notebooks/notebook_fair-profession_fasttext_kSAL.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from scipy import linalg\n",
    "from sklearn.metrics.pairwise import rbf_kernel, polynomial_kernel\n",
    "import random\n",
    "import time\n",
    "import copy\n",
    "from collections import defaultdict, Counter\n",
    "\n",
    "from typing import List\n",
    "import pickle\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.stats.stats import pearsonr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_TPR(y_pred, y_true, p2i, i2p, y_p_test):\n",
    "    gender = np.where(y_p_test == 1, 'f', 'm')\n",
    "    scores = defaultdict(Counter)\n",
    "    prof_count_total = defaultdict(Counter)\n",
    "    \n",
    "    for y_hat, y, g in zip(y_pred, y_true, gender):\n",
    "        \n",
    "        if y == y_hat:\n",
    "            \n",
    "            scores[i2p[y]][g] += 1\n",
    "        \n",
    "        prof_count_total[i2p[y]][g] += 1\n",
    "    \n",
    "    tprs = defaultdict(dict)\n",
    "    tprs_change = dict()\n",
    "    tprs_ratio = []\n",
    "    \n",
    "    for profession, scores_dict in scores.items():\n",
    "        \n",
    "        good_m, good_f = scores_dict[\"m\"], scores_dict[\"f\"]\n",
    "        prof_total_f = prof_count_total[profession][\"f\"]\n",
    "        prof_total_m = prof_count_total[profession][\"m\"]\n",
    "        if prof_total_m > 0 and prof_total_f > 0:\n",
    "            tpr_m = (good_m) / prof_total_m\n",
    "            tpr_f = (good_f) / prof_total_f\n",
    "\n",
    "            tprs[profession][\"m\"] = tpr_m\n",
    "            tprs[profession][\"f\"] = tpr_f\n",
    "            tprs_ratio.append(0)\n",
    "            tprs_change[profession] = tpr_f - tpr_m\n",
    "        else:\n",
    "            print(f\"profession {profession} missed in tpr-calc:\\n\\\n",
    "                  number of {profession} male in test set {prof_total_m}\\n\\\n",
    "                  number of {profession} female in test set {prof_total_f}\\n\\n\")\n",
    "        \n",
    "    return tprs, tprs_change, np.mean(np.abs(tprs_ratio))\n",
    "    \n",
    "def similarity_vs_tpr(tprs, word2vec, title, measure, prof2fem):\n",
    "    \n",
    "    professions = list(tprs.keys())\n",
    "    #\n",
    "    \"\"\" \n",
    "    sims = dict()\n",
    "    gender_direction = word2vec[\"he\"] - word2vec[\"she\"]\n",
    "    \n",
    "    for p in professions:\n",
    "        sim = word2vec.cosine_similarities(word2vec[p], [gender_direction])[0]\n",
    "        sims[p] = sim\n",
    "    \"\"\"\n",
    "    tpr_lst = [tprs[p] for p in professions]\n",
    "    sim_lst = [prof2fem[p] for p in professions]\n",
    "\n",
    "    #professions = [p.replace(\"_\", \" \") for p in professions if p in word2vec]\n",
    "    \n",
    "    plt.plot(sim_lst, tpr_lst, marker = \"o\", linestyle = \"none\")\n",
    "    plt.xlabel(\"% women\", fontsize = 13)\n",
    "    plt.ylabel(r'$GAP_{female,y}^{TPR}$', fontsize = 13)\n",
    "    for p in professions:\n",
    "        x,y = prof2fem[p], tprs[p]\n",
    "        plt.annotate(p , (x,y), size = 7, color = \"red\")\n",
    "    plt.ylim(-0.4, 0.55)\n",
    "    z = np.polyfit(sim_lst, tpr_lst, 1)\n",
    "    p = np.poly1d(z)\n",
    "    plt.plot(sim_lst,p(sim_lst),\"r--\")\n",
    "    plt.savefig(\"{}_vs_bias_{}_bert\".format(measure, title), dpi = 600)\n",
    "    print(\"Correlation: {}; p-value: {}\".format(*pearsonr(sim_lst, tpr_lst)))\n",
    "    plt.show()\n",
    "\n",
    "def rms_diff(tpr_diff):\n",
    "    \n",
    "    return np.sqrt(np.mean(tpr_diff**2))\n",
    "    \n",
    "# def save_vecs_and_words(vecs, words):\n",
    "#     def to_string(arr):\n",
    "#         return \"\\t\".join([str(x) for x in arr])\n",
    "    \n",
    "#     with open(\"vecs.txt\", \"w\") as f:\n",
    "#         for v in vecs:\n",
    "#             assert len(v) == 300\n",
    "#             f.write(to_string(v) + \"\\n\")\n",
    "    \n",
    "#     with open(\"labels.txt\", \"w\") as f:\n",
    "#             f.write(\"Profession\\n\")\n",
    "#             for w in words:\n",
    "#                 f.write(w + \"\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_dataset(path):\n",
    "    \n",
    "    with open(path, \"rb\") as f:\n",
    "        \n",
    "        data = pickle.load(f)\n",
    "    return data\n",
    "\n",
    "def load_dictionary(path):\n",
    "    \n",
    "    with open(path, \"r\", encoding = \"utf-8\") as f:\n",
    "        \n",
    "        lines = f.readlines()\n",
    "        \n",
    "    k2v, v2k = {}, {}\n",
    "    for line in lines:\n",
    "        \n",
    "        k,v = line.strip().split(\"\\t\")\n",
    "        v = int(v)\n",
    "        k2v[k] = v\n",
    "        v2k[v] = k\n",
    "    \n",
    "    return k2v, v2k\n",
    "    \n",
    "def count_profs_and_gender(data: List[dict]):\n",
    "    \n",
    "    counter = defaultdict(Counter)\n",
    "    for entry in data:\n",
    "        gender, prof = entry[\"g\"], entry[\"p\"]\n",
    "        counter[prof][gender] += 1\n",
    "        \n",
    "    return counter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load original data (train.pickle etc.) and pre-calculated BERT CLS states (\"train_cls.npy\" etc.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.44541700643683746\n",
      "{'professor': 0.4533699535765501, 'psychologist': 0.627327974906881, 'pastor': 0.2583668005354752, 'comedian': 0.21605667060212513, 'nurse': 0.9126009126009126, 'yoga_teacher': 0.825, 'attorney': 0.3766122913505311, 'photographer': 0.3563658099222953, 'composer': 0.16329625884732052, 'model': 0.7858407079646018, 'surgeon': 0.12832108535895986, 'physician': 0.4158485273492286, 'software_engineer': 0.17130434782608694, 'poet': 0.5133630289532294, 'painter': 0.47116788321167885, 'dj': 0.1636828644501279, 'journalist': 0.5159589626674266, 'architect': 0.23372781065088757, 'paralegal': 0.8618677042801557, 'dentist': 0.3672911787665886, 'personal_trainer': 0.4410377358490566, 'teacher': 0.5943396226415094, 'accountant': 0.3950091296409008, 'interior_designer': 0.7874493927125507, 'dietitian': 0.934412955465587, 'filmmaker': 0.3533007334963325, 'chiropractor': 0.3069908814589666, 'rapper': 0.09438775510204081}\n"
     ]
    }
   ],
   "source": [
    "train = load_dataset(\"../data/biasbios/train.pickle\")\n",
    "dev = load_dataset(\"../data/biasbios/dev.pickle\")\n",
    "test = load_dataset(\"../data/biasbios/test.pickle\")\n",
    "counter = count_profs_and_gender(train+dev+test)\n",
    "prof2fem = dict()\n",
    "\n",
    "f,m = 0., 0.\n",
    "for k, values in counter.items():\n",
    "    f += values['f']\n",
    "    m += values['m']\n",
    "    prof2fem[k] = values['f']/(values['f'] + values['m'])\n",
    "\n",
    "print(f / (f + m))\n",
    "print(prof2fem)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "saved_dataset = np.load(f\"../data/saved_models/fair_biography_prof_gender/FastText/all.npz\")\n",
    "\n",
    "x_train = saved_dataset['x_train']\n",
    "y_m_train = saved_dataset['y_m_train']\n",
    "y_p_train = saved_dataset['y_p_train']\n",
    "\n",
    "x_dev = saved_dataset['x_dev']\n",
    "y_p_dev = saved_dataset['y_p_dev']\n",
    "y_m_dev = saved_dataset['y_m_dev']\n",
    "\n",
    "x_test = saved_dataset['x_test']\n",
    "y_p_test = saved_dataset['y_p_test']\n",
    "y_m_test = saved_dataset['y_m_test']\n",
    "\n",
    "y_p_train_2d = np.asarray([y_p_train, - y_p_train + 1]).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "p2i, i2p = load_dictionary(\"../data/biasbios/profession2index.txt\")\n",
    "g2i, i2g = load_dictionary(\"../data/biasbios/gender2index.txt\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Option 1: Train the model from scratch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# N = 15000\n",
    "# N_test = 5000\n",
    "N = 4000\n",
    "N_test = 1000\n",
    "\n",
    "ratio = \"FastText\"\n",
    "\n",
    "# kernel_fn = \"rbf\"\n",
    "# kernel_coef = 0.1\n",
    "kernel_fn = \"poly\"\n",
    "kernel_coef = 2\n",
    "\n",
    "assert x_train[:N, :].shape[1] == x_train[:N, :].shape[1]\n",
    "\n",
    "K_y = y_p_train_2d[:N, :] @ y_p_train_2d[:N, :].T\n",
    "if kernel_fn == \"linear\":\n",
    "    K_x = x_train[:N, :] @ x_train[:N, :].T\n",
    "    K_x_test_tr = x_test[:N_test, :] @ x_train[:N, :].T\n",
    "elif kernel_fn == \"rbf\":\n",
    "    K_x = rbf_kernel(x_train[:N, :], x_train[:N, :], kernel_coef)\n",
    "    K_x_test_tr = rbf_kernel(x_test[:N_test, :], x_train[:N, :], kernel_coef)\n",
    "elif kernel_fn == \"poly\":\n",
    "    n_features = x_train[:N, :].shape[1]\n",
    "    K_x = rbf_kernel(x_train[:N, :], x_train[:N, :], kernel_coef)\n",
    "    K_x_test_tr = polynomial_kernel(x_test[:N_test, :], x_train[:N, :], degree = kernel_coef, gamma = n_features, coef0=1)\n",
    "\n",
    "print(f\"K_x.shape, K_x_test_tr.shape {K_x.shape, K_x_test_tr.shape}\")\n",
    "eigenval, W = np.linalg.eig(np.dot(K_y, K_x))\n",
    "order = np.argsort(np.abs(np.real(eigenval)))\n",
    "W = W[:, order]\n",
    "W = np.real(W)\n",
    "A = np.concatenate((K_x, K_x_test_tr), axis=0)\n",
    "u, s, vh = np.linalg.svd(A)\n",
    "s = s**(0.5)\n",
    "s = np.diag(s)\n",
    "print(f\"A.shape, u.shape, s.shape, vh.shape {A.shape, u.shape, s.shape, vh.shape}\")\n",
    "\n",
    "K_x_sqrt_all = np.dot(np.dot(u[:, :s.shape[0]], s), vh[:s.shape[1], :])\n",
    "K_x_sqrt2 = K_x_sqrt_all[:N, :]\n",
    "K_x_test_sqrt2 = K_x_sqrt_all[N:, :]\n",
    "print(f\"K_x.shape, K_x_test_sqrt2.shape, K_x_sqrt_all.shape {K_x.shape, K_x_test_sqrt2.shape, K_x_sqrt_all.shape}\")\n",
    "\n",
    "remove_eigen_num = 2\n",
    "U_proj_ker = linalg.null_space((K_x_sqrt2 @ W[:, :remove_eigen_num]).T)\n",
    "debiased_x_train = np.dot(K_x_sqrt2, U_proj_ker)\n",
    "debiased_x_test = np.dot(K_x_test_sqrt2, U_proj_ker)\n",
    "print(f\"debiased_x_train.shape, debiased_x_test.shape {debiased_x_train.shape, debiased_x_test.shape}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "random.seed(0)\n",
    "np.random.seed(0)\n",
    "\n",
    "clf_original = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"saga\", multi_class = 'multinomial', fit_intercept = False,\n",
    "                         verbose = 5, n_jobs = 90, random_state = 1, max_iter = 7)\n",
    "                         \n",
    "\n",
    "clf_original.fit(x_train[:N, :], y_m_train[:N])\n",
    "print(f\"Score of profession classifier on original(biased) dataset \\n{clf_original.score(x_test[:N_test, :], y_m_test[:N_test])}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "clf_debiased = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"saga\", multi_class = 'multinomial', fit_intercept = False,\n",
    "                         verbose = 5, n_jobs = 90, random_state = 1, max_iter = 7)\n",
    "\n",
    "clf_debiased.fit(debiased_x_train, y_m_train[:N])\n",
    "print(f\"Score of profession classifier on debiased dataset \\n{clf_debiased.score(debiased_x_test, y_m_test[:N_test])}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_before = clf_original.predict(x_test[:N_test])\n",
    "tprs_before, tprs_change_before, mean_ratio_before = get_TPR(y_pred_before, y_m_test[:N_test], p2i, i2p, y_p_test[:N_test])\n",
    "similarity_vs_tpr(tprs_change_before, None, \"before\", \"TPR\", prof2fem)\n",
    "\n",
    "y_pred_after = clf_debiased.predict(debiased_x_test)\n",
    "tprs, tprs_change_after, mean_ratio_after = get_TPR(y_pred_after, y_m_test[:N_test], p2i, i2p, y_p_test[:N_test])\n",
    "similarity_vs_tpr(tprs_change_after, None, \"after\", \"TPR\", prof2fem)\n",
    "\n",
    "change_vals_before = np.array(list((tprs_change_before.values())))\n",
    "change_vals_after = np.array(list(tprs_change_after.values()))\n",
    "\n",
    "print(\"rms-diff before: {}; rms-diff after: {}\".format(rms_diff(change_vals_before), rms_diff(change_vals_after)))\n",
    "\n",
    "# np.savez(f\"../../data/saved_models/fair_biography_prof_gender/{ratio}/cleaned_Train{N}_Test{N_test}.npz\", debiased_x_train = debiased_x_train, debiased_x_test = debiased_x_test)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Option 2: load directly"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train and save the model by src/ksal/ksal.py \n",
    "\n",
    "Then load it to the notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tpr_exp(x_test, debiased_x_train, y_m_train, debiased_x_test, y_m_test, y_p_test, N_test, p2i, i2p, clf_original):\n",
    "    y_p_test_string = np.where(y_p_test == 1, 'f', 'm')\n",
    "\n",
    "    y_pred_before = clf_original.predict(x_test[:N_test])\n",
    "    tprs_before, tprs_change_before, mean_ratio_before = get_TPR(y_pred_before, y_m_test[:N_test], p2i, i2p, y_p_test[:N_test])\n",
    "    similarity_vs_tpr(tprs_change_before, None, \"before\", \"TPR\", prof2fem)\n",
    "\n",
    "    clf_debiased = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                             solver = \"saga\", multi_class = 'multinomial', fit_intercept = False,\n",
    "                             verbose = 5, n_jobs = 90, random_state = 1, max_iter = 7)\n",
    "\n",
    "    clf_debiased.fit(debiased_x_train, y_m_train[:N])\n",
    "    print(f\"\\n\\n\\n================================================\\\n",
    "                \\nSCORE: of profession classifier on debiased dataset\\\n",
    "                \\n{clf_debiased.score(debiased_x_test, y_m_test[:N_test])}\\\n",
    "                \\n================================================\\n\\n\\n\")\n",
    "\n",
    "    y_pred_after = clf_debiased.predict(debiased_x_test)\n",
    "    tprs, tprs_change_after, mean_ratio_after = get_TPR(y_pred_after, y_m_test[:N_test], p2i, i2p, y_p_test[:N_test])\n",
    "    similarity_vs_tpr(tprs_change_after, None, \"after\", \"TPR\", prof2fem)\n",
    "\n",
    "    change_vals_before = np.array(list((tprs_change_before.values())))\n",
    "    change_vals_after = np.array(list(tprs_change_after.values()))\n",
    "\n",
    "    print(\"rms-diff before: {}; rms-diff after: {}\".format(rms_diff(change_vals_before), rms_diff(change_vals_after)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "N = 15000\n",
    "N_test = 5000\n",
    "ratio = \"FastText\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 5 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:    5.4s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Score of profession classifier on original(biased) dataset \n",
      "0.7022\n"
     ]
    }
   ],
   "source": [
    "random.seed(0)\n",
    "np.random.seed(0)\n",
    "\n",
    "clf_original = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"saga\", multi_class = 'multinomial', fit_intercept = False,\n",
    "                         verbose = 5, n_jobs = 90, random_state = 1, max_iter = 7)\n",
    "                         \n",
    "\n",
    "clf_original.fit(x_train[:N, :], y_m_train[:N])\n",
    "print(f\"Score of profession classifier on original(biased) dataset \\n{clf_original.score(x_test[:N_test, :], y_m_test[:N_test])}\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Linear Kernel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "kernel_fn = \"linear\"\n",
    "kernel_coef = 0\n",
    "\n",
    "saved_model = np.load(f\"../data/saved_models/fair_biography_prof_gender/{ratio}/cleaned_{kernel_fn}_{kernel_coef}_Train{N}_Test{N_test}.npz\")\n",
    "debiased_x_train_0 = saved_model['debiased_x_train_0']\n",
    "debiased_x_test_0 = saved_model['debiased_x_test_0']\n",
    "debiased_x_train_1 = saved_model['debiased_x_train_1']\n",
    "debiased_x_test_1 = saved_model['debiased_x_test_1']\n",
    "debiased_x_train_2 = saved_model['debiased_x_train_2']\n",
    "debiased_x_test_2 = saved_model['debiased_x_test_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removal 0 BERT Linear\n",
    "tpr_exp(x_test, debiased_x_train_0, y_m_train, debiased_x_test_0, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removal 1 BERT Linear\n",
    "tpr_exp(x_test, debiased_x_train_1, y_m_train, debiased_x_test_1, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.773655072810786; p-value: 6.29340039229481e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 298 seconds\n",
      "\n",
      "\n",
      "\n",
      "================================================                \n",
      "SCORE: of profession classifier on debiased dataset                \n",
      "0.6974                \n",
      "================================================\n",
      "\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:  5.0min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.28151941939644465; p-value: 0.21635439221736918\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rms-diff before: 0.18337572303336608; rms-diff after: 0.11473460987575866\n"
     ]
    }
   ],
   "source": [
    "# Removal 2 BERT Linear\n",
    "tpr_exp(x_test, debiased_x_train_2, y_m_train, debiased_x_test_2, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. RBF kernel at gamma = 0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "kernel_fn = \"rbf\"\n",
    "kernel_coef = 0.1\n",
    "\n",
    "saved_model = np.load(f\"../data/saved_models/fair_biography_prof_gender/{ratio}/cleaned_{kernel_fn}_{kernel_coef}_Train{N}_Test{N_test}.npz\")\n",
    "debiased_x_train_0 = saved_model['debiased_x_train_0']\n",
    "debiased_x_test_0 = saved_model['debiased_x_test_0']\n",
    "debiased_x_train_1 = saved_model['debiased_x_train_1']\n",
    "debiased_x_test_1 = saved_model['debiased_x_test_1']\n",
    "debiased_x_train_2 = saved_model['debiased_x_train_2']\n",
    "debiased_x_test_2 = saved_model['debiased_x_test_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removal 0 BERT RBF at gamma = 0.1\n",
    "tpr_exp(x_test, debiased_x_train_0, y_m_train, debiased_x_test_0, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Removal 1 BERT RBF at gamma = 0.1\n",
    "tpr_exp(x_test, debiased_x_train_1, y_m_train, debiased_x_test_1, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.773655072810786; p-value: 6.29340039229481e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 287 seconds\n",
      "\n",
      "\n",
      "\n",
      "================================================                \n",
      "SCORE: of profession classifier on debiased dataset                \n",
      "0.1372                \n",
      "================================================\n",
      "\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:  4.8min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.450143871049959; p-value: 0.04059786566020312\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rms-diff before: 0.18337572303336608; rms-diff after: 0.08420485250044037\n"
     ]
    }
   ],
   "source": [
    "# Removal 2 BERT RBF at gamma = 0.1\n",
    "tpr_exp(x_test, debiased_x_train_2, y_m_train, debiased_x_test_2, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. Polynomial kernel at degree=2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "kernel_fn = \"poly\"\n",
    "kernel_coef = 2\n",
    "\n",
    "saved_model = np.load(f\"../data/saved_models/fair_biography_prof_gender/{ratio}/cleaned_{kernel_fn}_{kernel_coef}_Train{N}_Test{N_test}.npz\")\n",
    "debiased_x_train_0 = saved_model['debiased_x_train_0']\n",
    "debiased_x_test_0 = saved_model['debiased_x_test_0']\n",
    "debiased_x_train_1 = saved_model['debiased_x_train_1']\n",
    "debiased_x_test_1 = saved_model['debiased_x_test_1']\n",
    "debiased_x_train_2 = saved_model['debiased_x_train_2']\n",
    "debiased_x_test_2 = saved_model['debiased_x_test_2']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.773655072810786; p-value: 6.29340039229481e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 298 seconds\n",
      "\n",
      "\n",
      "\n",
      "================================================                \n",
      "SCORE: of profession classifier on debiased dataset                \n",
      "0.4952                \n",
      "================================================\n",
      "\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:  5.0min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correlation: 0.576632539140049; p-value: 0.08098230117672431\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rms-diff before: 0.18337572303336608; rms-diff after: 0.04961962564035084\n"
     ]
    }
   ],
   "source": [
    "# Removal 0 BERT Poly 2\n",
    "tpr_exp(x_test, debiased_x_train_0, y_m_train, debiased_x_test_0, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.773655072810786; p-value: 6.29340039229481e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZsAAAEJCAYAAABCNoqwAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAA4mklEQVR4nO3de5zU8/7A8de7mzbKSql0UaicELFKbrl0JST9jltU5NYhRIQc4VBEiRxOiDguOUokRJRLrptKJd1UtElRidpq233//njPtLOzs9tuzXdmtt7Px2MfO/P9fuY77x2Z9/dzF1XFOeecC1K5ZAfgnHNu9+fJxjnnXOA82TjnnAucJxvnnHOB82TjnHMucBWSHUAqqFGjhjZs2DDZYTjnXJkxY8aM31S1ZknLe7IBGjZsSGZmZrLDcM65MkNElpemvDejOeecC5wnG+ecc4HzZOOccy5wnmycc84FzpONc865wHmycc45FzhPNs455wLnycY551zgPNk455wLnCcb55xzgfNk45xzLnCebJxzzgXOk41zzrnAebJxzjkXOE82zjnnAufJxjnnXOA82TjnnAucJxvnnHOB82TjnHMucJ5snHPOBc6TjXPOucB5snHOORc4TzbOOecC58nGOedc4DzZOOecC5wnG+ecc4HzZOOccy5wnmycc84FzpONc865wHmycc45F7gyl2xEpKOILBCRxSIyoJhy54uIikhGIuNzzjlXWJlKNiJSHngC6AQ0Ay4SkWYxylUFbgC+SmyEzjnnYilTyQZoCSxW1R9VdSvwKnBujHL3AQ8CmxMZnHPOudjKWrKpC/wc8XxF6Nh2InIMUF9VJxV3IRG5SkQyRSRzzZo18Y/UOefcdmUt2RRLRMoBw4Cbd1RWVUepaoaqZtSsWTP44Jxzbg9W1pJNFlA/4nm90LGwqsARwDQRWQYcD7zlgwSccy65ylqy+QZoLCKNRKQScCHwVvikqv6hqjVUtaGqNgS+BM5R1czkhOuccw7KWLJR1W3AdcBkYD7wmqrOE5F7ReSc5EbnnHOuKBWSHUBpqeo7wDtRx/5ZRNlTExGTc86VOar2WyQhb1emajbOOed20Z9/QuvWUK6c/SRImavZOOec2wkLFlgtJiPDEg7AZZcl7O29ZuOcc7ur7Gy48EJLMocdBsuXw9VXwyefQF4ejBmTsFC8ZuOcc7ubRYugadP8fhmAwYOhRQto1y4pIXnNxjnnEqlbt9IdLylVWLEC/v1vaNIkP9Fccw1s3QoDBkCNGrv2HrvAazbOObeznn8epk6FtDSoUwdycmDuXOjVCyZOhM2bYdAg2LYN7r7bksBff9lr778ffvvN+k+GD9/5GLKyoH17+P57qFcPxo2DW26B7t3hqKPi8VfGhScb55wrhQkzsxg6eQEr12fTe8n3dGhxHBl3XAdnnAEffggPPACPPWaPly+Hxx+3WsaQIVC/viWG+fOt36R1a0tI8+eXPpA334QuXQoeu/deaypr2TIuf2s8ebJxzrkSmjAzi9vHzyE7JxeAdZtyePa731kxM4su4TUWK1XKf4FIfnNWpUpQvjxUqGCd84cfbrWe0sjJsU7/Sy+Ft97KP/7EE9Cnz87/YQngycY550po6OQF2xNN2JZteQydvIAukQfbtrVRX5s2wV13QW6uNZsdfLCdP/xwm+PSr58ljzvuKP6Nv/7aOvarVLFEc/zxcMIJVrNp2jSOf2FwRCNHK+yhMjIyNDPTl09zzhWv0YBJxPrGFGDpkLPi+2aqVvO59978Y/XqwXPPWTJLMhGZoaolXuTYazbOOVdCB6ankbU+O+bxuNm4EVauhHPOgR9+yD8+ZYr1C5VRPvTZOedKqH+HpqRVLF/gWFrF8vTvEIemrJdfhkaNoFYt+OMPe/zgg7B2rdVyynCiAa/ZOOdciXVpYRsDh0ejHZieRv8OTbcfL7WNG+GSS2xkWVjPnpCeDu+8U9SryiRPNs45VwpdWtTd+eQSlpMDr70GV15pAwQAate24dCNG+96kCnIm9Gccy4R8vLg9tttDszxx8N++1nT2MMPW/L55ZfUTjR5ebv0cq/ZOOdckJYvh9NPhx9/zD92222WaM48MykhFZyY+jHdtyznoOOOsOHYM2fC66/De+/BqlX2gilTbLXo3Fz7e/bd146LXAs0AdKBu1BdUdR7es3GOeeCkJcH110HDRvmJ5p27WDdOltNYK+9khJWeGJq1vpsFJuYOqLCwUzofLkttRNLp05w442wbJklnb59qWb54zLgD2A9cExx7+s1G+eci5c//4Trr4fFi23CZXgAQJs21j+TAmJNTF1frpJNTFXN31Bt48b8AuGazIgR8M030KsXFWx6URaqg0ryvp5snHNuV02fDiedVPDYVVfBzTcnbNvlkloZY55QgePNm9v6bkuWwMknFyz00EO2eGj16myAXOBrRB7HEs9oVL8t6n19BQF8BQHn3E5Qtf6LG28sOHT5pptg6FBbBy0FnTjko5gTU+umpzF9wOklvk5pVxDwPhvnnCuNzz6z2kq5ctbUNGsWdO4M06ZZAho2LGUTDQQ8MbUY3ozmnHMlceWV8MwzBY/l5Fjnf7myc98e94mpJeTJxjnnirJliyWZF18sePzsswsu8V/GxGViail5snHOuWiDB8de9n/GDDim2BG+rgiebJxzDmDrVhuJtWBBweOvv27DmFO4H6YsKDsNjc45F4S1a6FBA5tkGZlohgyxDv/zz/dEEwdes3HO7Zk6d4ZJkwofX7MGatRIfDy7OU82zrk9x6+/2urK0aZPt22WXWC8Gc05t/v7+mubGxOdaD77zJrKPNEEzms2zrnd05YtUK+eLa8SqWZNW87f+2ESyms2zrndS7gWU7lyfqJJT7eZ/qqwerUnmiTwmo1zbvcwejRccUXh47//DtWrJz4eV4DXbJxzZdeyZVaLESmYaG691Woxqp5oUoQnG+dcQc8/D2+/XfrXTZsGI0eWrOwubjHMs89agmnUKP/Y3XfDTz9ZgnnwwV27vos7b0ZzzgH5WwX3feUV9pNtVFv8G8cvmwXZ2TaxsW5deOklGz7cuzccdxz06QNVq8KRR9qOlO+/D0uXWuf8yJEwdix8+SVs2GCbik2YYJMoW7SAXr1KF2Buri3f//jjBY9XrGgxej9MSvOajStat272u3//5MbhAhe5VfA39Zrx6qEnkf3vJ/khuxzUqWOd7pUqwebNUKuWLUz5/vuWcB55BHr2tAu1amXP16yx5yNH2i6PBxxg1wC44ILSJZpPPrFaTIUKBRPNuHFWi9m61RNNGRCXmo2IZKhq4LuPiUhHYARQHnhGVYdEne8H9Aa2AWuAy1V1edBxJVP4blSWL2fgly9z6BEHc+jVl8KoUVagTx/7ghg+HA45xO4C09IgMxOeew5WrLBmk23boHVraN/e7kAPPdTOgd2pgn1xLFli/2M//LB9wTRuDFlZtnZU+/ZJ+ARcPERuFZwX2llSc/O4qkEnPrmznRW6/nq78VC1JqvILYTDwtsHh3enTEuDQYPyzw8alF9mRy6/3P6NRnrgAejbF/beu+R/nEsJ8WpGu1FENgO3q+qaOF2zABEpDzwBtANWAN+IyFuq+n1EsZlAhqpuEpFrgYeAC4KIJxWE70azc3IZ8O3bDDnu//i1VgPef+AR6o9/xQpdeaV1nLZqZavYtmtnd6SvvGIT2saPtyQEMHMmrFsH11wDp51md5SRcnOhShVrm1+92o717m3HbrnFk00ZFrlV8PwDGvGPz19j/BGn0/fVB+H3dyEjw/5NPPig1WzA/nv/4x+wcKEtYFk3xpL13bvb9shpaXDWWTsOZN066NrV/o1FuvpqeOqpnf8DXdLFJdmoancRORn4n4i8CTymqrnxuHaElsBiVf0RQEReBc4FticbVZ0aUf5LoHucY0gpkXejAuRJObJzcpm/cgP1w3eW4W2/q1Wz3/vvb3edlSpZu/rWrXDDDbDffnb+8cftHNjChGG//w6zZ9vw0l69YNMmO7733ta8sWVLsH+sC9SB6Wnbtwqef8DBXNdlAAAzW3egW+RWwV27FnxhuAYdduqp9vvVV+33ZZfZT1hRNyRPPQXXXlv4+Pffw9/+VsK/wqWyuA0QUNVPRaQr8B/gShHpq6pT4nV9oC7wc8TzFUCrYspfAbxb1EkRuQq4CqBBgwbxiC/hIu9GX2xxJjdNf5nVe1fnP0d0ov0119iJPn0soRTlttuseaRWLevgveACGDjQ2tfXr88vl55uCebhh+1O1u1W+ndour2WHBb4VsFbtlgz7M8/Fzz+0ENw881lavdLt2Oi4TvfXbmIyAtAc+Av4Cvgc+BU4A9VHbjLb2Dv0Q3oqKq9Q88vBVqp6nUxynYHrgPaqOoOb7kzMjI0MzPwLqe4O3HIR9vvRiPVTU9jeuTdqHMlEO7/C3yr4AUL4LDDCh9/800455z4v58LhIjMUNWMkpaPV83mKWBG1Bf7OBF5I07XB8gC6kc8rxc6VoCItAXupISJpixLyt2o220FulWwKlx6qQ2djvbEE7afTOfORb++Z08boLLPPjt+r/79YejQgseef962DSjuPVyg4tVn83kRpwrVOnbBN0BjEWmEJZkLgYsjC4hIC6wZr6Oqro7je6ek8BdDQu5GnYshXBtq/elEzvhlHg1OzuDwevvBN99Ys2zLlvB//1fwRe3b2wjJjAxo08bm3bz2Grz3njXn3nUX3HOPDRb44w/4z3/yX/vmmzBxor1+0CDIybGyTZpY0+9779noyTVrbPBCw4aWYD77LL+f0RNOUgQ6qVNVC9U8duFa20TkOmAyNvR5tKrOE5F7gUxVfQsYCuyDDVQA+ElVd+t6eaB3o84VI3I0JMDkus2ZXO0Upkx+nLrbNsL//lf4RU89ZaMaGze2kZFgtY4OHeCSS6zPcMMGW4bmuedgzBgbPRn2wgs2v2b5chvMEl4toG5du0bYH3/YEP0uXWxI/48/es0myeKebESkPpYQJqlqXGcDquo7wDtRx/4Z8bhtPN/POVe0yNGQAJVztvDx8Is5YOO6ggXPOMNqGBs22FD8J5/c8fycMJH8EZXFHQ+vjxZ26KHw2GM2tP+DDwoua+OSIu7JRlV/BpqJyIHxvrZzLnWER0O2+mkOD7/zaMGT9epZE9pFF9lKFK++anvIlCtnNZgbb7QkcPLJhS9crRocdJCNSFu71prRxo2zc92725ybTZusuS03FwYMsGa0yImec+bY+mmbN0PbtlaTuv9+m7zcpUsQH4fbgbiMRgMQkYOAcqq6NC4XTKCyOhrNuaRR5cc6B3Pwr8sKnWo3aCIf3B3RXDV9OowYYV/+VavGN461a+HRR20e2BlnFJ4H5AKTlNFoIvII0Ag4RESOAbqp6th4XNs5l0J++GH7JMuDIw4/euJFPHrSJaRVLM/gc44s+JoTT7SfIFSvDvfeG8y1XVzFqxntOFU9RUSmqmquiFwBeLJxbnfRoUPBjvqQqS+/x8DlFVm5Ppu6PhrSFSNeyWariFQGwm1yleJ0XedcsmRn27p30dLTrflKhNOA6YmOy5VJ8VoP4mHgfaCWiFwYp2s655LhjTdsZFd0onngARsBtm5d4VFjzu3ALtVsRKQX8C0wBVgEXAQ0Ay7d9dCccwkTa7uAsHXrrDbj3C7Y1ZrN3kBf4DPgv0BD4FcKLivjnEtVP/xgtZToRNOnjyUgVU80Li52qWajqts3HBeRisDhQAtsD5milrBxziXb8cfDV18VPv7GGz4PxQUinlsM5ACzQj/OuVSzapVt8RytcWOr4fiS/i5AcfnXJSIPiMgsEZktIi+KyA3xuK5zLg6ee86ayqITzYMPWjPZwoWeaFzg4lWz6aCqR4vIVGA8cF6cruuc2xl5eTZbP7zScaRffoHatRMfk9ujxet25rfQ7zxVfYOAV5N2zkUZNAjmzoUJE6wWU758wUTTtWt+h78nGpcE8UoK74jI3sBCEemBDRJwzpVC5E6ZvZd8zKWbl9LgpAxLHD/9BBUq2EKSjz0GY8fCl1/aSsrXXw+DB9u+LtFmzIBjjkn8H+NclLjUbFR1hKpuBAZg82xujcd1ndtThPeGyVqfjQLrNuUwsnwjJrTvbkll331h2DDYf3+rwYwcCRUrwujR0KIFbN1a8IJ5eVaL8UTjUkSRyUZEaovI9SLSIeJYxcjnEcerisgJQGVVvU1VJwYUr3O7pei9YQDycnIYOnmBJY3wjH0RuOUW23kyeuvjf/wjv6nMZ/i7FFNcM9p4bNmjc0TkCOAYoDXwBbY5WqRPgO+AmiJSG8hS1bMDiNe53VJ4b5hIJy+dyWGrl0K31taMVlQCmTjRmtpefRU+/ti2WnYuxRSXbHJUtb+I7AX8AlyiqpcUUXaNqvYIP/GN05wrnQPT08iKSjhvNmvDXjWq03vAjbFfFL0XVadOwQTnXBwUl2z2FZFDVHWJiMxW1XeLKfuSiJwdbj5T1ZXxDdO53Vv/Dk25ffyc7U1phXa+DHv/fWjXLnGBORcnxSWbD4AnReRgoIaIPAHMAeaq6mdRZc8GjhaRf2ELcmaq6pBAInZuN9SlRV3SVq2kw5ktYxfIy/N+GFemFTca7V/A68AjWDJ5G0gH+sQoW19VD1XVo7ARacvjHKdzsXXrtuMy06bZ6C2Ap56CJUuKvk5Jrhepf//SHY91/pxzQKRworn++tgd/nHayt25RCquZjMRmIdN2OwN3F1MbWWKiDRU1WWquhhYHOc4nds+D6XykoXc8dVY6rU6iqZ//QX33w+//QZ//gnDh8NFF8Gpp9oQ4X79bOTW55/b6sWrVvHht8v4vu9Qqv66kkM3/sYhB+xDnfmzoH17Wz/s9tvh+++tw/3aa+HJJ+GPP+x9Wre25FW1qu3vsnSpBffCCzB9OqSlwcMPw9Sp0KMHrF9vS/Tn5kJGhsXStSuMG2flYjn9dOjVCw4/HG64webWtG4N9erZCLQTT4TbbrNBAc6VEcUlG1HVawFEZCi2zfPbRZRtD/QWkQ+Br7FmtOimNud2WngeSnZOLnfOnszdJ3Rn/f61mTrra2p+8ol9GW/eDPPn2wv69YN58+Dtt+GkkyzRdO/OD1f3Y8Qvi+i6Kovv6jRm6ZY6vLX3vtzQvCJ1q1e1WsPgwfDoozBrFlx2Gbz0EqxcCVdfbasiN28O554Le+2VH+AbbzDhnyNtUubAybyz6k82djmWjKpqtamNG6FVK1vwcuDA2H/kAw/Y3zF3LnTvDpdeCoccYudmzrRkc8IJcMcdAX7SzgWjuGRTVUSOBJao6gYRKXKrZ1U9TkQOBaoBRwH/h+1x41xcRM9DySlfgb+2KcvWb6Hmycfbci1hlSvbwpIVK8KWLQUWmZy+5Hc2H5TLPW2v4qhfFvHAeyN56ehOLPj1L+ruW9lqLOFrbNlitYznnrMkdMQR9jN7tjWD3Xff9uuu/CO7QAf/1m15vPvZD9Tcay0H1a5ttaHhwwv/Ya1a2aTNH36wmlTkgphbt1rNZr/97Pm0aTa507kyaEcDBB4BjhKRHKCKiNwHfKKqH0QWFJFHgEbAIdh8nMKTBpzbBZHzUF5t3oHrvniNn/atzeZtoR0m+/WD7OzYd/1Nmtjs+7335s/NOQBc89U49tu0gU0V96LVirnU/v1n2F9ibxR2+OHQtKk9HjUKFi2y99x//+1FnqvenNsnjWRTxb0Y2sZmAZy84CsOWvZt7D+oSxdLHH/9VWSs3Hab9dvUqgUNG8KRR5bw03Iu9YgW0dkoIvup6rrQ4zrA0VitpbmqXhxV9hNVPUVEpqrqaSLyvqq2Dzj2uMnIyNDMzMxkh+GKceKQjwrNQwGom57G9AGn7/R1us2Zwtq0aizIaBP7OhMnwmuvwZgxxS7D32jAJML/J3372MVUz95QuNDAgQVqQ86VZSIyQ1UzSlq+uJrNOOB0AFX9BZvYWdRcm60iUhm2//9WZJObczsjeh4KQFrF8vTv0HSXrvP6kW1Jq1iewUVd5+yz7WcH6letyCcDC63kZFavhpo1SxWnc7ub4oY+l2ZQ/8PA+0AtEblw10JyrrAuLeoyuOuR1E1PQ7AazeCuR9KlRd2kXGe7f/4TRGImmr8NfJcJ367wROMcxTej/QI8DcwGvlPVRTHKtFLVr0KPDwEuAvYCRqnqz4FFHWfejOZKrYgJltdcPYLJ6YdwYHoa/Ts03fkk5lyKi2cz2q/AMuAU4PrQSgJZWOK5OlTmX0A7EblfVe8MPXdu9/Tjj/lDkaOFbtqeSmA4zpUlxSWbdao6OvKAiNTHBgmE/SUiDwCdRORZVf0xiCCdS6oKFWxSZrTzz4fXX098PM6VQcUlm5ejD4SaxiKbx/6OTejsDjwUmmuzGpihqrfHM1DnEmrLFptrE8uff8I++yQ2HufKuCKTjao+vaMXq2oOMElEzlHVWQAisj9wbNwidC6Rnn4arroq9jlfk8y5nVZczaaAUK1lKFAZeElV/xs+F040oce/YyPTnCs7ilpR+b33oEMRQ5qdcyVW4mSDjUyrhG0zMFJEclR1bDBhOZcA33wDLYtY0t9rMc7FVXHzbKIdA3RU1Wuwtc/+EUxIzu3AjrYBiNxSIJa99rKaTHSiufTS/CX9i/L887a4Z0nk5ZWsnHN7gNLUbERV/ww9/gh4LoB4nAPytxNYuT6b3ks+5tLNS2lwUoYtq794ccFtAGbPhn//217Yp0/BLQUOPdTObdsGr7wS+81atbK1yk4/3ZaT+e03qF3b3qNzZzjtNHvPa6+18q+9Zs1rtWrBXXfZFgQLF9p2AvfdZ8vShNcyO//84D8s58qA0tRsKohIYwBVzQWqBBNS8USko4gsEJHFIjIgxvm9RGRs6PxXItIwCWG6XRDeTqD80h8ZNe5erpn0H0aWb0TWB59AZqbNyB882JLArFnw+OO2QOaoUfDEE7alwJln2jL9PXrAiy/GTjT16lktpmpVGDDA9pvJy4Nq1WD8eCuTmwt9+9ry/088Ycc6dLCa09y5tpDmCy/Yoprp6fBtaOHNK6/0RONchNIkm/XADyKyRkQmAnuJyBkickAwoRUmIuWBJ4BOQDPgIhFpFlXsCmyO0KHAcODBRMXn4uOeifPIzsml+8x3qJCXy9q0auTl5LDw1z8hJ8f2drnrLksiY8fCF19YDWf5cnj/ffjPf2y1ZBGrcURq0sSSRnjTs8xMe/zHH3DJJfDuu5aA9tnHVntesgQWLLD3DQsv8y9iZevWtS0Ohg+3XTcjyzjngFI0o6nqgSJyIHBc6KcS8BqQLiKrsZUFgh620xJYHJ48KiKvAucC30eUORcYFHr8OjaYQbSodXlcSpkwM4t1m+yL/aNDj2PFvgdw/pwPOXnpTBr+shi6n2XNZPfdZ0mgZk1LOt27Q6NGtujlq68WvnDXrra/Ta1alhR++AE2bLAaz1FHwcsv27yaDRtg0iTbS+bAA+3Y2LG2rcAdd+TXXMKqVrW+n/AWzpdfHvsP69HDmuYejNO9T15esatQO5dqStNng6quBN4M/QDb10Q7DijxGjm7oC4FJ5WuAFoVVUZVt4nIH8D+2PbW24nIVcBVAA0aNAgqXldKQycviHn8zWZtqFY+j0a33WajyMD6WA47zJrNAJYtK/iifv1si+gpUyAryzY827YNPv7Yds485BD49VfbPmDMGDjvPNsoLSw8ECFyW4DmzfMfd+xoSaRZMzjoIHttvXrWt3PttdYc17y51ZC++Qauu85qZc8/X3Cr59GjLQn26wePPWY7jtavb8/79rUVDLZts3OdO9tunccdB+3a7dJn7Vwi7fStkYjUFZFzgEuBi4Eytdqzqo5S1QxVzajpq/KmjJUx9qzJ2vcAPjq0JU1qVc0/mJ1tX+innVb4Ihs2wAUXwCOP5B877TSrVfz3v3D00dYMFx4QcPXVVjP63/8sId1wQ/7ropajmTAzixOHfESjAZO4f9L3ZDY+1jY5y8y0mtCAAZZYLr/cmuumTLFk2KyZDV4YNsx23qxZ0xLPzz9borrySqhTB1assGTSu7f1Ce23n71m//3teV4e3HqrJxpX5pRmUuf52PDnY4AWQPgbWoEJ2K6eQcsC6kc8rxc6FqvMChGpAOwL/J6A2FwcHJietn1zsy8bNOfLBs154dizSU+rSN0pk+Dhh2HcOPuJFtlSGm5Ky8iwH7CmtEgPP2y/X3rJfr/4YsHzMRJN5F446zblMO7r5ayYmUWXnJz8LaVVi54kGr3VM1gT3aOPWk1q9GiYPh0uvNBqTj/9ZGXC/UNVqlhNJ5aSNK0tW2aDG/r1s+bBm24qeH7QIIvjiCOKv45zpVSaZrT/YV/ar2LJ5TtgLrAQ6KOqq+MeXWHfAI1FpBGWVC7EalWR3gJ6AF8A3YCPvL+m7Chqk7RZgzrk98RFmjXL+lwSYOjkBQXiAjh+8Qyyr78Bzm6Z37x30UX2JT5+fOGaV/RWz3XqwJdf2rDpevVY1Lsv3yz9nW0by/Hzpyu4esl8ajQLjYE56ihr/rvlFkuUAwbANddYgggPtR4+3JJqZqaVWbfOkumvv1ptqU4du9bWrda0OH8+3HMPNGgAl11m/WGrV8PFF+c3TzoXB6VJNkOAfkADYHhEJ30QccUU6oO5DpgMlAdGq+o8EbkXyFTVt4BngRdFZDGwljLWvJdQxd2BJ0l4/5ehkxeQtnghU569NnbBJNw/xGrie7NZG6Ye2pILbzsr/2CNGkXXko45xpryIoX6hibMzOL2BueSXccSWrc5U1i1ai3fjhpD+86t4f/+z+b0xJqweuWVNirumWes1jJxovVNZWRYH1CtWhbTrbcWfN3q1VbLuvhiq82cdJLXbFwgSjMa7Q4ReQEYCcwVkWHAA4FFVnQc7wDvRB37Z8TjzdgKBykncqJiC93AiLmvU/+whtYxPWqUFerTx74chg+3DuyKFSEtze5Un3vO+gE2bLAmm0MOseaWq6+2obbVq9vdbo8e1vF84ol2Z3333fnzSf71L+uwvvRSuwNu0iSpn0ksXe6+li4TJxY+cdddcO+9iQ8oJLKJD2xLabDdPuMhVs1J8/L4z6dLLdmI2OoH27bZyY0b8wuGh1rvvbf9rljRVq5+7DHrh1K1fwfR2rSBgw+GZ5+1ybE+ws0FpLSj0X4A2orIRdhW0D2AqsW/ykHh9v72017nyowzuaZ7R7o8fKv9zw52h3rFFTar/Y47rCP4/fdtiO5nn1mZrl3tDvTvf7cEc8op0KuXvS47276EOnWy408/bcfq1LE5J+EhvQMKzYctnXDbf7jfoyglrT1t3WpfpLFs2QKVKpU6xHgrqomvf4emcbl+dM1JVBl3xBl0n/QMrP/Ymudq17Y5QcOGwXff7fii4YERtWrFPj91qtWC1q61f2sHHABDh9oAhzZt4vBXOWdKlWzCVPUVEXkbuA9bI+0ZEblcVX/bwUv3WNF3rQJs2qYMnbyALpFfyOHmoWrV7Pf++9u5SpXsSxfy72xzcgp/mYtYc8n779tQ2yOPhLPOyp9sCAUmHBa7LEx2dsGlW666yuI6/HC7c/78c0s4HTtav0HlynD22da3cMkl9viKK2zkVVEmTiwYW9gFF8SeL5NEkU18K9dnx33r58iaU/VNf9Bh4ecMbP8PPmh7AecNOD2/YHh4dr9+9vv55/PPhZvrOnbMP1bUwIjw7+h+pU6ddv6PcK4IO5VsAELrpN0oIs8B/8YGC9SOV2C7m+i71hdbnMlN019m9ezqMOwm6+gFa0bburX4i732mo0kOu88aN/eXjtnjs3N2LjRllYpX96SQvfulnQ+/dSuO2LE9svEGl01smIjTmjfnS6Db7LXh5duuf12WLUKzj0XzjjDHs+aZde+5RYbQtyokfUrHHWUDfUtrvZUVG1n2TIbCpyiurSoG7fkEi2y5rS2yr707nY3aRXLMzhONSfnkqk0Q59rA39T1amRx1V1NnCiiPSKd3C7k+j2/hXptbn5rH7UTU/j2hNOsLkVkU491X6H7+7DEwwHDbKEFNmBGzkREWytsEhFdFbH6iPIy8mx2taaNZYQ7rsvv/lu7FhrdunZEx56KP9FkbWr8O9Yy7XMmxe74/nYY61Pag8XdM3JuWQqTc2mL1AdmAogIquwZWIyQz//LfqlLm7t/YMGxS2mWKOrTl46k8NWL4XOp1qH8cMPw5o11mR38802WKFJE2taW7zY+g6uvto676tUsX6FaCNGwI03Fj7+3/9ac1s8RfYlPf+8jQzr3Lno8im27EuQNSfnkklKOgVFRGYBXSOGPG8AHsW2gG4F/KOsbqaWkZGhmQm4s47sH0mFu9YTh3xUoLbVbc4U1qZVY0FGG6ZH9hHsjI0bbTHLWLZutT6fXRD5WZ645Vfu+mMmTWWTzTd5910byjtrFmzaZI9zc61/aPNmS9iffWaz+zMyrFzjxjbvJLzVwJ13Wh9Zbq41Bw4dahMvhw2zZWZat96l+J0r60RkhqqWeJmy0tzS1QsnmpBcVf2nqp4F9ATifIu6++nSoi7TB5zO0iFnMX3A6Um/g+3foSlpFctvf/76kW354m+td2101fPPW1NadKIZPTp/Y7I4JJrbx88ha302CqzYuI0vvs9igVaxFaBPOMH6kk46yRJN5862DcAzz8D998NTT9mFOnXKr3H17m2jtsaNgw8+sBpSerptIZCbawly3Tr4+uuCiWbZMuuzKqmePe2azu1hStOMVlFEKqpqeK31myPOTcVqOa4MiWsfQdu28OGHhY8vWWLzOOIouq+p54yJjGrVlbeqVmb8snksWv0XPYd8xPGffUed8ts4e/xkms6YYYMkIjc6a9fO5iRlZtrunkccYbWdZctsRn3PntYk16CBjczr0sWa3e6/P78WtHp1/qi8004rPFv/sststN28eTbnZdYsG8ARnkTp3B6iNMlmNtAemASgqqPDJ1R1o4j4apZl0C71EcycaTPio1WtahNPAxLd1/RFg+Zc8+U4fts7nZ/3qcbyr76j44ptfN7wKDpPG0NWhQrsc0gz6p51lo2Se/llSxDTptloPREbHj5woC3SOXasNauJ2FIy4Rn48+ZB06aWSKpUsRrN7bfnj8pbtKjwbP3DD7dyN99sNZqjj7b5U0U1MTq3mypNsnkSeFREvlbVNZEnQtsMrI1rZC51/f3vtkJytJkz7cs0YNEj+yY3PYHJTU+gbnoa44Cso/LPLd+vDqNadaV21cqM/+4bmzTbpIn9bNxogwPC81TKl7eVoitUsCX877nHmtYqV7amtXPPhR9/tFn6FSpYn07k4IJYs/WjZ/Sn0GAE5xKpNMvVvCIibYF5InIf8JKqrg0NiR6JrVfmdlebN9tItFhycopeiTgAxY3su2nsrAJlI2s9tG8PX31lzWjNm9tIuBtvtERy8smFR9VFztbfZx9rEoseDRg5Km9Hs/XB+ntuvdXmRkXujePcbq7Eo9EAxFbd7A/cCewDZANpwCygY3SNp6xI1Gi0MumRR2J3gA8YAIMHJz6ekKJG9kWPsAurm5626yPsnHPblXY0WqmSTcSbpAGnAnWAZcCnEQMHyhxPNlFUi27uWbPG5q6kqOhVEcBqPYO7Hpn00X/O7U5Km2x2dm20bODdnXmtS2FZWbZadCxlZEsgn4XvXGpKXEO7S10nn5y/JE2kKVNsHbQyxmfhO5d6PNnsqf78M39l6UiHHgoLFvioKedcXPk3yp5m4kSbPxKdaK6/3prKFi3yROOcizuv2ewJiuvwX7Wq+KG6zjkXB34LuzubNctqMdGJ5oor8tcp80TjnEsATza7o+uusyTTokXB49OmWYJ55pmkhOWc23N5M9ruYt06qF499rncXO+Hcc4llX8DlXX/+5/VYqITzahR+U1lnmicc0nmNZuySBXq17dJmNFWr4aavgC3cy61+C1vWTJvXn6Hf2Siueyy/FpMaRJN9KoAgwbB3LkFj/Xvb7+7dSvZNUu7mZhzbo/gNZuy4Npr83eXjDRrFhx1VJEvCy9WufeiH+j+42e02TePg267wZLB2WfbysYDB8IBB8B559mLRo2yCZ+nnAK9esHSpba6wPffw7/+ZfNx7r7bElXVqnasf3/Yts02GatZM38zseuuC+bzcM6VOZ5sUlVRM/zBdosUKfblkQtSNixfga1/beK9rZU5t/+d1G7WzFZtvvVW+Oc/baMwsOVpLrgATjzRfvfqZcdPOsk2HRs40Ha6zM62XSgXLoTvvoNKlWDoUCu7bFn+ZmLOORfizWip5pVXYs/wHzMmv6lsB4kGCm6d3HPGREYfdy4vNO/At9kVYd99rVCswQPhc9HvEX6el2e7Wg4aZDteli9f8Bo+GME5F4PXbFKBqm3OtWlT4XPr1+cngFKI3Do5cgOxTVvzl96nTx9LGnXqwDnnFH/BihVtY7FbbrFay6efwtatMGKE1XT694eDDoKrrsrfTKxfv1LH7ZzbPe3Ufja7m6TtZ7N6dewZ/JdfDs8+u0uX9k3EnHNBKu1+Nt7mkQz33GPNUtGJZs4cq+XsYqIB2zo5rWL5AsfCWyc751yiebJJlM2b4f774ZBDCu5jP2iQ9YOowhFHxO3turSoy+CuR1I3PQ3BajS+W6VzLlm8zyZoq1bBTTfBq6/a8+rVrUYzZgx06BDoW/smYs65VOHJJgiqNl/l5putEz0nx4736mVDh/faK7nxOedcgnmyiadVq+Dpp229sjlz7NiNN9qkzCZNkhqac84lkyebeJg/H44+2moxYI9HjbJZ+TVqJDMy55xLCT5AYFc89xw0bWqz68OJZvhw+PZbuPJKTzTOORfiNZtdMXq0LdkyeLDNjTnggGRH5JxzKanMJBsRqQ6MBRoCy4C/q+q6qDJHA08C1YBc4H5VHRtYUJ9+WuLlY5xzbk9WlprRBgAfqmpj4MPQ82ibgMtU9XCgI/CoiKQHGpUnGuec26GylGzOBcaEHo8BukQXUNWFqroo9HglsBrwncSccy7JylKyqaWqv4QerwJiLCqWT0RaApWAJUWcv0pEMkUkc82aNfGN1DnnXAEp1WcjIlOA2jFO3Rn5RFVVRIpcQVRE6gAvAj1UNS9WGVUdBYwCW4hzp4N2zjm3QymVbFS1bVHnRORXEamjqr+EksnqIspVAyYBd6rqlwGF6pxzrhTKUjPaW0CP0OMewJvRBUSkEvAG8IKqvp7A2JxzzhWjLCWbIUA7EVkEtA09R0QyROSZUJm/A6cAPUVkVujn6KRE65xzbjvfPI0kbp7mnHNllG+e5pxzLuV4snHOORc4TzbOOecC58nGOedc4DzZOOecC5wnG+ecc4HzZOOccy5wnmycc84FzpONc865wHmycc45FzhPNs455wLnycY551zgPNk455wLnCcb55xzgfNk45xzLnCebJxzzgXOk41zzrnAebJxzjkXOE82zjnnAufJxjnnXOA82TjnnAucJxvnnHOB82TjnHMucJ5snHPOBc6TjXPOucB5snHOORc4TzZJNGFmFicO+YhGAyZx4pCPmDAzq3ChTz+FCy6AZ56BW27ZuTfKy9u1QIvSv/+uXyOo2JxzKaVCsgPYU02YmcXt4+eQnZPLOd9P4/if5rBuwt7MOqMlR6//CTZsgEcfhREjoFIlWLkSPv8cRo6ExYvhnnss+bRpA5UrQ4UKULEifPwxrF4Nw4bBW2/BlCmQkQGHHw6TJkF2Npx/PrRvXzCgJ5+EhQth/Xq47z4YOBAaN4asLOjSBU45Ba64Aho1gnfftfdZutRe27kznHoqzJ0L/frBQQfB3XeDKlStCv/6F9x/P/z2G/z5JwwfDhddBCecAMcdB+3aJfbDd84lnCebJBk6eQHZObkAHLTuF36o2ZAPGh/PkDefhbmf2pf5K6/YF3mNGnDEEZaArrsOXn/dajzVqsHs2bBlCwwaBF9/bTWFnBxLMgCdOsEll8DZZ8Oxx0J6upWLTDZ//QUvvAAdOtjzb7+13717Q5UqltS2bLHX9OgBM2cW/oP69YN58+Dtt+09srOhTh1LYLNnwyefQOvWsHkzzJ9vcd56qyVJ59xuz/9PT5KV67O3P378xIv42+ofuWPqaDbnbrODIlYzCCsX0eLZpg1cfbV9+S9aBGvXQvXqVjt5800YMwY2bbKy++5rv/PyrLYS68tdFerWtYQVNn487L23ld+ypWB5kYLPK1e2+CpWtLJ5eXDWWXDOOXZ+3jyrWUVev0oVTzTO7UH8//YkOTA9jaxQwrlo1ns0XLeSPCnHskbNoG9fWLfOmpveftteULu2NZ8NG2a1iEWL4IEHrKYQ7vdo1syaq+bPh7ZtC75h375WU6le3ZrVLr44/1zVqtCyJVx/vSWeyy8vHHC7dvb6xYutSa9y5aL/uO7drQb26aewdas1BZYrZ3FnZ8Mdd+zCJ+ecK4tEI++e91AZGRmamZmZ0PeM7LMJS6tYnsFdj6RLi7oJjaXEHnkEVq2y5ru77kp2NM65JBKRGaqaUdLyXrNJknBCGTp5ASvXZ3Ngehr9OzRNTKLZvBmGDMl/fvzx0LHjjl93883BxeSc2615zYbk1Gycc64sK23NxufZOOecC5wnG+ecc4HzZOOccy5wZSbZiEh1EflARBaFfu9XTNlqIrJCREYmMkbnnHOxlZlkAwwAPlTVxsCHoedFuQ/4JCFROeec26GylGzOBcaEHo8BusQqJCLHArWA9xMTlnPOuR0pS/NsaqnqL6HHq7CEUoCIlAMeAboDbaPPR5W9Crgq9PQvEVkQcboG8NsuRxwMj630UjUu8Nh2lsdWevGO66DSFE6pZCMiU4DaMU7dGflEVVVEYk0Q6gO8o6orJHr9riiqOgoYVUQcmaUZP55IHlvppWpc4LHtLI+t9JIdV0olG1UtsjYiIr+KSB1V/UVE6gCrYxRrDZwsIn2AfYBKIvKXqhbXv+Occy5gKZVsduAtoAcwJPT7zegCqnpJ+LGI9AQyPNE451zylaUBAkOAdiKyCOuPGQIgIhki8kyc3ytm81qK8NhKL1XjAo9tZ3lspZfUuHxtNOecc4ErSzUb55xzZZQnG+ecc4Hbo5ONiHQUkQUislhECg0kEJFTRORbEdkmIt1SKK5+IvK9iHwnIh+KSKnGuwcc2zUiMkdEZonIZyLSLFViiyh3voioiCRsGGgJPreeIrIm9LnNEpHeqRJbqMzfQ//m5onIy6kQl4gMj/i8ForI+kTEVcLYGojIVBGZGfr/9MwUiu2g0PfGdyIyTUTqJSQwVd0jf4DywBLgYKASMBtoFlWmIdAceAHolkJxnQZUCT2+FhibQrFVi3h8DvBeqsQWKlcVW8roS2y0YkrEBvQERiYinp2IrTEwE9gv9PyAVIgrqvz1wOgU+sxGAdeGHjcDlqVQbP8DeoQenw68mIjY9uSaTUtgsar+qKpbgVexJXG2U9VlqvodkJdicU1V1U2hp18CibkzKVlsGyKe7g0kagTKDmMLuQ94ENicoLhKE1sylCS2K4EnVHUdgKrGmuOWjLgiXQS8koC4oGSxKVAt9HhfYGUKxdYM+Cj0eGqM84HYk5NNXeDniOcrQseSrbRxXQG8G2hE+UoUm4j8Q0SWAA8BfVMlNhE5BqivqpMSFFNYSf+bnh9q2nhdROonJrQSxdYEaCIi00XkSxEpwR7iCYkLsGYhoBH5X6BBK0lsg4DuIrICeAereSVCSWKbDXQNPT4PqCoi+wcd2J6cbMo8EekOZABDkx1LJFV9QlUPAW4DBiY7Hti+bt4w4OZkx1KEiUBDVW0OfED+orOpoALWlHYqVoN4WkTSkxlQlAuB11U1N9mBRLgIeF5V6wFnAi+G/g2mgluANiIyE2gDZAGBf3ap8scnQxYQefdYL3Qs2UoUl4i0xdaMO0dVt6RSbBFepYjVuQOwo9iqAkcA00RkGXA88FaCBgns8HNT1d8j/js+AxybgLhKFBt2d/yWquao6lJgIZZ8kh1X2IUkrgkNShbbFcBrAKr6BVAZWwgz6bGp6kpV7aqqLQitO6mq6wOPLBEdQ6n4g92t/YhVv8MdaYcXUfZ5EjdAYIdxAS2wTsDGqfaZRcYEnA1kpkpsUeWnkbgBAiX53OpEPD4P+DKFYusIjAk9roE10+yf7LhC5Q4DlhGaoJ5Cn9m7QM/Q479hfTaBx1jC2GoA5UKP7wfuTcjnlqj/QKn4g1VvF4a+uO8MHbsXqy0AHIfd1W0EfgfmpUhcU4BfgVmhn7dS6DMbAcwLxTW1uC/8RMcWVTZhyaaEn9vg0Oc2O/S5HZZCsQnWBPk9MAe4MBXiCj0fBAxJ1GdVis+sGTA99N9zFtA+hWLrBiwKlXkG2CsRcflyNc455wK3J/fZOOecSxBPNs455wLnycY551zgPNk455wLnCcb55xzgfNk45xzLnCebJzbARG5VkR+FZGfRaRL1Ll3o4855wrzeTbOFUNEamGTLY8GagPvAbVUNVdEegCdVPXCJIboXJngNRvnincQsEhVV6hqJrAN2F9EamPrSu1wNd/QSs4Xhx6nichmEXkh4vw7InJr6HEVERkRqkX9JiITRKRBRNlpIjJMRN4QkT9FZImInCEibUVkrohsCJ2rGvGa/UXk2dA114jIa6EkGj6/TETuCG2o9VfoOifE4bNzbjtPNs4VbzHQKLS7YUss2awBngAGqeqaElxjCtA29PgUbG2xMwBEpFLo2JTQ+eHYIqHHY4nuN2CiiJSPuN6lwBAgHRgLvAhcFbpOQ6Apoa0dRESACdj+KkeErvknEL3b5uWh1+xL6q067XYD3ozm3A6ISFdgALA19Ls20AO4BHgM2xVxBtBfVbfFeP2ZwJOqepCIDMW+7C/B9hSpAYwDamJrkG0CzlbVD0Kv3QdYC7RR1S9EZBq2Rt8/QuebYc18LVX1m9Cxh7AFUc8LrWr9CbbL5pbQ+f2xJFZfVVeEVsF+QlWHhs4fDswF0lX1jzh9jG4P5zUb53ZAVceraktVPQlbjPJ+4BrgdmC5qp4CHAD0KuISHwN1RKQJVsP5AKvJtAs9/0jtrq8msBewNOK9/wJWU3DZ+F8iHm8q4li4Ga1R6Jq/ish6EVmPLdC4GWgQ8ZrI128M/a6Kc3FSIdkBOFfGPAo8rKpZInIUtso1wKfY1g+FqOpGEfkS23elIfA1VjvqDVQHngsVXQNsCZVZDNtrNgdQcPfF0liOJY/qqprI7c2dK8BrNs6VkIh0Auqq6tOhQ0uAjiJSEWhPKEEUYQq2S+g0tR0lpwInYzutTgEIJYMXgPtE5EARqQI8AvyAJaidkYktc/9YeOtfEakpIj6CziWUJxvnSiA0umsYcGXE4cHYjpq/Yf8v/aeYS0wBqmFNaKjtjPgD8LOq/hhR7iYsQXwD/ATUwfYh2alte0MJ7FysP2iGiPwJfIlt8excwvgAAeecc4Hzmo1zzrnAebJxzjkXOE82zjnnAufJxjnnXOA82TjnnAucJxvnnHOB82TjnHMucJ5snHPOBe7/ATsAz1F6INgLAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 301 seconds\n",
      "\n",
      "\n",
      "\n",
      "================================================                \n",
      "SCORE: of profession classifier on debiased dataset                \n",
      "0.5806                \n",
      "================================================\n",
      "\n",
      "\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:  5.0min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correlation: 0.3789776883508932; p-value: 0.14772867602704037\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rms-diff before: 0.18337572303336608; rms-diff after: 0.1492207928515152\n"
     ]
    }
   ],
   "source": [
    "# Removal 1 BERT Poly 2\n",
    "tpr_exp(x_test, debiased_x_train_1, y_m_train, debiased_x_test_1, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "profession dietitian missed in tpr-calc:\n",
      "                  number of dietitian male in test set 0\n",
      "                  number of dietitian female in test set 59\n",
      "\n",
      "\n",
      "Correlation: 0.773655072810786; p-value: 6.29340039229481e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=90)]: Using backend ThreadingBackend with 90 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 296 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/lustre/home/sc066/shunshao/miniconda3/envs/cca/lib/python3.7/site-packages/sklearn/linear_model/_sag.py:329: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "[Parallel(n_jobs=90)]: Done   1 out of   1 | elapsed:  4.9min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "\n",
      "================================================                \n",
      "SCORE: of profession classifier on debiased dataset                \n",
      "0.5762                \n",
      "================================================\n",
      "\n",
      "\n",
      "\n",
      "Correlation: 0.08446044696320024; p-value: 0.7472326984514821\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "rms-diff before: 0.18337572303336608; rms-diff after: 0.10341280718742507\n"
     ]
    }
   ],
   "source": [
    "# Removal 2 BERT Poly 2\n",
    "tpr_exp(x_test, debiased_x_train_2, y_m_train, debiased_x_test_2, y_m_test, \\\n",
    "        y_p_test, N_test, p2i, i2p, clf_original)"
   ]
  }
 ],
 "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}