign-AE_all-before_2022\", is_trained=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## or Custom setup of EXE-scanner"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load train set (see README for details on how to obtain the dataset)\n",
    "X_train = pd.read_csv(\"data/X_train_benign-AE_all-before_2022.csv\")\n",
    "y_train = pd.read_csv(\"data/y_train_benign-AE_all-before_2022.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train EXE-scanner\n",
    "exe_scanner_model = GBDTMalwareClassifier(\"EXE_scanner_GBDT-dataset=benign-AE_all-before_2022\", is_trained=False)\n",
    "training_info = {}\n",
    "exe_scanner_model.train(X_train, y_train[\"y\"], training_info)\n",
    "print(exe_scanner_model.name, \"training info:\", training_info)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load pretrained main classifier (see README for details on how to download pretrained models)\n",
    "main_classifier_model = GBDTMalwareClassifier(\"CLF_GBDT-dataset=EMBER_train\", is_trained=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def finalize_predictions(y_true, y_pred, target_fpr, y_pred_second=None, verbose=False):\n",
    "    if verbose:\n",
    "        print(\"#\" * 100)\n",
    "        print(\"Dataset distribution:\", np.unique(y_true, return_counts=True))\n",
    "        print(\"Processing results...\")\n",
    "    \n",
    "    fixed_predictions = {}\n",
    "\n",
    "    fpr_arr, _, thresholds_roc = roc_curve(y_true, y_pred)\n",
    "    \n",
    "    threshold = thresholds_roc[fpr_arr < target_fpr][-1] # Get the threshold for the target FPR\n",
    "        \n",
    "    y_pred_binary = y_pred >= threshold\n",
    "\n",
    "    standalone_pred_stats = np.unique(y_pred_binary, return_counts=True)\n",
    "    if verbose:\n",
    "        print(\"Y-true stats: \", np.unique(y_true, return_counts=True))\n",
    "        print(\"Prediction standalone: \", standalone_pred_stats)\n",
    "        print(\"Threshold standalone: \", threshold)\n",
    "    \n",
    "    # If the second model (EXE-scanner) is provided, combine the predictions\n",
    "    if y_pred_second is not None:\n",
    "        y_pred_combined = np.zeros_like(y_pred)\n",
    "\n",
    "        # Replace the 0 prediction (benign) with the second model prediction score\n",
    "        for i, pred in enumerate(y_pred_binary):\n",
    "            if pred == False:\n",
    "                y_pred_combined[i] = y_pred_second[i]\n",
    "            else:\n",
    "                y_pred_combined[i] = 1.0\n",
    "\n",
    "        # Recalculate new thresholds\n",
    "        fpr_arr_combined, _, thresholds_roc_combined = roc_curve(y_true, y_pred_combined)\n",
    "        threshold_combined = thresholds_roc_combined[fpr_arr_combined < target_fpr][-1]\n",
    "            \n",
    "        combined_pred_stats = np.unique(y_pred_combined, return_counts=True)\n",
    "        if verbose:\n",
    "            print(\"Threshold combined: \", threshold_combined)\n",
    "            print(\"Prediction combined: \", combined_pred_stats)\n",
    "\n",
    "        # Update the binary prediction (leave the first model malicious=1 prediction)\n",
    "        num_malicious = 0\n",
    "        for i, pred in enumerate(y_pred_combined):\n",
    "            if y_pred_binary[i] == 1:\n",
    "                num_malicious += 1\n",
    "                continue\n",
    "            else:\n",
    "                y_pred_binary[i] = y_pred_combined[i] >= threshold_combined\n",
    "\n",
    "        # Same for this assert\n",
    "        assert num_malicious == combined_pred_stats[1][-1], f\"Malicious predictions should be transfered to combined model with EXE-scanner, {num_malicious} != {combined_pred_stats[1][-1]}\"\n",
    "\n",
    "    # Results for the binary prediction\n",
    "    y_pred_binary = y_pred_binary.astype(int)\n",
    "    fpr = np.sum(y_true[y_pred_binary == 1] == 0) / np.sum(y_true == 0)\n",
    "    fnr = np.sum(y_true[y_pred_binary == 0] == 1) / np.sum(y_true == 1)\n",
    "\n",
    "    fixed_predictions[target_fpr] = {\n",
    "        \"y_pred\": y_pred_binary,\n",
    "        \"y_pred_combined\": y_pred_combined if y_pred_second is not None else None,\n",
    "        \"threshold\": threshold,\n",
    "        \"threshold_combined\": threshold_combined if y_pred_second is not None else None,\n",
    "    }\n",
    "\n",
    "    if verbose:\n",
    "        fig, [ax_roc, ax_ap] = plt.subplots(1, 2, figsize=(11, 5))\n",
    "        PrecisionRecallDisplay.from_predictions(y_true, y_pred, ax=ax_ap, name=\"Main Classifier\")\n",
    "        RocCurveDisplay.from_predictions(y_true, y_pred, ax=ax_roc, name=\"Main Classifier\")\n",
    "\n",
    "        if y_pred_second is not None:\n",
    "            PrecisionRecallDisplay.from_predictions(y_true, y_pred_second, ax=ax_ap, name=\"EXE-Scanner\")\n",
    "            RocCurveDisplay.from_predictions(y_true, y_pred_second, ax=ax_roc, name=\"EXE-Scanner\")\n",
    "\n",
    "            PrecisionRecallDisplay.from_predictions(y_true, y_pred_combined, ax=ax_ap, name=\"Combined Detector\")\n",
    "            RocCurveDisplay.from_predictions(y_true, y_pred_combined, ax=ax_roc, name=\"Combined Detector\")\n",
    "\n",
    "        ax_roc.grid(linestyle=\"--\")\n",
    "        ax_ap.grid(linestyle=\"--\")\n",
    "        \n",
    "        # Zoom ax_ap to the top right corner\n",
    "        ax_ap.set_xlim([0.8, 1.0])\n",
    "        ax_ap.set_ylim([0.8, 1.0])\n",
    "\n",
    "        # Zoom ax_roc to the top left corner\n",
    "        ax_roc.set_xlim([0.0, 0.2])\n",
    "        ax_roc.set_ylim([0.8, 1.0])\n",
    "\n",
    "        # Move y-axis of ax_ap to the right\n",
    "        ax_ap.yaxis.tick_right()\n",
    "        ax_ap.yaxis.set_label_position(\"right\")\n",
    "\n",
    "        plt.show()\n",
    "\n",
    "    return fixed_predictions\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load test set (see README for details on how to obtain the dataset)\n",
    "X_test = pd.read_csv(\"data/X_test_benign-AE_all-before_2022.csv\")\n",
    "y_test = pd.read_csv(\"data/y_test_benign-AE_all-before_2022.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "####################################################################################################\n",
      "Dataset distribution: (array([0, 1]), array([24202, 16323]))\n",
      "Processing results...\n",
      "Y-true stats:  (array([0, 1]), array([24202, 16323]))\n",
      "Prediction standalone:  (array([False,  True]), array([25406, 15119]))\n",
      "Threshold standalone:  0.4413384963629881\n",
      "Threshold combined:  0.9999999950656784\n",
      "Prediction combined:  (array([1.54755413e-11, 2.39238896e-11, 2.71761080e-11, ...,\n",
      "       1.00000000e+00, 1.00000000e+00, 1.00000000e+00]), array([    1,     1,     1, ...,     1,     1, 15119]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1100x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "y_test_pred_main_clf = main_classifier_model.predict_proba(X_test)\n",
    "y_test_pred_exe_scanner = exe_scanner_model.predict_proba(X_test)\n",
    "\n",
    "fixed_predictions_standalone = finalize_predictions(y_test[\"y\"], y_test_pred_main_clf, target_fpr=0.01, verbose=False)\n",
    "fixed_predictions_combined = finalize_predictions(y_test[\"y\"], y_test_pred_main_clf, target_fpr=0.01, y_pred_second=y_test_pred_exe_scanner, verbose=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare acurracy, precision, recall and F1-score of the standalone and combined models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "accuracy_standalone = accuracy_score(y_test[\"y\"], fixed_predictions_standalone[0.01][\"y_pred\"])\n",
    "precision_standalone = precision_score(y_test[\"y\"], fixed_predictions_standalone[0.01][\"y_pred\"])\n",
    "recall_standalone = recall_score(y_test[\"y\"], fixed_predictions_standalone[0.01][\"y_pred\"])\n",
    "f1_standalone = f1_score(y_test[\"y\"], fixed_predictions_standalone[0.01][\"y_pred\"])\n",
    "\n",
    "accuracy_combined = accuracy_score(y_test[\"y\"], fixed_predictions_combined[0.01][\"y_pred\"])\n",
    "precision_combined = precision_score(y_test[\"y\"], fixed_predictions_combined[0.01][\"y_pred\"])\n",
    "recall_combined = recall_score(y_test[\"y\"], fixed_predictions_combined[0.01][\"y_pred\"])\n",
    "f1_combined = f1_score(y_test[\"y\"], fixed_predictions_combined[0.01][\"y_pred\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "width = 0.35\n",
    "x = np.arange(4)\n",
    "labels = [\"Accuracy\", \"Precision\", \"Recall\", \"F1-score\"]\n",
    "values_standalone = [accuracy_standalone, precision_standalone, recall_standalone, f1_standalone]\n",
    "values_combined = [accuracy_combined, precision_combined, recall_combined, f1_combined]\n",
    "\n",
    "rects1 = ax.bar(x - width/2, values_standalone, width, label='Standalone', color=\"blue\")\n",
    "rects2 = ax.bar(x + width/2, values_combined, width, label='Combined', color=\"orange\")\n",
    "\n",
    "ax.set_ylabel('Scores')\n",
    "ax.set_title('Scores by model')\n",
    "ax.set_xticks(x)\n",
    "ax.set_xticklabels(labels)\n",
    "ax.legend(loc='lower right')\n",
    "\n",
    "# Add values to the bars\n",
    "for rect in rects1:\n",
    "    height = rect.get_height()\n",
    "    ax.text(rect.get_x() + rect.get_width() / 2., height,\n",
    "            f'{height:.2f}', ha='center', va='bottom')\n",
    "\n",
    "for rect in rects2:\n",
    "    height = rect.get_height()\n",
    "    ax.text(rect.get_x() + rect.get_width() / 2., height,\n",
    "            f'{height:.2f}', ha='center', va='bottom')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculate bad churn (regressions in model accuracy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_different_indices(array1, array2):\n",
    "    if len(array1) != len(array2):\n",
    "        raise ValueError(\"Arrays must have the same length\")\n",
    "    \n",
    "    different_indices = [i for i in range(len(array1)) if array1[i] != array2[i]]\n",
    "    return different_indices\n",
    "\n",
    "# Calculates \"bad churn\" - when the new model makes a mistake that the old model did not make\n",
    "def calculate_bad_churn(y_true, y_pred_standalone, y_pred_combined, verbose=False):\n",
    "    # Verify inputs\n",
    "    if len(y_true) != len(y_pred_standalone) or len(y_true) != len(y_pred_combined):\n",
    "        raise ValueError(\"Input arrays must have the same length\")\n",
    "\n",
    "    if verbose:\n",
    "        print(f\"y_true: {y_true}\")\n",
    "        print(f\"y_pred_standalone: {y_pred_standalone}\")\n",
    "        print(f\"y_pred_combined: {y_pred_combined}\")\n",
    "\n",
    "    # Calculate totals\n",
    "    total_benign = total_malicious = 0\n",
    "    for y in y_true:\n",
    "        if y == 0:\n",
    "            total_benign += 1\n",
    "        else:\n",
    "            total_malicious += 1\n",
    "\n",
    "    if verbose:\n",
    "        print(f\"Total benign: {total_benign}\")\n",
    "        print(f\"Total malicious: {total_malicious}\")\n",
    "\n",
    "    # Calculate indices where the combined model made a mistake\n",
    "    incorrect_standalone = find_different_indices(y_true, y_pred_standalone)\n",
    "    incorrect_combined = find_different_indices(y_true, y_pred_combined)\n",
    "\n",
    "    if verbose:\n",
    "        print(f\"Incorrect standalone: {y_true != y_pred_combined}\")\n",
    "        print(f\"Incorrect combined: {y_true != y_pred_standalone}\")\n",
    "        print(f\"Incorrect standalone indices: {incorrect_standalone}\")\n",
    "        print(f\"Incorrect combined indices: {incorrect_combined}\")\n",
    "\n",
    "    # Calculate bad churn, where the new model made a mistake that the standalone model did not\n",
    "    bad_churn_benign = bad_churn_malicious = 0\n",
    "    for idx in incorrect_combined:\n",
    "        if idx not in incorrect_standalone:\n",
    "            if y_true[idx] == 0:\n",
    "                bad_churn_benign += 1\n",
    "            else:\n",
    "                bad_churn_malicious += 1\n",
    "\n",
    "\n",
    "    results = {\n",
    "        \"benign_churn_rate\": bad_churn_benign / total_benign,\n",
    "        \"malicious_churn_rate\": bad_churn_malicious / total_malicious,\n",
    "        \"benign_churn\": bad_churn_benign,\n",
    "        \"malicious_churn\": bad_churn_malicious,\n",
    "    }\n",
    "\n",
    "    if verbose:\n",
    "        print(f\"Bad churn benign: {bad_churn_benign}/{total_benign}\")\n",
    "        print(f\"Bad churn malicious: {bad_churn_malicious}/{total_malicious}\")\n",
    "\n",
    "    return results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Goodware Flip Rate: 0.00\n",
      "Malware Flip Rate: 0.00\n",
      "Num. Goodware Flips: 0\n",
      "Num. Malware Flips: 0\n"
     ]
    }
   ],
   "source": [
    "bad_churn_results = calculate_bad_churn(y_test[\"y\"], fixed_predictions_standalone[0.01][\"y_pred\"], fixed_predictions_combined[0.01][\"y_pred\"], verbose=False)\n",
    "\n",
    "print(f\"Goodware Flip Rate: {bad_churn_results['benign_churn_rate']:.2f}\")\n",
    "print(f\"Malware Flip Rate: {bad_churn_results['malicious_churn_rate']:.2f}\")\n",
    "print(f\"Num. Goodware Flips: {bad_churn_results['benign_churn']}\")\n",
    "print(f\"Num. Malware Flips: {bad_churn_results['malicious_churn']}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Real-world executable files"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "class WrapperEXEscanner:\n",
    "    \"\"\"\n",
    "    A wrapper class for the use of EXE scanner with standalone malware classifier.\n",
    "\n",
    "    Parameters:\n",
    "    - main_classifier: The main classifier model.\n",
    "    - EXE_scanner: The EXE scanner model.\n",
    "    - main_classifier_threshold: The threshold value for the main classifier. (optional)\n",
    "\n",
    "    Methods:\n",
    "    - predict_sample(bytez, return_score=False): Predicts the label for a given sample.\n",
    "\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, main_classifier, EXE_scanner):\n",
    "        self.main_classifier = main_classifier\n",
    "        self.EXE_scanner = EXE_scanner\n",
    "\n",
    "    def predict_sample(self, bytez, return_score=False):\n",
    "        \"\"\"\n",
    "        Predicts the label for a given sample.\n",
    "\n",
    "        Parameters:\n",
    "        - bytez: The input sample in byte format.\n",
    "        - return_score: Whether to return the prediction score. (default: False)\n",
    "\n",
    "        Returns:\n",
    "        - predicted_label: The predicted label for the sample.\n",
    "        - score: The prediction score (if return_score is True).\n",
    "\n",
    "        \"\"\"\n",
    "\n",
    "        score = self.main_classifier.get_score(bytez)\n",
    "        is_malicious = score > self.main_classifier.threshold\n",
    "        print(f\"Main classifier prediction: {is_malicious} with score {score}\")\n",
    "        if not is_malicious:\n",
    "            score = self.EXE_scanner.get_score(bytez)\n",
    "            is_malicious = score > self.EXE_scanner.threshold\n",
    "            print(f\"EXE scanner prediction: {is_malicious} with score {score}\")\n",
    "\n",
    "        predicted_label = int(is_malicious)\n",
    "        if return_score:\n",
    "            return predicted_label, score\n",
    "        else:\n",
    "            return predicted_label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "main_classifier_model = GBDTMalwareClassifier(\"CLF_GBDT-dataset=EMBER_train\", is_trained=True)\n",
    "main_classifier = MalwareDetectorWrapper(main_classifier_model, threshold=0.4413384963629881) # Threshold saved in \"data/models/fixed_thresholds/CLF_GBDT-dataset=EMBER_train-test_dataset=before2022/...\"\n",
    "exe_scanner = EXEScanner(target_fpr=0.01)\n",
    "exe_scanner_pipeline = WrapperEXEscanner(main_classifier, exe_scanner)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Main classifier prediction: False with score 0.00192148027492247\n",
      "EXE scanner prediction: False with score 0.7716771419831085\n",
      "HelloWorld: 0 (0.7716771419831085)\n"
     ]
    }
   ],
   "source": [
    "# Requires placing executables inside \"samples\" folder in the root of this repository\n",
    "SAMPLE_FOLDER = \"samples\"\n",
    "for file in os.listdir(SAMPLE_FOLDER):\n",
    "    with open(os.path.join(SAMPLE_FOLDER, file), \"rb\") as f:\n",
    "        bytez = f.read()\n",
    "        label, score = exe_scanner_pipeline.predict_sample(bytez, return_score=True)\n",
    "        print(f\"{file}: {label} ({score})\")"
   ]
  }
 ],
 "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": "3.11.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}