analyze_medical_imaging_thresholds.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "from tqdm import tqdm\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "from sklearn.metrics import balanced_accuracy_score \n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_theme()\n",
    "\n",
    "from source.constants import RESULTS_PATH, PLOTS_PATH\n",
    "from source.data.medical_imaging import get_chexpert\n",
    "\n",
    "os.makedirs(PLOTS_PATH, exist_ok=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "method_seeds = [42, 142, 242, 342, 442]\n",
    "dseed = 42\n",
    "\n",
    "model = [\"resnet18\", \"resnet34\", \"resnet50\"][2]\n",
    "\n",
    "verbose = False\n",
    "\n",
    "pas = [\"old\", \"woman\", \"white\"]\n",
    "pa = 0 # 0, 1, 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "# patients general 65401\n",
      "# patients with race 58010\n",
      "24638 24638\n"
     ]
    }
   ],
   "source": [
    "full_ds, _, _ = get_chexpert(load_to_ram=False)\n",
    "\n",
    "run_path = os.path.join(RESULTS_PATH, f\"chexpert_{model}_mseed{method_seeds[0]}_dseed{dseed}\")\n",
    "fair_inds = torch.load(os.path.join(run_path, \"fair_inds.pt\"))\n",
    "val_inds = torch.load(os.path.join(run_path, \"val_inds.pt\"))\n",
    "\n",
    "print(len(fair_inds), len(val_inds))\n",
    "\n",
    "y_fair_t = full_ds.targets[fair_inds]\n",
    "a_fair_t = full_ds.protected_attributes[pa, fair_inds]\n",
    "y_val_t = full_ds.targets[val_inds]\n",
    "a_val_t = full_ds.protected_attributes[pa, val_inds]\n",
    "\n",
    "# switch 0 / 1 in protected attribute\n",
    "pas = [\"young\", \"man\", \"non-white\"]\n",
    "a_fair_t = 1 - a_fair_t\n",
    "a_val_t = 1 - a_val_t\n",
    "\n",
    "# switch label\n",
    "# y_fair_t = 1 - y_fair_t\n",
    "# y_val_t = 1 - y_val_t\n",
    "\n",
    "p_a_val = a_val_t.float().mean().item() * 100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load probits\n",
    "fair_probits, val_probits = list(), list()\n",
    "for mseed in method_seeds:\n",
    "    path = os.path.join(RESULTS_PATH, f\"chexpert_{model}_mseed{mseed}_dseed{dseed}\")\n",
    "\n",
    "    # don't do fairness ensemble on medical imaging - use this split as test dataset\n",
    "    fp = torch.load(os.path.join(path, f\"fair_probits.pt\"))\n",
    "    #fp = 1 - fp\n",
    "    fair_probits.append(fp)\n",
    "    vp = torch.load(os.path.join(path, f\"val_probits.pt\"))\n",
    "    #vp = 1 - vp\n",
    "    val_probits.append(vp)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Optimize Thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "thresholds = np.linspace(0, 1, 1001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/5 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 5/5 [06:00<00:00, 72.02s/it]\n"
     ]
    }
   ],
   "source": [
    "ensemble_best_thresholds = list()\n",
    "member_best_thresholds = list()\n",
    "ensemble_thresholds, ensemble_accs = list(), list()\n",
    "member_thresholds, member_accs = list(), list()\n",
    "\n",
    "for m in tqdm(range(len(method_seeds))):\n",
    "    threshs, accs = [[], [], []], [[], [], []]\n",
    "    best_thresholds, best_accs = [0, 0, 0], [0, 0, 0]\n",
    "\n",
    "    val_m_probits = torch.mean(val_probits[m], dim=0)\n",
    "\n",
    "    X = np.array(range(len(y_val_t)))\n",
    "    y = y_val_t.numpy()\n",
    "    group=a_val_t.numpy()\n",
    "    y_scores = val_m_probits[:, -1].numpy()\n",
    "\n",
    "    for thr in thresholds:\n",
    "        preds = y_scores > thr\n",
    "        for g in range(3):\n",
    "            if g == 2:\n",
    "                acc = balanced_accuracy_score(y, preds)\n",
    "            else:\n",
    "                acc = balanced_accuracy_score(y[group == g], preds[group == g])\n",
    "            if acc >= best_accs[g]:\n",
    "                best_accs[g] = acc\n",
    "                best_thresholds[g] = thr\n",
    "            threshs[g].append(thr)\n",
    "            accs[g].append(acc)\n",
    "    \n",
    "    ensemble_thresholds.append(threshs)\n",
    "    ensemble_accs.append(accs)\n",
    "    \n",
    "    ensemble_best_thresholds.append(best_thresholds)\n",
    "\n",
    "    for me, member_probits in enumerate(val_probits[m]):\n",
    "        threshs, accs = [[], [], []], [[], [], []]\n",
    "        best_thresholds, best_accs = [0, 0, 0], [0, 0, 0] \n",
    "        \n",
    "        y_scores = member_probits[:, -1].numpy()\n",
    "        for thr in thresholds:\n",
    "            preds = y_scores > thr\n",
    "            for g in range(3):\n",
    "                if g == 2:\n",
    "                    acc = balanced_accuracy_score(y, preds)\n",
    "                else:\n",
    "                    acc = balanced_accuracy_score(y[group == g], preds[group == g])\n",
    "                if acc > best_accs[g]:\n",
    "                    best_accs[g] = acc\n",
    "                    best_thresholds[g] = thr\n",
    "                threshs[g].append(thr)\n",
    "                accs[g].append(acc)\n",
    "\n",
    "        member_thresholds.append(threshs)\n",
    "        member_accs.append(accs)\n",
    "\n",
    "        member_best_thresholds.append(best_thresholds)\n",
    "\n",
    "ensemble_best_thresholds = np.asarray(ensemble_best_thresholds)\n",
    "member_best_thresholds = np.asarray(member_best_thresholds)\n",
    "ensemble_thresholds = np.asarray(ensemble_thresholds)\n",
    "ensemble_accs = np.asarray(ensemble_accs)\n",
    "member_thresholds = np.asarray(member_thresholds)\n",
    "member_accs = np.asarray(member_accs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------------\n",
      "$0.162_{\\pm 0.007}$\n",
      "Group 0\n",
      "$0.147_{\\pm 0.009}$\n",
      "Group 1\n",
      "$0.216_{\\pm 0.022}$\n",
      "------------------------------\n",
      "$0.138_{\\pm 0.014}$\n",
      "Group 0\n",
      "$0.124_{\\pm 0.014}$\n",
      "Group 1\n",
      "$0.231_{\\pm 0.038}$\n",
      "------------------------------\n"
     ]
    }
   ],
   "source": [
    "ensemble_best_thresholds = np.asarray(ensemble_best_thresholds)\n",
    "member_best_thresholds = np.asarray(member_best_thresholds)\n",
    "\n",
    "print(\"-\"*30)\n",
    "for thrs in [ensemble_best_thresholds, member_best_thresholds]:\n",
    "    print(f\"${np.mean(thrs[:, -1]):.3f}_{'{'}\\pm {np.std(thrs[:, -1]):.3f}{'}'}$\")\n",
    "    for i in range(2):\n",
    "        print(f\"Group {i}\")\n",
    "        print(f\"${np.mean(thrs[:, i]):.3f}_{'{'}\\pm {np.std(thrs[:, i]):.3f}{'}'}$\")\n",
    "    print(\"-\"*30)  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 3)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 700x300 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_both = False\n",
    "\n",
    "f, axs = plt.subplots(1, 2, figsize=(7, 3), sharey=True)\n",
    "\n",
    "axs[0].plot(thresholds, np.mean(ensemble_accs[:, 0], axis=0), label=f\"$A=0$ ({(100 - p_a_val):.0f}%)\", color=\"C4\")\n",
    "axs[0].fill_between(thresholds, \n",
    "                    np.mean(ensemble_accs[:, 0], axis=0) - np.std(ensemble_accs[:, 0], axis=0), \n",
    "                    np.mean(ensemble_accs[:, 0], axis=0) + np.std(ensemble_accs[:, 0], axis=0), \n",
    "                    alpha=0.3, color=\"C4\")\n",
    "axs[0].plot(thresholds, np.mean(ensemble_accs[:, 1], axis=0), label=f\"$A=1$ ({(p_a_val):.0f}%)\", color=\"C2\")\n",
    "axs[0].fill_between(thresholds,\n",
    "                    np.mean(ensemble_accs[:, 1], axis=0) - np.std(ensemble_accs[:, 1], axis=0), \n",
    "                    np.mean(ensemble_accs[:, 1], axis=0) + np.std(ensemble_accs[:, 1], axis=0), \n",
    "                    alpha=0.3, color=\"C2\")\n",
    "if plot_both:\n",
    "    axs[0].plot(thresholds, np.mean(ensemble_accs[:, 2], axis=0), label=\"$Both$\", color=\"black\")\n",
    "    axs[0].fill_between(thresholds,\n",
    "                        np.mean(ensemble_accs[:, 2], axis=0) - np.std(ensemble_accs[:, 2], axis=0), \n",
    "                        np.mean(ensemble_accs[:, 2], axis=0) + np.std(ensemble_accs[:, 2], axis=0), \n",
    "                        alpha=0.3, color=\"black\")\n",
    "axs[1].plot(thresholds, np.mean(member_accs[:, 0], axis=0), label=f\"$A=0$ ({(100 - p_a_val):.0f}%)\", color=\"C4\")\n",
    "axs[1].fill_between(thresholds, \n",
    "                    np.mean(member_accs[:, 0], axis=0) - np.std(member_accs[:, 0], axis=0), \n",
    "                    np.mean(member_accs[:, 0], axis=0) + np.std(member_accs[:, 0], axis=0), \n",
    "                    alpha=0.3, color=\"C4\")\n",
    "axs[1].plot(thresholds, np.mean(member_accs[:, 1], axis=0), label=f\"$A=1$ ({(p_a_val):.0f}%)\", color=\"C2\")\n",
    "axs[1].fill_between(thresholds,\n",
    "                    np.mean(member_accs[:, 1], axis=0) - np.std(member_accs[:, 1], axis=0), \n",
    "                    np.mean(member_accs[:, 1], axis=0) + np.std(member_accs[:, 1], axis=0), \n",
    "                    alpha=0.3, color=\"C2\")\n",
    "if plot_both:\n",
    "    axs[1].plot(thresholds, np.mean(member_accs[:, 2], axis=0), label=\"$Both$\", color=\"black\")\n",
    "    axs[1].fill_between(thresholds,\n",
    "                        np.mean(member_accs[:, 2], axis=0) - np.std(member_accs[:, 2], axis=0), \n",
    "                        np.mean(member_accs[:, 2], axis=0) + np.std(member_accs[:, 2], axis=0), \n",
    "                        alpha=0.3, color=\"black\")\n",
    "    \n",
    "y_min = 0.7\n",
    "y_max = plt.gca().get_ylim()[1]\n",
    "\n",
    "print(ensemble_best_thresholds.shape)\n",
    "axs[0].vlines(np.mean(ensemble_best_thresholds, axis=0)[0], y_min, y_max, color=\"C4\", linestyle=\"--\")\n",
    "# fill between two vlines\n",
    "axs[0].axvspan(np.mean(ensemble_best_thresholds, axis=0)[0] - np.std(ensemble_best_thresholds, axis=0)[0],\n",
    "                np.mean(ensemble_best_thresholds, axis=0)[0] + np.std(ensemble_best_thresholds, axis=0)[0],\n",
    "                color=\"C4\", alpha=0.2)\n",
    "axs[0].vlines(np.mean(ensemble_best_thresholds, axis=0)[1], y_min, y_max, color=\"C2\", linestyle=\"--\")\n",
    "axs[0].axvspan(np.mean(ensemble_best_thresholds, axis=0)[1] - np.std(ensemble_best_thresholds, axis=0)[1],\n",
    "                np.mean(ensemble_best_thresholds, axis=0)[1] + np.std(ensemble_best_thresholds, axis=0)[1],\n",
    "                color=\"C2\", alpha=0.2)\n",
    "axs[1].vlines(np.mean(member_best_thresholds, axis=0)[0], y_min, y_max, color=\"C4\", linestyle=\"--\")\n",
    "axs[1].axvspan(np.mean(member_best_thresholds, axis=0)[0] - np.std(member_best_thresholds, axis=0)[0],\n",
    "                np.mean(member_best_thresholds, axis=0)[0] + np.std(member_best_thresholds, axis=0)[0],\n",
    "                color=\"C4\", alpha=0.2)\n",
    "axs[1].vlines(np.mean(member_best_thresholds, axis=0)[1], y_min, y_max, color=\"C2\", linestyle=\"--\")\n",
    "axs[1].axvspan(np.mean(member_best_thresholds, axis=0)[1] - np.std(member_best_thresholds, axis=0)[1],\n",
    "                np.mean(member_best_thresholds, axis=0)[1] + np.std(member_best_thresholds, axis=0)[1],\n",
    "                color=\"C2\", alpha=0.2)\n",
    "\n",
    "axs[1].legend(fontsize=\"x-small\")\n",
    "axs[0].set_title(\"Ensemble\")\n",
    "axs[1].set_title(\"Individual Members\")\n",
    "axs[0].set_xlabel(\"Threshold\")\n",
    "axs[1].set_xlabel(\"Threshold\")\n",
    "axs[0].set_ylabel(\"Balanced Accuracy ($\\\\rightarrow$)\")\n",
    "plt.ylim(y_min, y_max)\n",
    "plt.tight_layout()\n",
    "plt.savefig(os.path.join(PLOTS_PATH, f\"n_members_medical_imaging_pa{pa}_thresholds.pdf\"))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "quam",
   "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": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}