notebooks/notebook_fair-profession_fasttext_INLP.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "import sys\n",
    "sys.path.append(\"../src\")\n",
    "sys.path.append(\"../data/embeddings\")\n",
    "sys.path.append(\"../data/biasbios\")\n",
    "sys.path.append(\"../data/embeddings/biasbios\")\n",
    "import classifier\n",
    "import debias\n",
    "import gensim\n",
    "import codecs\n",
    "import json\n",
    "from gensim.models.keyedvectors import Word2VecKeyedVectors\n",
    "from gensim.models import KeyedVectors\n",
    "import numpy as np\n",
    "import random\n",
    "import sklearn\n",
    "from sklearn import model_selection\n",
    "from sklearn import cluster\n",
    "from sklearn import metrics\n",
    "from sklearn.manifold import TSNE\n",
    "from sklearn.svm import LinearSVC, SVC\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from sklearn.feature_extraction import DictVectorizer\n",
    "from pytorch_transformers import BertTokenizer, BertModel, BertForMaskedLM\n",
    "\n",
    "import scipy\n",
    "from scipy import linalg\n",
    "from scipy import sparse\n",
    "from scipy.stats.stats import pearsonr\n",
    "import tqdm\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.linear_model import SGDClassifier, SGDRegressor, Perceptron, LogisticRegression\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "%matplotlib inline\n",
    "matplotlib.rcParams['agg.path.chunksize'] = 10000\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "\n",
    "import pickle\n",
    "from collections import defaultdict, Counter\n",
    "from typing import List, Dict\n",
    "\n",
    "import torch\n",
    "from torch import utils\n",
    "\n",
    "# import pytorch_lightning as pl\n",
    "# from pytorch_lightning import Trainer\n",
    "import copy\n",
    "import pandas as pd\n",
    "from gensim.models import FastText\n",
    "import time\n",
    "from gensim.scripts.glove2word2vec import glove2word2vec\n",
    "\n",
    "STOPWORDS = set([\"i\", \"me\", \"my\", \"myself\", \"we\", \"our\", \"ours\", \"ourselves\", \"you\", \"your\", \"yours\", \"yourself\", \"yourselves\", \"he\", \"him\", \"his\", \"himself\", \"she\", \"her\", \"hers\", \"herself\", \"it\", \"its\", \"itself\", \"they\", \"them\", \"their\", \"theirs\", \"themselves\", \"what\", \"which\", \"who\", \"whom\", \"this\", \"that\", \"these\", \"those\", \"am\", \"is\", \"are\", \"was\", \"were\", \"be\", \"been\", \"being\", \"have\", \"has\", \"had\", \"having\", \"do\", \"does\", \"did\", \"doing\", \"a\", \"an\", \"the\", \"and\", \"but\", \"if\", \"or\", \"because\", \"as\", \"until\", \"while\", \"of\", \"at\", \"by\", \"for\", \"with\", \"about\", \"against\", \"between\", \"into\", \"through\", \"during\", \"before\", \"after\", \"above\", \"below\", \"to\", \"from\", \"up\", \"down\", \"in\", \"out\", \"on\", \"off\", \"over\", \"under\", \"again\", \"further\", \"then\", \"once\", \"here\", \"there\", \"when\", \"where\", \"why\", \"how\", \"all\", \"any\", \"both\", \"each\", \"few\", \"more\", \"most\", \"other\", \"some\", \"such\", \"no\", \"nor\", \"not\", \"only\", \"own\", \"same\", \"so\", \"than\", \"too\", \"very\", \"s\", \"t\", \"can\", \"will\", \"just\", \"don\", \"should\", \"now\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "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\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "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",
    "\n",
    "p2i, i2p = load_dictionary(\"../data/biasbios/profession2index.txt\")\n",
    "g2i, i2g = load_dictionary(\"../data/biasbios/gender2index.txt\")\n",
    "counter = count_profs_and_gender(train+dev+test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'g': 'f',\n",
       " 'p': 'professor',\n",
       " 'text': 'Dr. Jessica Orlofske is an Assistant Professor of Biology at the University of Wisconsin-Parkside. Her research program areas include invertebrate biodiversity and conservation and biomonitoring and assessment of ecological integrity. She obtained her PhD from the University of New Brunswick under the direction of Dr. Donald Baird. Dr. Orlofske was also a post-doctoral fellow with the CRI before taking up her current position, making her the first CRI student to become a Science Director.',\n",
       " 'start': 98,\n",
       " 'hard_text': 'Her research program areas include invertebrate biodiversity and conservation and biomonitoring and assessment of ecological integrity. She obtained her PhD from the University of New Brunswick under the direction of Dr. Donald Baird. Dr. Orlofske was also a post-doctoral fellow with the CRI before taking up her current position, making her the first CRI student to become a Science Director.',\n",
       " 'text_without_gender': '_ research program areas include invertebrate biodiversity and conservation and biomonitoring and assessment of ecological integrity. _ obtained _ PhD from the University of New Brunswick under the direction of Dr. Donald Baird. Dr. _ was also a post-doctoral fellow with the CRI before taking up _ current position, making _ the first CRI student to become a Science Director.',\n",
       " 'hard_text_tokenized': 'Her research program areas include invertebrate biodiversity and conservation and biomonitoring and assessment of ecological integrity . She obtained her PhD from the University of New Brunswick under the direction of Dr. Donald Baird . Dr. Orlofske was also a post - doctoral fellow with the CRI before taking up her current position , making her the first CRI student to become a Science Director .'}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'accountant': 0, 'architect': 1, 'attorney': 2, 'chiropractor': 3, 'comedian': 4, 'composer': 5, 'dentist': 6, 'dietitian': 7, 'dj': 8, 'filmmaker': 9, 'interior_designer': 10, 'journalist': 11, 'model': 12, 'nurse': 13, 'painter': 14, 'paralegal': 15, 'pastor': 16, 'personal_trainer': 17, 'photographer': 18, 'physician': 19, 'poet': 20, 'professor': 21, 'psychologist': 22, 'rapper': 23, 'software_engineer': 24, 'surgeon': 25, 'teacher': 26, 'yoga_teacher': 27}\n"
     ]
    }
   ],
   "source": [
    "print(p2i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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": [
    "counter = count_profs_and_gender(train+dev+test)\n",
    "f,m = 0., 0.\n",
    "prof2fem = dict()\n",
    "\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": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'accountant': 0, 'architect': 1, 'attorney': 2, 'chiropractor': 3, 'comedian': 4, 'composer': 5, 'dentist': 6, 'dietitian': 7, 'dj': 8, 'filmmaker': 9, 'interior_designer': 10, 'journalist': 11, 'model': 12, 'nurse': 13, 'painter': 14, 'paralegal': 15, 'pastor': 16, 'personal_trainer': 17, 'photographer': 18, 'physician': 19, 'poet': 20, 'professor': 21, 'psychologist': 22, 'rapper': 23, 'software_engineer': 24, 'surgeon': 25, 'teacher': 26, 'yoga_teacher': 27}\n"
     ]
    }
   ],
   "source": [
    "print(p2i)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### get input representatons "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "def load_word_vectors(fname):\n",
    "    \n",
    "    model = KeyedVectors.load_word2vec_format(fname, binary=False)\n",
    "    vecs = model.vectors\n",
    "    words = list(model.vocab.keys())\n",
    "    return model, vecs, words\n",
    "\n",
    "\n",
    "def get_embeddings_based_dataset(data: List[dict], word2vec_model, p2i, filter_stopwords = False):\n",
    "    \n",
    "    X, Y = [], []\n",
    "    unk, total = 0., 0.\n",
    "    unknown = []\n",
    "    vocab_counter = Counter()\n",
    "    \n",
    "    for entry in tqdm.tqdm_notebook(data, total = len(data)):\n",
    "        \n",
    "        y = p2i[entry[\"p\"]]\n",
    "        words = entry[\"hard_text_tokenized\"].split(\" \")\n",
    "        if filter_stopwords:\n",
    "            words = [w for w in words if w.lower() not in STOPWORDS]\n",
    "            \n",
    "        vocab_counter.update(words) \n",
    "        bagofwords = np.sum([word2vec_model[w] if w in word2vec_model else word2vec_model[\"unk\"] for w in words], axis = 0)\n",
    "        #print(bagofwords.shape)\n",
    "        X.append(bagofwords)\n",
    "        Y.append(y)\n",
    "        total += len(words)\n",
    "        \n",
    "        unknown_entry = [w for w in words if w not in word2vec_model]\n",
    "        unknown.extend(unknown_entry)\n",
    "        unk += len(unknown_entry)\n",
    "    \n",
    "    X = np.array(X)\n",
    "    Y = np.array(Y)\n",
    "    print(\"% unknown: {}\".format(unk/total))\n",
    "    return X,Y,unknown,vocab_counter\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "word2vec, vecs, words = load_word_vectors(\"../data/embeddings/crawl-300d-2M.vec\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d82c5b5735b94948a205af1c9bc31045",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/74882 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "% unknown: 0.010167678378823583\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "758f6e69806e4f1783d8d59b2b35d425",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/11550 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "% unknown: 0.010116209992979287\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3f5a9ad077a0417baac596d738a3332d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/28842 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "% unknown: 0.01018950257509165\n"
     ]
    }
   ],
   "source": [
    "X_train, Y_train, unknown_train, vocab_counter_train = get_embeddings_based_dataset(train, word2vec, p2i)\n",
    "X_dev, Y_dev, unknown_dev, vocab_counter_dev =  get_embeddings_based_dataset(dev, word2vec, p2i)\n",
    "X_test, Y_test, unknown_test, vocab_counter_test =  get_embeddings_based_dataset(test, word2vec, p2i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Run only at initialization\n",
    "# def save_in_word2vec_format(vecs: np.ndarray, words: np.ndarray, fname: str):\n",
    "\n",
    "\n",
    "#     with open(fname, \"w\", encoding = \"utf-8\") as f:\n",
    "\n",
    "#         f.write(str(len(vecs)) + \" \" + \"300\" + \"\\n\")\n",
    "#         for i, (v,w) in tqdm.tqdm_notebook(enumerate(zip(vecs, words))):\n",
    "\n",
    "#             vec_as_str = \" \".join([str(x) for x in v])\n",
    "#             f.write(w + \" \" + vec_as_str + \"\\n\")\n",
    "\n",
    "\n",
    "# vocab_bios, _ = list(zip(*vocab_counter_train.most_common(120000)))\n",
    "# words_set = set(words)\n",
    "# vocab_bios = [w for w in tqdm.tqdm_notebook(vocab_bios) if w in words_set]\n",
    "# vecs_for_vocab = np.array([word2vec[w] for w in tqdm.tqdm_notebook(vocab_bios)])\n",
    "# print(\"here\")\n",
    "# save_in_word2vec_format(vecs_for_vocab, vocab_bios, \"vecs.vocab.bios.txt\")\n",
    "\n",
    "# word2vec_bios, _, _ = load_word_vectors(\"vecs.vocab.bios.txt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "len train: 74882; len dev: 11550; len test: 28842\n"
     ]
    }
   ],
   "source": [
    "print(\"len train: {}; len dev: {}; len test: {}\".format(len(train), len(dev), len(test)))\n",
    "mean_train = np.mean(X_train, axis = 0, keepdims = True)\n",
    "mean_dev = np.mean(X_dev, axis = 0, keepdims = True)\n",
    "mean_test = np.mean(X_test, axis = 0, keepdims = True)\n",
    "\n",
    "#X_train -= mean_train\n",
    "#X_dev -= mean_dev\n",
    "#X_test -= mean_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classify by profession"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Using backend ThreadingBackend with 64 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 17 seconds\n",
      "time: 17.3991961479187\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Done   1 tasks      | elapsed:   17.3s\n",
      "[Parallel(n_jobs=64)]: Done   1 out of   1 | elapsed:   17.3s finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7554112554112554\n"
     ]
    }
   ],
   "source": [
    "#clf = LinearSVC(max_iter = 10) #LogisticRegression()\n",
    "#clf = SGDClassifier(warm_start = True, loss = 'log', n_jobs = 64, max_iter = 250, random_state = 0)\n",
    "\n",
    "clf = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"sag\", multi_class = 'multinomial', fit_intercept = True,\n",
    "                         verbose = 10, max_iter = 4, n_jobs = 64, random_state = 1)\n",
    "#clf = SGDClassifier(loss= 'log', max_iter= 3000, fit_intercept= True, class_weight= None, n_jobs= 100)\n",
    "#clf = LinearSVC(max_iter = 50) #LogisticRegression()\n",
    "\n",
    "start = time.time()\n",
    "idx = np.random.rand(X_train.shape[0]) < 1.0\n",
    "clf.fit(X_train[idx], Y_train[idx])\n",
    "print(\"time: {}\".format(time.time() - start))\n",
    "print(clf.score(X_dev, Y_dev))\n",
    "#print(clf.score(X_test, Y_test))\n",
    "clf_original = copy.deepcopy(clf)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7553914430344636\n"
     ]
    }
   ],
   "source": [
    "print(clf.score(X_test, Y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### perform debiasing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "iteration: 149, accuracy: 0.6567164179104478: 100%|█████████████████████████████████████████████████████████████████████████████████████| 150/150 [05:30<00:00,  2.20s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "time: 330.5958001613617\n"
     ]
    }
   ],
   "source": [
    "def get_projection_matrix(num_clfs, X_train, Y_train, X_dev, Y_dev, Y_train_task, Y_dev_task, dim, all_data_prob, by_class = False):\n",
    "\n",
    "    is_autoregressive = True\n",
    "    min_acc = 0.\n",
    "    dim = 300\n",
    "    n = num_clfs\n",
    "    random_subset = 1\n",
    "    start = time.time()\n",
    "    TYPE= \"svm\"\n",
    "    penalty = \"l2\"\n",
    "    MLP = False\n",
    "    \n",
    "    if MLP:\n",
    "        x_train_gender = np.matmul(X_train, clf.coefs_[0]) + clf.intercepts_[0]\n",
    "        x_dev_gender = np.matmul(X_dev, clf.coefs_[0]) + clf.intercepts_[0]\n",
    "    else:\n",
    "        x_train_gender = X_train.copy()\n",
    "        x_dev_gender = X_dev.copy()\n",
    "        \n",
    "    \n",
    "    if TYPE == \"sgd\":\n",
    "        print(\"using sgd\")\n",
    "        gender_clf = SGDClassifier\n",
    "        params = {'alpha': 0.01, 'penalty': penalty, 'loss': 'hinge', 'fit_intercept': True, 'class_weight': \"balanced\", 'n_jobs': 16}\n",
    "    elif TYPE == \"svm\":\n",
    "        gender_clf = LinearSVC\n",
    "        params = {'fit_intercept': True, 'C': 0.3, 'class_weight': None, \"dual\": False}\n",
    "    elif TYPE == \"perceptron\":\n",
    "        gender_clf = Perceptron\n",
    "        params = {'max_iter': 1000, 'fit_intercept': True, 'class_weight': None}\n",
    "    elif TYPE == \"logistic\":\n",
    "        gender_clf = LogisticRegression\n",
    "        params = {}\n",
    "        \n",
    "    result = debias.get_debiasing_projection(gender_clf, params, n, dim, is_autoregressive, min_acc,\n",
    "                                              x_train_gender, Y_train, x_dev_gender, Y_dev,\n",
    "                                       Y_train_main=Y_train_task, Y_dev_main=Y_dev_task, \n",
    "                                        by_class = by_class)\n",
    "    print(\"time: {}\".format(time.time() - start))\n",
    "    return result\n",
    "\n",
    "# was c=0.15, num_clfs=130\n",
    "num_clfs = 150\n",
    "Y_dev_gender = np.array([g2i[d[\"g\"]] for d in dev])\n",
    "Y_test_gender = np.array([g2i[d[\"g\"]] for d in test])\n",
    "Y_train_gender = np.array([g2i[d[\"g\"]] for d in train])\n",
    "P, rowspace_projs, Ws = get_projection_matrix(num_clfs, X_train, Y_train_gender, X_dev, Y_dev_gender, Y_train, Y_dev, 300, 0.0, by_class= True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.savez(\"P_FastText.npz\", P = P, rowspace_projs = rowspace_projs, Ws = Ws)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(\"P.bios.fasttext.{}.svm.pickle\".format(num_clfs), \"wb\") as f:\n",
    "    \n",
    "#     pickle.dump(P, f)\n",
    "    \n",
    "# np.linalg.norm(Ws[-1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Ws[-50].dot(Ws[0].T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Using backend ThreadingBackend with 64 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 4 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Done   1 tasks      | elapsed:    4.2s\n",
      "[Parallel(n_jobs=64)]: Done   1 out of   1 | elapsed:    4.2s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.5925541125541125"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gender_clf = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"sag\", multi_class = 'multinomial', fit_intercept = True,\n",
    "                         verbose = 10, max_iter = 7, n_jobs = 64, random_state = 1)\n",
    "\n",
    "gender_clf.fit(P.dot(X_train.T).T, Y_train_gender)\n",
    "gender_clf.score(P.dot(X_dev.T).T, Y_dev_gender)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5894741059266579"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gender_clf.score(P.dot(X_train.T).T, Y_train_gender)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Perform finetuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Using backend ThreadingBackend with 64 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max_iter reached after 27 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=64)]: Done   1 tasks      | elapsed:   26.8s\n",
      "[Parallel(n_jobs=64)]: Done   1 out of   1 | elapsed:   26.8s finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "LogisticRegression(max_iter=6, multi_class='multinomial', n_jobs=64,\n",
       "                   random_state=0, solver='sag', verbose=10, warm_start=True)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "                         solver = \"sag\", multi_class = 'multinomial', fit_intercept = True,\n",
    "                         verbose = 10, max_iter = 6, n_jobs = 64, random_state = 0, class_weight = None)\n",
    "\n",
    "clf.fit(P.dot(X_train.T).T, Y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy 0.7203466787746053\n",
      "\n",
      "Test accuracy 0.7149989598502184\n"
     ]
    }
   ],
   "source": [
    "print(f\"Train accuracy {clf.score(P.dot(X_train.T).T, Y_train)}\\n\")\n",
    "print(f\"Test accuracy {clf.score((P.dot(X_test.T)).T, Y_test)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clf = LogisticRegression(warm_start = True, penalty = 'l2',\n",
    "#                          solver = \"sag\", multi_class = 'multinomial', fit_intercept = True,\n",
    "#                          verbose = 10, max_iter = 3, n_jobs = 64, random_state = 1)\n",
    "# clf.fit(X_train, Y_train)\n",
    "\n",
    "# print(f\"Biased Train accuracy {clf.score(X_train, Y_train)}\")\n",
    "# print(f\"Biased Train accuracy {clf.score(X_test, Y_test)}\")\n",
    "# print(f\"Train accuracy {clf.score(P.dot(X_train.T).T, Y_train)}\")\n",
    "# print(f\"Test accuracy {clf.score((P.dot(X_test.T)).T, Y_test)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### test again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_TPR(y_pred, y_true, p2i, i2p, gender):\n",
    "    \n",
    "    scores = defaultdict(Counter)\n",
    "    prof_count_total = defaultdict(Counter)\n",
    "    \n",
    "    assert len(y_pred) == len(y_true) == len(gender)\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",
    "        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",
    "        \n",
    "    return tprs, tprs_change, np.mean(np.abs(tprs_ratio))\n",
    "\n",
    "def get_FPR(y_pred, y_true, p2i, i2p, gender):\n",
    "    \n",
    "    fp = defaultdict(Counter)\n",
    "    neg_count_total = defaultdict(Counter)\n",
    "    pos_count_total = defaultdict(Counter)\n",
    "    \n",
    "    \n",
    "               \n",
    "    for y_hat, y, g in zip(y_pred, y_true, gender):\n",
    "        \n",
    "        if y != y_hat:\n",
    "            \n",
    "            fp[i2p[y_hat]][g] += 1 # count false positives for y_hat\n",
    "        \n",
    "        neg_count_total[i2p[y_hat]][g] += 1 # count \"not y_hat\"\n",
    "                \n",
    "    fprs = defaultdict(dict)\n",
    "    fprs_change = dict()\n",
    "    fprs_ratio = []\n",
    "    \n",
    "    #print(prof_count_total)\n",
    "    \n",
    "    for profession, scores_dict in fp.items():\n",
    "        \n",
    "        good_m, good_f = scores_dict[\"m\"], scores_dict[\"f\"]\n",
    "        prof_total_f = neg_count_total[profession][\"f\"]\n",
    "        prof_total_m = neg_count_total[profession][\"m\"]\n",
    "        print(profession)\n",
    "        fpr_m = (good_m+good_f) / prof_total_m\n",
    "        fpr_f = (good_m+good_f) / prof_total_f\n",
    "        \n",
    "        fprs[profession][\"m\"] = fpr_m\n",
    "        fprs[profession][\"f\"] = fpr_f\n",
    "        fprs_ratio.append(0)\n",
    "        fprs_change[profession] = fpr_m - fpr_f\n",
    "        \n",
    "    return fprs, fprs_change, np.mean(np.abs(fprs_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",
    "    \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),\"b--\")\n",
    "\n",
    "    plt.savefig(\"{}_vs_bias_{}_fasttext\".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": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Correlation: 0.8229158334674403; p-value: 7.657770186244778e-08\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": [
      "Correlation: 0.4625654647200439; p-value: 0.015122273151764936\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: 0.20125138721368724; rms-diff after: 0.09203048936545488\n"
     ]
    }
   ],
   "source": [
    "y_pred_before = clf_original.predict(X_test)\n",
    "test_gender = [d[\"g\"] for d in test]\n",
    "dev_gender = [d[\"g\"] for d in dev]\n",
    "train_gender = [d[\"g\"] for d in train]\n",
    "\n",
    "tprs_before, tprs_change_before, mean_ratio_before = get_TPR(y_pred_before, Y_test, p2i, i2p, test_gender)\n",
    "similarity_vs_tpr(tprs_change_before, word2vec, \"before\", \"TPR\", prof2fem)\n",
    "#y_pred_after = clf.predict(X_test.dot(P))\n",
    "y_pred_after = clf.predict(X_test.dot(P))\n",
    "tprs, tprs_change_after, mean_ratio_after = get_TPR(y_pred_after, Y_test, p2i, i2p, test_gender)\n",
    "similarity_vs_tpr(tprs_change_after, word2vec, \"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)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tsne_by_gender(vecs, labels, title, words = None):\n",
    "\n",
    "  tsne = TSNE(n_components=2, random_state=0)\n",
    "  vecs_2d = tsne.fit_transform(vecs)\n",
    "  num_labels = len(set(labels.tolist()))\n",
    "\n",
    "  names = [\"class {}\".format(i) for i in range(num_labels)]\n",
    "  plt.figure(figsize=(6, 5))\n",
    "  colors = 'r', 'b', 'orange'\n",
    "  for i, c, label in zip(set(labels.tolist()), colors, names):\n",
    "    print(len(vecs_2d[labels == i, 0]))\n",
    "    plt.scatter(vecs_2d[labels == i, 0], vecs_2d[labels == i, 1], c=c,\n",
    "                label=label, alpha = 0.6)\n",
    "  plt.legend(loc = \"upper right\")\n",
    "  plt.title(title)\n",
    "  \n",
    "  if words is not None:\n",
    "        k = 60\n",
    "        for i in range(k):\n",
    "            \n",
    "            j = np.random.choice(range(len(words)))\n",
    "            label = labels[i]\n",
    "            w = words[j]\n",
    "            x,y = vecs_2d[i]\n",
    "            plt.annotate(w , (x,y), size = 10, color = \"black\" if label == 1 else \"black\")\n",
    "            \n",
    "  plt.show()\n",
    "  return vecs_2d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "prof = \"professor\"\n",
    "idx = np.random.rand(X_dev.shape[0]) < 0.1\n",
    "prof_idx = Y_dev == p2i[prof]\n",
    "n = 800\n",
    "tsne_by_gender(X_dev[prof_idx][:n], np.array(dev_gender)[prof_idx][:n], \"tsne by gender, before, {}\".format(prof))\n",
    "tsne_by_gender((X_dev[prof_idx].dot(P))[:n], np.array(dev_gender)[prof_idx][:n], \"tsne by gender, after. {}\".format(prof))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def count_profs(data_y, i2p):\n",
    "    d = Counter()\n",
    "    for y in data_y:\n",
    "        d[i2p[y]] += 1\n",
    "    return d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_profs(Y_train, i2p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "count_profs(Y_dev, i2p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "p0 = all_Ps[0]\n",
    "p1 = all_Ps[1]\n",
    "p2 = all_Ps[2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "ename": "SyntaxError",
     "evalue": "invalid syntax (<ipython-input-34-bb1006037f2f>, line 1)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;36m  File \u001b[0;32m\"<ipython-input-34-bb1006037f2f>\"\u001b[0;36m, line \u001b[0;32m1\u001b[0m\n\u001b[0;31m    p2.dot(p2.T)<\u001b[0m\n\u001b[0m                 ^\u001b[0m\n\u001b[0;31mSyntaxError\u001b[0m\u001b[0;31m:\u001b[0m invalid syntax\n"
     ]
    }
   ],
   "source": [
    "p2.dot(p2.T)<"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "p0 = all_Ps[0]\n",
    "p1 = all_Ps[1]\n",
    "p2 = all_Ps[2]\n",
    "\n",
    "p2.dot(p2.T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_common_words(data: List[dict], word2vec_model):\n",
    "    \n",
    "    words_counter = Counter()\n",
    "    vecs = []\n",
    "    all_words = []\n",
    "    \n",
    "    \n",
    "    for entry in tqdm.tqdm(data, total = len(data)):\n",
    "        \n",
    "        y = p2i[entry[\"p\"]]\n",
    "        words = entry[\"hard_text\"].split(\" \")\n",
    "        all_words.extend(words)\n",
    "            \n",
    "    \n",
    "    words_counter = Counter(all_words)\n",
    "    common_words = [w for w in words_counter if words_counter[w] > 10 and (w in word2vec_model)]\n",
    "    common_vecs = [word2vec_model[w] for w in common_words]\n",
    "    \n",
    "    return common_words, common_vecs\n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "words, vecs = get_common_words(train,word2vec)\n",
    "\n",
    "vecs_normed = vecs / np.linalg.norm(vecs, keepdims = True)\n",
    "ws_normed = [w/np.linalg.norm(w) for w in all_ws]\n",
    "k = 1000\n",
    "groups, labels = [], []\n",
    "\n",
    "for i,w in enumerate(ws_normed):\n",
    "    print(\"INLP ITERATION: {}\".format(i))\n",
    "    sims_i = w.dot(vecs_normed.T).squeeze(0)\n",
    "    zipped = zip(words, vecs_normed, sims_i)\n",
    "    zipped = sorted(list(zipped), key = lambda tuple: -abs(tuple[2]))\n",
    "    ws,vs, sims = list(zip(*zipped))\n",
    "    print(ws[:k])\n",
    "    groups.append(vs[:k])\n",
    "    labels.append(i)\n",
    "    print(\"------------------------------------------------------------------\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_gender_subspace(vecs, labels):\n",
    "    \n",
    "        N = len(labels)\n",
    "        \n",
    "        all_vecs = []\n",
    "        all_labels = []\n",
    "        \n",
    "        for vecs_list,l in zip(vecs, labels):\n",
    "            \n",
    "            all_labels.append(np.ones(len(vecs_list)) * l)\n",
    "        \n",
    "        all_vecs_np = np.concatenate(vecs, axis = 0)\n",
    "        all_labels_np = np.concatenate(all_labels, axis = 0)\n",
    "        tsne = TSNE(n_components=2, random_state=0)\n",
    "        vecs_2d = tsne.fit_transform(all_vecs_np)\n",
    "    \n",
    "        fig, ax = plt.subplots()\n",
    "        # define the colormap\n",
    "        cmap = plt.cm.jet\n",
    "        # extract all colors from the .jet map\n",
    "        cmaplist = [cmap(i) for i in range(cmap.N)]\n",
    "        # create the new map\n",
    "        cmap = cmap.from_list('Custom cmap', cmaplist, cmap.N)\n",
    "        # define the bins and normalize\n",
    "        bounds = np.linspace(0, N, N + 1)\n",
    "        norm = matplotlib.colors.BoundaryNorm(bounds, cmap.N)\n",
    "        print(\"here\")\n",
    "        print(all_labels_np.shape)\n",
    "        scat = ax.scatter(vecs_2d[:,0], vecs_2d[:,1], c=all_labels_np, cmap=cmap, norm=norm, alpha=0.15)\n",
    "        cb = plt.colorbar(scat, spacing='proportional')#, ticks=bounds)\n",
    "        cb.set_label(\"INLP iteration number\")\n",
    "        plt.savefig(\"INLP progress\", dpi = 600)\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#visualize_gender_subspace(groups, labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "word2vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 362,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 0.\n",
      " most_similar male: ('himself', 'his', 'he', 'Fatherhood', 'His', 'vExpert', 'Freemasonry', 'He', 'Topman', 'fatherhood', 'Grails', 'homoerotic', 'everyman', 'FanDuel', 'Masonry', 'VMworld', 'varicocele', '.His', 'journeyman', 'APEX')\n",
      "; most_similar female: ('herself', 'she', 'her', 'She', 'Her', 'hers', 'Carolyn', 'Colleen', 'Susan', 'Diane', '.She', 'Ms.', 'Kathleen', 'Chairwoman', 'Kathryn', 'Laura', 'Kathy', 'Alyssa', 'Karen', 'Tricia')\n",
      "=====================================\n",
      "gender direction 1.\n",
      " most_similar male: ('shortwave', 'E.F.', 'Henceforth', 'R.T.', 'Rattling', 'CMG', 'SGM', 'GMR', 'microstructure', 'faithfully', 'periodically', 'Wyner', 'SPT', 'impulse', 'W.R.', 'henceforth', 'J.E.', 'mesoscale', 'maestro', 'urges')\n",
      "; most_similar female: ('Childbirth', 'Yoga', 'Hatha', 'Midwifery', 'Pregnancy', 'pregnancy', 'yoga', 'Maternity', 'yoginis', 'Childbearing', 'Asana', 'YOGA', 'childbirth', 'midwifery', 'YogaWorks', 'Intimates', 'Jivamukti', 'AcroYoga', 'motherhood', 'maternity')\n",
      "=====================================\n",
      "gender direction 2.\n",
      " most_similar male: ('2410', 'Ichimura', 'Br', 'Bienkowski', 'Hedden', 'Fr', 'Furuya', 'Tsutsui', 'Uchiyama', 'Streu', 'Tirone', 'Eisch', 'Desper', 'Kawamoto', 'Oberdick', 'Sma', 'Demler', 'Bollinger', 'Dubos', 'Hepato')\n",
      "; most_similar female: ('Feminists', 'Whoopi', 'Herstory', 'feminists', 'Hillary', 'Latinos', 'dressage', 'Hispanics', 'Skechers', 'Footwear', 'Portia', 'blacks', 'Africans', 'gays', 'shoe', 'victimization', 'transgenders', 'Women', 'victimisation', 'Nkechi')\n",
      "=====================================\n",
      "gender direction 3.\n",
      " most_similar male: ('talking', 'talk', 'talked', 'saying', 'engineer', 'Gordon', 'electromagnetics', 'secondly', 'stressing', 'arguing', 'CTO', 'Electric', 'loner', 'said', 'speaking', 'explaining', 'Electromagnetics', 'electromagnetism', 'Aldous', 'Swendsen')\n",
      "; most_similar female: ('Lentils', 'Kayte', 'PopSugar', 'Blessings', 'endowments', 'heirlooms', 'ornaments', 'VAL', 'KYM', 'Scarves', 'Peppers', 'Halftime', 'Culinary', '//', 'Peas', 'Aashirwad', 'bustlines', 'Portfolios', 'Hers', 'ornament')\n",
      "=====================================\n",
      "gender direction 4.\n",
      " most_similar male: ('RSA', 'O', 'Mantle', 'softly', 'unattended', 'hushed', 'alone', 'WSC', 'undisturbed', 'Reverie', 'Lament', 'serenely', 'preindustrial', 'slower', 'multiplied', 'slowly', 'Goodnight', 'unguarded', 'Rexroth', 'Alice')\n",
      "; most_similar female: ('Pacitti', 'Athletica', 'bronchitis', 'yoga', 'myofascial', 'Paltrow', 'Chloë', 'Kurti', 'Drumline', 'Corburn', 'Mattocks', 'Sweatpants', 'Adames', 'Kesha', 'shetty', 'Vitalia', 'Antimicrobials', 'Activism', 'Fighter', '9244')\n",
      "=====================================\n",
      "gender direction 5.\n",
      " most_similar male: ('Uwe', 'Hoctor', '−', 'Ralf', 'Sadly', 'Susanne', 'Unfortunately', 'M3', '64', 'Rostock', '7154', 'Bernd', 'Jochen', '4029', 'Günter', '6914', 'Torben', 'Jürgen', 'Mickey', 'drivers')\n",
      "; most_similar female: ('CPMC', 'NEDO', 'Polyclinics', 'valorem', 'propos', 'telepresence', 'HBS', 'Hermeneutics', 'clinic-', 'Vietnam', 'tug', 'FDI', 'Manchuria', 'Sarang', 'adjudged', 'Shreveport', 'ICBC', 'valuation', 'Studies-', 'antimonopoly')\n",
      "=====================================\n",
      "gender direction 6.\n",
      " most_similar male: ('Davidians', 'Davidian', 'ensuring', 'cans', 'OD', 'persecuted', 'insuring', 'TOD', 'VO', 'knob', 'vowel', 'ameliorating', 'Michalak', 'Lip', 'ignition', 'Pedal', 'EIN', 'Tung', 'correct', 'assuring')\n",
      "; most_similar female: ('php', 'ruby', 'Embedded', 'surf', 'RPC', 'webs', 'defences', 'Hypervisor', 'Guilt', 'TCL', 'marvel', 'mommy', 'sentinel', 'ICD', 'backend', 'daydream', 'PHP', 'antivirus', 'Moorestown', 'domesticity')\n",
      "=====================================\n",
      "gender direction 7.\n",
      " most_similar male: ('whilst', 'ITS', 'ISSUES', 'FRUIT', 'Jurisdictions', 'size', 'oppressions', 'DIFFERENT', 'Heresies', 'Histories', 'inequality', 'Essays', 'Patriarchy', 'parcel', 'Myths', 'circumnavigating', 'Lavinia', 'Unpacking', 'Wherein', 'fleet')\n",
      "; most_similar female: ('Minneapolis', 'Wayzata', 'Lucado', 'Bemidji', 'Metrodome', 'HealthPartners', 'AHCA', 'Politte', 'Chanhassen', 'Novocaine', 'Wichita', 'Robbinsdale', 'KC', 'Lindstrom', 'Hazanavicius', 'SLP', 'Owatonna', 'Minn.', 'audiologists', 'otolaryngologist')\n",
      "=====================================\n",
      "gender direction 8.\n",
      " most_similar male: ('Masur', 'Clemens', 'bark', 'Gibbon', 'rowers', 'commands', 'mute', 'bin', 'pauses', 'Olberding', 'rower', 'wood', 'Holmgren', 'CHORD', 'chord', 'rowing', 'WASH', 'geese', 'Sacchi', 'SFS')\n",
      "; most_similar female: ('Scoreland', '34DD', 'Striptease', 'Curvy', 'Adalynn', 'Alona', 'Brazzers', 'Voluptuous', 'Glamorous', 'Busty', 'Sexy', 'Redhead', 'Heels', 'Boobs', 'bombshell', 'Optometrists', '36D', 'Blonde', 'Disclosure', 'Brunette')\n",
      "=====================================\n",
      "gender direction 9.\n",
      " most_similar male: ('92', '82', '84', '86', '91', '83', '76', '66', '78', '79', '94', '87', '68', '89', '88', '74', '63', '93', '71', '67')\n",
      "; most_similar female: ('OpenID', 'ITRC', 'Malware', 'Symantec', 'OSCON', 'committer', 'Xeon', 'Undertow', 'Greenplum', 'cybercrime', 'Itanium', 'Solr', 'Zeitgeist', 'PostgreSQL', 'filmography', 'committers', 'Sideshow', 'URLs', 'CERIAS', 'MySpace')\n",
      "=====================================\n",
      "gender direction 10.\n",
      " most_similar male: ('Fog', 'E-40', 'DMZ', 'Rade', 'Shaper', 'Philips', 'Profoto', 'fog', 'Sharp', 'grinder', 'Elster', 'sharpness', 'magnifies', 'Ree', 'Quit', 'Mifune', 'macros', 'Bogie', 'Doc', 'Kirk')\n",
      "; most_similar female: ('RBD', 'POMCO', 'Storytime', 'caregivers', 'Associati', 'Builders', 'preschool', 'Populus', 'GPR', 'schizotypy', 'completion', 'Nursery', 'Constructions', 'nursery', 'Pawsitively', 'Debnam', 'Paez', 'Aadya', 'BCMG', 'Tlc')\n",
      "=====================================\n",
      "gender direction 11.\n",
      " most_similar male: ('Bug', 'gamma', 'Scissorhands', 'Redmond', 'X-', 'D.from', 'America-', 'Gizmo', 'hyperactive', 'Fauvist', 'tyrosine', 'messes', 'scolded', 'Gamma', 'Hulk', 'Georgia-', 'Ladybug', 'Merah', 'bakes', 'CNET')\n",
      "; most_similar female: ('polyamory', 'intersectional', 'PrEP', 'ECHS', 'ODU', 'Equality', 'endourology', 'Archivaria', '7x7', 'Mutuality', 'photovoice', 'bisexuality', 'GBV', 'epidemiology', 'SFA', 'fabricating', 'PhotoVoice', 'GIPA', 'EdM', 'JAAPA')\n",
      "=====================================\n",
      "gender direction 12.\n",
      " most_similar male: ('Schumpeter', 'thought', 'GST', 'occurred', 'Krakatoa', 'remark', 'Question', 'INSET', 'encountered', 'What', 'HST', 'abolished', 'Pentecost', 'remembered', 'Bathtub', 'wondered', 'Remarks', 'assumed', 'Sugata', 'Exactly')\n",
      "; most_similar female: ('Stalley', 'Hartgrove', 'Ravens', 'ravens', 'falcons', 'Exempla', 'Hemby', 'caseworker', 'consignment', 'Consignment', 'Nightjar', 'Gourds', 'MCSA', 'staking', 'Waxwing', 'Stonebridge', 'Skins', 'BRIT', 'Coldplay', 'Patience')\n",
      "=====================================\n",
      "gender direction 13.\n",
      " most_similar male: ('Largent', 'Ensminger', 'Whorley', 'Kouns', 'Gooch', 'Mcquaid', 'uniforms', 'Hass', 'jerseys', 'cunt', 'glove', 'McElfresh', 'Narus', 'Moby', 'SOE', 'Pihos', 'Musil', 'sigmoidoscopy', \"'em\", 'cloth')\n",
      "; most_similar female: ('bachelors', 'newlyweds', 'Spousal', 'Brides.com', 'Kiara', 'build', 'UPR', 'couples', 'Meera', 'bachelor', 'Couples', 'Pi', 'variable', 'Camila', 'Essence.com', '⠀', 'Lifespan', 'regenerating', 'healthily', 'DIY')\n",
      "=====================================\n",
      "gender direction 14.\n",
      " most_similar male: ('mid-80s', 'EPD', 'Fuji', 'definition', 'podium', 'crews', 'Corning', 'Schindler', 'glossary', 'stairs', 'Libbey', 'Dow', 'mid-1980', 'Ballinger', 'Schwing', 'Near', 'Gorge', 'USGS', 'quarry', 'cetera')\n",
      "; most_similar female: ('untrusted', 'Abyei', 'MLB', 'Antinomianism', 'WeddingWire', 'ligament', 'Scrappy', 'cupcake', 'Vodou', 'Broncos', 'Somaliland', 'Premal', 'LLM', 'CCM', 'PURCHASE', 'Criminology', 'Yankees', 'Him', 'Ruminate', 'Peacebuilding')\n",
      "=====================================\n",
      "gender direction 15.\n",
      " most_similar male: ('flaw', 'HIDE', 'closeness', 'Owosso', 'Him', 'Lint', '\\u200b', 'predictability', 'Bonifas', 'underwear', 'Beaufort', 'THEN', 'HE', 'Allegan', 'INSTEAD', 'Audit', 'SHE', 'embolism', 'Petoskey', 'COST')\n",
      "; most_similar female: ('Mixcloud', 'XLR8R', 'soundcloud', 'dietetic', 'Vocalo', 'ill', 'Southasian', 'reportage', 'undertaking', 'Bookslut', 'modding', 'earthbound', 'hemisphere', 'PopMatters', 'JVC', 'KCRW', 'videogaming', 'febrile', 'Kerrang', 'Populous')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 16.\n",
      " most_similar male: ('Azab', 'Assi', 'Headliners', 'rammed', 'main', 'Loconte', 'AFP', 'Faux', 'Antiwar.com', 'millimeter', 'Zig', 'expounds', 'Yad', 'animating', 'MERS', 'Propagation', 'rails', 'propagation', 'extrema', 'collapsed')\n",
      "; most_similar female: ('Nest', 'poems', 'Haiku', 'Poems', 'poem', 'POEM', 'solitude', 'PANK', 'poet', 'LARK', 'Reveries', 'sweetheart', 'elegies', 'moniker', 'Hawthornden', 'Peranakan', 'sobriquet', 'Ramanujan', 'mama', 'smart')\n",
      "=====================================\n",
      "gender direction 17.\n",
      " most_similar male: ('Idiom', 'Raiders', 'Rodel', 'Swalwell', 'Raka', 'McNair', 'Spindrift', 'Mcnair', 'Vikings', 'Whiz', 'Dunmore', 'Hampton', 'OJ', 'Merriman', 'NHBA', '7011', 'Baffert', 'Tulley', 'Harlene', 'Excelsior')\n",
      "; most_similar female: ('Lambda', 'RNAs', 'Granade', 'VDI', 'Curves', 'MIRA', 'Glamour', 'LiDAR', 'Inked', 'PLOS', 'heterosexual', 'lobe', 'carnivore', 'Ceph', 'MAV', 'sexes', 'Palimpsest', 'Complutense', 'MATLAB', 'Département')\n",
      "=====================================\n",
      "gender direction 18.\n",
      " most_similar male: ('Neurofeedback', 'Ratatouille', 'pizza', 'amusement', 'XD', 'Berardinelli', 'Koontz', 'cheesy', 'Lucifer', 'neurofeedback', 'salad', 'Voted', 'Timpano', 'Drago', 'behaviorists', 'chaos', 'blaze', 'Geauga', '\\ufeff', 'Dragonfall')\n",
      "; most_similar female: ('Marginalized', 'ADU', 'BME', '0280', 'Intimacies', '0279', '28304', 'Pasifika', '54401', 'intimacies', 'VUP', '53233', 'Marshallese', 'Identities', 'Hyphen', 'macro-', 'data', 'linkages', 'Herstory', 'rowing')\n",
      "=====================================\n",
      "gender direction 19.\n",
      " most_similar male: ('Wex', 'Sherkin', 'Isley', 'Rooks', 'vultures', 'Vultures', 'Cafaro', 'externing', 'Blackfish', 'Manos', 'Gulfcoast', 'Nordlund', 'Roda', 'Buffett', 'HHC', 'Graul', 'Oddball', 'mainland', 'Healthsouth', 'Kurre')\n",
      "; most_similar female: ('cardiomyocyte', 'cystectomy', 'Querying', 'terrified', 'frustrated', 'sympathectomy', 'contractility', 'cardiomyocytes', 'scared', 'myocytes', 'amenorrhea', 'PCOS', 'Arabe', 'dysmenorrhea', 'hypertrophy', 'EHESS', 'intimidated', 'CIBSE', 'radicalised', 'intravesical')\n",
      "=====================================\n",
      "gender direction 20.\n",
      " most_similar male: ('Juraj', 'Lalande', 'Halina', 'Szent', 'Jerold', 'Keven', 'Simma', 'Luc', 'Gyorgy', 'Praha', 'Endre', 'Orlin', 'Laci', 'Floriano', 'Wery', 'Malec', 'Sonnenburg', 'Steenbergen', 'unneeded', 'Lebar')\n",
      "; most_similar female: ('100-Mile', 'MEND', 'AFSP', 'Fairey', 'Creatives', 'PlayMakers', 'hydrogel', 'Bravery', 'Tees', 'hydrogels', 'polymer', 'Toughest', 'Activism', 'mettle', 'nonviolence', 'paddleboarding', 'militancy', 'Activists', 'activism', 'flyer')\n",
      "=====================================\n",
      "gender direction 21.\n",
      " most_similar male: ('oro', 'Sumitomo', 'tution', 'immunity', 'Riken', 'Trembling', 'Cooling', 'mint', 'Below', 'Kuro', 'Mitsubishi', 'Mint', 'cooling', 'vivos', 'victor', 'Voltage', 'dispositive', 'compensation', 'Mamiya', 'orchid')\n",
      "; most_similar female: ('Mongo', 'nesting', 'charting', 'Wayfarer', 'Saundra', 'Fb', 'Angular', 'HTML5', 'Coasters', 'Greasemonkey', 'Bhagwat', 'gestation', 'Chabon', 'Shona', 'concurrently', 'Jaswinder', 'populating', 'jQuery', 'JavaScript', 'CIDR')\n",
      "=====================================\n",
      "gender direction 22.\n",
      " most_similar male: ('Impossibility', 'EMMA', 'con-', 'hus', 'Sketch', 'LIM', 'AMOS', 'in-', 'Fancy', 'the-', 'Stitch', 're-', 'REX', 'SIR', 'silico', 'tion', 'sharply', 'securely', 'sketch', 'of-')\n",
      "; most_similar female: ('Regardless', 'K-6', 'mastery', 'cyberharassment', 'geared', 'mystique', 'fan', 'sourcebook', 'resurgence', 'melee', 'Shadowrun', 'MAcc', 'Celene', 'intrigue', 'MSLS', 'prefer', 'allure', 'Internists', 'Jaclyn', 'reignite')\n",
      "=====================================\n",
      "gender direction 23.\n",
      " most_similar male: ('MoreRead', '\\u200b', 'reservations', 'consults', 'presides', 'immobilization', 'hikes', 'steps', 'Dipl', 'Hendersonville', 'walks', 'conducts', 'negotiations', 'Sandals', 'perishables', 'consultative', 'treks', 'Ca', 'Saturdays', 'weekends')\n",
      "; most_similar female: ('youtube', 'username', 'microsoft', 'Youtube', 'censored', 'yours', 'muppet', 'flickr', 'creationists', 'machinima', 'haters', 'Orkut', 'barbie', 'alright', 'deviantART', 'deniers', 'mr', 'Diddy', 'Blogging', 'xxxx')\n",
      "=====================================\n",
      "gender direction 24.\n",
      " most_similar male: ('Allegro', 'move', 'Nathans', 'Schild', 'Motta', 'typing', 'Morsani', 'VBA', 'Chromatin', 'Sparr', 'PeopleSoft', 'VB', 'Traurig', 'Nucleus', 'strings', 'handles', 'Seibel', 'IMG', 'moves', 'Weitzenhoffer')\n",
      "; most_similar female: ('Khoon', 'interventionism', 'fanciful', 'bailouts', 'Tidepools', 'Catwalk', 'fantastical', 'Keynesian', 'empiric', 'bazaars', 'Vientiane', 'Mun', 'Spooked', 'Thuy', 'TARP', 'DSGE', 'Blithe', 'Traveller', 'Selfie', 'storybooks')\n",
      "=====================================\n",
      "gender direction 25.\n",
      " most_similar male: ('Venturi', 'Orlando', 'cluttering', 'Ugo', 'scrutinized', 'McGruff', 'Nürnberg', 'Textron', 'AMG', 'Bucerius', 'loitering', 'HSV', 'DSM', 'Ricciardelli', 'Nuremberg', 'Deltona', 'Chirico', 'Bosch', 'Gotti', 'Schnee')\n",
      "; most_similar female: ('Carnatic', 'flautist', 'erhu', 'caving', 'Rupee', 'Karnatic', 'marimba', 'quartets', '1857', 'Govinda', 'NIC', 'Bansuri', 'sitar', '8-bit', 'Selah', 'Sharath', 'Metroid', 'Saira', 'rock', 'Sibelius')\n",
      "=====================================\n",
      "gender direction 26.\n",
      " most_similar male: ('Bertram', 'Bramwell', '1927', 'novelty', 'anachronism', 'foolishness', 'Cheriton', 'Lovegren', '1928', 'p', 'Fairbanks', '1908', 'inventory', 'Hirschberg', 'Blois', 'Chesterton', '1936', 'historicism', 'beehives', 'Blondel')\n",
      "; most_similar female: ('abdominoplasty', 'Banglalink', 'Pawsitively', 'Tummy', 'PYP', 'Beaconhouse', 'NOSM', 'Anam', 'VIVO', 'Colaiste', 'Seo', 'Sprouts', 'Squeeze', 'DIEP', 'inking', 'POSTECH', 'Tuna', 'Pak', 'cholangiocarcinoma', 'Critters')\n",
      "=====================================\n",
      "gender direction 27.\n",
      " most_similar male: ('stints', 'Trammell', 'presumably', 'Shoals', 'Garrigues', 'post-1980', 'Housecall', '41.2', 'orthopedist', '2009-present', 'Ambulatory', 'Emmy-', 'propelling', 'averaged', '12-time', 'dutifully', 'ESPN.com', 'tilts', 'nods', 'showrunner')\n",
      "; most_similar female: ('food', 'childish', 'sensitization', 'greedy', 'weather', 'firework', 'dermatologic', 'rain', 'hunger', 'hungry', 'science', 'itchy', 'preparedness', 'cancer', 'bitter', 'sad', 'anticancer', 'chemical', 'rainy', 'eczema')\n",
      "=====================================\n",
      "gender direction 28.\n",
      " most_similar male: ('alarmed', 'alarm', 'Convinced', 'Puzzled', 'Worried', 'postman', 'Upset', 'rung', 'clanking', 'dismay', 'shaken', 'checked', 'Initially', 'Enabled', 'sceptical', 'Undaunted', 'ringing', 'suitcases', 'suitcase', 'worried')\n",
      "; most_similar female: ('excels', 'wanna', 'transcends', 'PL', 'laude', 'Femme', 'objectively', 'did', 'idolize', 'NEED', 'exposures', '1-on-1', 'chose', 'manages', 'gender-', 'strive', 'exists', 'CF', 'crave', 'excelling')\n",
      "=====================================\n",
      "gender direction 29.\n",
      " most_similar male: ('Papierski', 'folklorist', 'ENSA', 'Wartman', 'Nogowski', 'muralist', 'WKBW', 'Kandy', 'Ngoi', 'MidAmerica', 'SPUR', 'Shereen', 'accordionist', 'Tongan', 'Raekwon', 'regionalism', 'griot', 'Castagnola', 'Dafford', 'locals')\n",
      "; most_similar female: ('Kamchatka', 'Parenthood', 'Admit', '6560', 'Concern', 'Omsk', 'Natacha', 'Watcher', 'Novosibirsk', 'Responsibility', 'FOF', 'observed', 'România', '7733', '6339', 'stalked', 'Creation', 'Numéro', 'Males', 'Females')\n",
      "=====================================\n",
      "gender direction 30.\n",
      " most_similar male: ('pic', 'mani', 'Popeye', 'hill', 'CG', 'layout', 'Charmed', 'Mcgraw', 'devour', 'eat', 'devours', 'TCM', 'Hercules', 'Thirumala', 'pics', 'Jaws', 'pickiest', 'movie', 'Hooters', 'Maxim')\n",
      "; most_similar female: ('RCAH', 'CLIR', 'P-16', 'WEFT', 'caucuses', 'Junctures', 'EGAP', 'StoryCorps', 'ELMCIP', 'Conversations', 'Pambazuka', 'Undercurrents', 'countercultures', 'Oikocredit', 'futures', 'Futures', 'DAAD', 'AICGS', 'WOLA', 'KUOW')\n",
      "=====================================\n",
      "gender direction 31.\n",
      " most_similar male: ('Maccabees', 'tank', 'crater', 'Phantoms', 'deeds', 'JNF', 'shadow', 'Avigdor', 'Bombers', 'tanks', 'spotter', 'Pravda', 'Menachem', 'Wiesel', 'Lancer', 'jets', 'Jets', 'Squadrons', 'faint', 'afford')\n",
      "; most_similar female: ('introduces', 'welcomes', 'Pinterest', 'introducing', 'imposes', 'kidnaps', 'Hardy', 'proposes', 'derails', 'Wiggs', 'teases', 'Banga', 'establishes', 'pinterest', 'adopts', '}', '{', 'appropriates', 'poses', 'proposing')\n",
      "=====================================\n",
      "gender direction 32.\n",
      " most_similar male: ('Delventhal', 'Baking', 'Littles', 'Coop', 'brainstormed', 'Crafting', 'Duesterhaus', 'Shivar', 'Grains', 'Telling', 'Trottier', 'DuVernay', 'Chaffin', 'Viticulture', 'Baranowski', 'DeMeo', 'Ashworth', 'Boydston', '@', 'Hagopian')\n",
      "; most_similar female: ('nude', 'BNA', 'iPSC', 'Venezuelan', 'BPMN', 'EPO', 'Ilia', 'Macedonian', 'humpback', 'Slovenian', 'Serbian', 'Revlon', 'Ruffian', 'Maracaibo', 'UTP', 'macaque', 'premature', 'autologous', 'PPP', 'Republic')\n",
      "=====================================\n",
      "gender direction 33.\n",
      " most_similar male: ('5871', 'Discolored', '9131', 'blistering', '6177', '6659', '7787', '4483', '7664', '2523', 'Repair', '6564', 'Boil', '6426', '5588', '2584', 'Topsham', 'Distinguish', '4888', '9059')\n",
      "; most_similar female: ('AVID', 'Anglicare', 'EMAP', 'ACAP', 'Shelter', 'cages', 'ACTF', 'Caritas', 'LSS', 'ZSL', 'SPM', 'Takahe', 'INGOs', 'cage', 'DIVE', 'shifters', 'MassHealth', 'welfare', 'Nari', 'SAS')\n",
      "=====================================\n",
      "gender direction 34.\n",
      " most_similar male: ('POLA', 'heuristic', 'Sunga', 'webmasters', 'surfers', 'heuristics', 'scripts', 'Ahdoot', 'Asian', 'roberto', 'BitTorrent', 'ruthless', 'empires', 'inexperienced', 'Umeko', 'Sapan', 'actresses', 'Atsuko', 'Savalia', 'Vipul')\n",
      "; most_similar female: ('celebrating', 'nuptials', 'Watertown', 'wedding', 'Moraine', 'Mass', 'celebratory', 'celebration', 'vouchers', 'MSP', 'Celebrate', 'celebrate', 'AF', 'tethered', 'festivities', 'baptized', 'Stillwater', 'jubilant', 'voucher', 'captured')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 35.\n",
      " most_similar male: ('poodle', 'fluffy', 'poodles', 'Goldendoodle', 'schnauzer', 'ditzy', 'darling', 'dreamy', 'delicate', 'dachshunds', 'Javaheri', 'leggy', 'angelic', 'Farshid', 'Julee', 'Zilker', 'Chelsi', 'scrumptious', 'Heidi', 'embellishments')\n",
      "; most_similar female: ('Ss', 'Corman', 'Handed', 'Sell', 'Fly', 'LD', 'Pay', 'dimorphism', 'Eynde', 'indicator', 'ACP', 'Coffin', 'Ca', 'Aware', 'Paid', 'Bottle', 'Differently', 'rites', 'pay', 'Detect')\n",
      "=====================================\n",
      "gender direction 36.\n",
      " most_similar male: ('Papanikolaou', 'Bersani', 'Chihara', 'Violi', 'Bernardi', 'Risch', 'Hamori', 'Appearing', 'Fabius', 'aplomb', 'Kokas', 'invitations', 'Papaconstantinou', 'Bartschi', 'Camilleri', 'Noblet', 'Yokomizo', 'Albanese', 'premiers', 'Frattini')\n",
      "; most_similar female: ('SAI', 'Burrough', 'MMS', 'RISC', 'OI', 'MMU', 'SLF', 'FLL', 'HMP', 'shorted', 'BAH', 'LEF', 'CMP', 'VCO', 'Tektronix', 'HUM', 'BPC', 'EME', 'DFN', 'NI')\n",
      "=====================================\n",
      "gender direction 37.\n",
      " most_similar male: ('Bang', 'Money', '6x6', '1015', '1017', 'mattered', 'Whole', 'Sufi', 'Picasso', 'Creativity', 'bang', 'Imagination', 'disclosed', 'Door', 'probed', 'Indiatimes', 'MPS', 'Tv', 'stimulated', 'Led')\n",
      "; most_similar female: ('Addison', 'Kingwood', 'Trish', 'ARMS', 'Ophthalmologist', 'Oncologist', 'Olmsted', 'birthmark', 'Skeeters', 'brachioplasty', 'Phoebe', 'Elissa', 'Machemer', 'SouthEast', 'hyperhidrosis', 'Histiocytosis', 'Pearland', 'Radburn', 'Litchman', 'Settlers')\n",
      "=====================================\n",
      "gender direction 38.\n",
      " most_similar male: ('Fatales', 'Adrien', 'Jackal', 'Vargas', 'Maldonado', 'colonia', 'Madara', 'Decree', 'Racecourse', 'Diaz', 'Alphas', 'ingrained', 'Minardi', 'ruins', 'Mayweather', 'stables', '1787', 'Velasquez', '1777', 'Adonis')\n",
      "; most_similar female: ('segmental', 'Siemens', 'Vinyasa', 'hatha', 'Hatha', 'Falun', 'Timpe', 'Davita', 'Raychem', 'interrelatedness', 'Gaiam', 'Nortel', 'CorePower', 'G.E.', 'Sears', 'Conveyor', 'offshoring', 'Wipro', 'Cathay', 'HPB')\n",
      "=====================================\n",
      "gender direction 39.\n",
      " most_similar male: ('Bar', 'CMJ', 'Porthole', 'Scribes', 'Bands', 'Jockey', 'Jukebox', 'Dein', 'Skull', 'Tribe', 'Medicus', 'Slippers', 'Morrisey', 'Fedora', 'Fan', 'Slack', 'Lesion', 'Cantor', 'Congressional', 'radiologists')\n",
      "; most_similar female: ('̶', 'ending', 'canola', 'frying', 'demystifying', 'repurposing', 'lifecycle', 'underbanked', 'personne', 'GLA', 'maki', 'Waititi', 'honey', 'noncommunicable', 'fluffy', 'countering', '\\u200b', 'mixture', 'aHUS', 'dripping')\n",
      "=====================================\n",
      "gender direction 40.\n",
      " most_similar male: ('Resurgent', 'wagons', 'formations', 'post-2003', 'Chargers', 'cowboys', 'Confederate', 'dismantled', 'marshal', 'subnational', 'Yeovil', 'Haverfordwest', 'Batteries', 'postbellum', 'Callies', 'Division', '01655', 'dyads', 'declining', 'institutionalization')\n",
      "; most_similar female: ('Gain', 'Achieve', 'Lifestyle', 'lifestyle', 'Ideal', 'Perfect', 'Faizan', 'Striving', 'IDEAL', 'Helping', 'Livelihood', 'Upscale', 'Worldly', 'Tangible', 'taste', 'Helps', 'comfort', 'Able', 'Viterbi', 'degrees')\n",
      "=====================================\n",
      "gender direction 41.\n",
      " most_similar male: ('Woon', 'investigating', 'Quek', 'Sims', 'Wait', 'Wadley', 'Balazs', 'Hmm', 'Tue', 'Rowell', 'Posted', 'Thu', 'glitch', 'Flux', 'dragging', 'Teo', 'underwater', 'investigate', 'Yawn', 'Josey')\n",
      "; most_similar female: ('donna', 'ArtScroll', 'ings', 'reprinted', 'Twiztid', 'jerseys', 'Bioidentical', 'telugu', 'bioidentical', 'marathi', 'Tradizionale', 'Sorelle', 'nce', 'Champe', 'clit', 'ISLS', 'cation', 'inserts', 'convent', '34DD')\n",
      "=====================================\n",
      "gender direction 42.\n",
      " most_similar male: ('broadsides', 'Kimmel', 'hammered', 'rapped', 'podium', 'Guggenheim', 'premieres', 'Premiered', 'emcee', 'NYSE', 'amNewYork', 'bearer', 'Verizon', 'Shorenstein', 'emcees', 'date', 'WNYC', 'news', 'engraved', 'syphilis')\n",
      "; most_similar female: ('Annai', 'River', 'gradient', 'Samiti', 'Territory', 'PATHS', 'Sturt', 'Unschooling', 'gradients', 'NLP', 'Warlpiri', 'Godavari', 'Sharavathi', 'Shivpuri', 'Siang', 'Shipibo', 'ILP', 'Rivers', 'Swamp', 'Mahavidyalaya')\n",
      "=====================================\n",
      "gender direction 43.\n",
      " most_similar male: ('Was', 'Tired', 'Got', 'Not', 'Did', 'Can', 'Wanted', 'Became', 'Want', 'Are', 'Should', 'Thrilled', 'Need', 'Reluctant', 'Largely', 'Never', 'Shall', 'sat', 'Hated', 'Would')\n",
      "; most_similar female: ('worded', 'generalize', 'generalization', 'classifies', 'contextualized', 'Senge', 'operationalize', 'Boutique', 'behave', 'heading', 'Mr.', 'implement', 'Ms.', 'Fashions', 'accordance', 'spelled', 'contextually', 'actionable', 'categorize', 'interpretable')\n",
      "=====================================\n",
      "gender direction 44.\n",
      " most_similar male: ('monotonous', 'restrictive', 'monotonic', 'furthermore', 'leukemic', 'stratified', 'TRIPS', 'hurry', 'Malawian', 'shapeless', 'STAM', 'Noth', 'deepness', 'abolish', 'PEL', 'burdensome', 'quantity', 'Andriessen', 'psychedelic', 'eternal')\n",
      "; most_similar female: ('athlete', 'ProActive', 'Jaguar', 'GoPro', 'sports', 'geolocation', 'botnets', 'Dasher', 'Giffin', 'analytics', 'Proactive', 'operations', 'arthroscopy', 'Analytics', 'Citi', 'USAA', 'botnet', 'Nike', 'KeyBank', 'Hoppes')\n",
      "=====================================\n",
      "gender direction 45.\n",
      " most_similar male: ('Recep', 'Foxworthy', 'Grainger', 'Tracie', '18th', 'Shanna', '19th', 'McKnight', 'Dorman', 'Seine', 'Ferrell', 'Rudisill', '20th', 'Selz', 'Gena', 'Garey', 'Shaun', 'Ronna', 'Ferguson', 'Shellie')\n",
      "; most_similar female: ('Populi', 'Portfolios', 'Modest', 'Espacio', 'applets', 'Impetus', 'substantial', 'Abierta', 'Ercol', 'educative', 'reforms', 'Magma', 'Veldt', 'compulsory', 'Finale', 'sizeable', 'ETFs', 'movable', 'Constructive', 'Nimbus')\n",
      "=====================================\n",
      "gender direction 46.\n",
      " most_similar male: ('Fade', 'Rated', 'Statistical', 'corrections', 'Rating', 'Correction', 'CCTV', 'Revealing', 'fugitive', 'handcuffed', 'Flash', 'aid', 'Imager', 'fading', 'aided', 'hiding', 'Exterior', 'handcuffs', 'Interior', 'manipulated')\n",
      "; most_similar female: ('Essig', 'Ellzey', 'Nespresso', 'Teeter', 'Zachman', 'Coles', 'Brockmeier', 'Bombeck', 'Herweg', 'Diedrich', 'coworking', 'Bendel', 'Malouf', 'kubernetes', 'sustainer', 'RBD', 'Friese', 'apt', 'savannas', 'Specht')\n",
      "=====================================\n",
      "gender direction 47.\n",
      " most_similar male: ('fraught', 'UCD', 'tense', 'earlier', 'Globes', 'metre', 'metres', 'Holi', '9NEWS', 'calmer', 'Firstpost', 'Starry', 'hackathon', 'Festive', 'Moments', 'contentious', 'tantrums', 'improvised', 'Sawant', 'Cochabamba')\n",
      "; most_similar female: ('handyman', 'Handyman', 'presses', 'Woodbury', 'Chipley', 'woodworker', 'Chubb', 'Brooksville', 'Woodworker', 'Wendell', 'Grandy', 'Batesville', 'Homeowner', 'watercolorist', 'Pawling', 'insures', 'carpenter', 'Bookkeeper', 'Clewiston', 'jobbing')\n",
      "=====================================\n",
      "gender direction 48.\n",
      " most_similar male: ('Elvira', 'juvenile', 'juveniles', 'PPU', 'Adolf', 'stork', 'Englund', 'Seghers', 'Francoist', 'Komala', 'Polanski', 'BAD', 'Kuczynski', 'Malathi', 'Showgirls', 'inmates', 'Böll', 'Picazo', 'homosexual', 'lesbianism')\n",
      "; most_similar female: ('Haptic', 'ADAS', 'Embedded', 'QTI', 'webcasts', 'Adaptive', 'Modality', 'Stream', 'webcast', 'podcasts', 'Planner', 'Cutoff', 'dominate', 'Sensory', 'Quizlet', 'podcast', 'mLearning', 'Cognitive', 'MassCUE', 'Analyzer')\n",
      "=====================================\n",
      "gender direction 49.\n",
      " most_similar male: ('Navodaya', 'Bahraini', 'Saudi', 'Hardinge', 'Corea', 'Amal', 'Mayas', 'Yucatan', 'Bahrain', 'Emirati', 'Filipinas', 'HBSC', 'Bengaluru', 'aspirants', 'Deccan', 'satires', 'Naipaul', 'Malabar', 'Sidhwa', 'Lakdawala')\n",
      "; most_similar female: ('Stimulus', 'inventory', 'Infusion', 'Wubs', 'increases', 'Sunnyview', 'quantities', 'Bonus', 'Yarn', 'infusions', 'overdoses', 'irrationally', 'NETA', 'Omega-3', 'Artie', 'Organismic', 'regenerating', 'decrease', 'baby', 'Collectibles')\n",
      "=====================================\n",
      "gender direction 50.\n",
      " most_similar male: ('reformulation', '￼', 'Celaya', 'neurotoxin', 'Cengage', 'Shh', 'option', 'Ayyash', 'explicit', 'Alembic', 'adding', 'adrenalectomy', 'adobe', 'panicked', 'Schrager', 'smell', 'AJE', 'D.E.A.', 'rhinoplasties', 'Answer')\n",
      "; most_similar female: ('imagers', 'Hitachi', 'GMD', 'Mirka', 'Portable', 'Imke', 'imager', 'Miniature', 'Appreciation', 'polymath', 'Princesses', 'Outings', 'Ribbons', 'Tennis', 'ODD', 'CTO', 'Roundtables', 'breakthroughs', 'holders', 'Slippers')\n",
      "=====================================\n",
      "gender direction 51.\n",
      " most_similar male: ('fats', '4100', '5900', '5650', '2555', '2562', '7050', '4231', 'body', '5500', '6700', '5600', '8200', 'Matkin', 'bodys', '1725', 'valves', '1820', '9200', 'Thomassen')\n",
      "; most_similar female: ('Guerilla', 'Posters', 'Peep', 'Penned', 'Guerrilla', 'Ads', 'Superbowl', 'guerrilla', 'Advertisements', 'Tagline', 'Informer', 'Poster', 'Political', 'guerilla', 'NOON', 'Apologies', 'TYPO', 'Monocle', 'Dumbest', 'glimpsed')\n",
      "=====================================\n",
      "gender direction 52.\n",
      " most_similar male: ('applets', 'implementations', 'pixie', 'Atomics', 'Atol', 'SSA', 'LineRate', 'SoCs', 'caching', 'carousel', 'arbitrarily', 'Semiconductor', 'Chubby', 'Architecting', 'FTF', 'workforces', 'Daiichi', 'Vanger', 'ActiveX', '1637')\n",
      "; most_similar female: ('M.d', 'HH', 'upcycling', 'canvass', 'URBZ', 'reserach', 'tous', 'Ranawat', 'C.S', 'lifestyle', 'eco-', 'Rav', 'passions', 'Studying', 'Kolb', 'leather', 'Choma', 'Involvement', 'Shailee', 'ArtLab')\n",
      "=====================================\n",
      "gender direction 53.\n",
      " most_similar male: ('blocked', 'injured', 'totaled', 'brakes', 'Nobis', 'Coffman', 'Gutterman', 'redirected', 'Majus', 'Meyers', 'Garns', 'Kaster', 'Karman', 'Bahr', 'ligaments', 'Alman', 'Amster', 'Shoulder', 'Lumbar', 'angles')\n",
      "; most_similar female: ('Piagetian', 'blackboard', 'Melanesian', 'Unpacking', 'SciPy', 'precollege', 'eResearch', 'numeracy', 'Statistical', 'macrobiotic', 'ISME', 'Nonparametric', 'Honiara', 'Behaviours', 'Dietary', 'Inference', 'lice', 'Representational', 'powerpoint', 'schoolwide')\n",
      "=====================================\n",
      "gender direction 54.\n",
      " most_similar male: ('Linings', 'threaded', 'Drenner', 'scrapers', 'donning', 'gray', 'newsreaders', 'Keaton', 'Fracturing', 'CFTC', 'opaque', 'Geoscientists', 'Ridgeway', 'Carbide', 'piecing', 'CJA', 'tungsten', 'dominate', 'processors', 'Drilling')\n",
      "; most_similar female: ('wanderlust', 'Francophile', 'Philosophie', 'hazard', 'Condé', 'Toots', 'concomitance', 'Docent', 'Lettres', 'una', 'Voltaire', 'Gallico', 'Passi', 'CNAP', 'VD', 'foodie', 'Montesquieu', 'Stendhal', 'Giannuzzi', 'Galatians')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 55.\n",
      " most_similar male: ('Cenac', 'Gilfillan', 'Ghouse', 'Rothberg', 'Talwar', 'McGillivray', 'Loughlin', 'Maclean', 'Purdey', 'Schorr', 'Spinella', 'Saperstein', 'manifold', 'Carrell', 'Mazzei', 'Valiant', 'Rigg', 'Chisholm', 'Advani', 'Ruffalo')\n",
      "; most_similar female: ('AUP', 'cyberpunk', 'tos', 'Troopers', 'interlocal', 'assignement', 'netlabel', 'GNU', 'PED', 'Aomori', 'Paintsville', 'po', 'intermedia', 'carpal', 'github', 'Symfony2', 'ville', 'Lona', 'Kaiapoi', 'parapsychological')\n",
      "=====================================\n",
      "gender direction 56.\n",
      " most_similar male: ('Palatka', 'Slocumb', 'Wests', 'Opdyke', 'LaCrosse', 'Wasco', 'Designers', 'Kucha', 'Dowdy', 'Kaukauna', 'Westfield', 'Lamson', 'Should', 'Brookville', 'Goodyear', 'Wainscott', 'Bechers', 'Elco', 'Allworth', 'Millville')\n",
      "; most_similar female: ('caseworker', 'Sociale', 'penpal', 'Volunteering', 'monitor', 'adjust', 'Reintegration', 'LGBTI', 'monitoring', 'Aftercare', 'clinic-', 'volunteering', 'see', 'waitlists', 'reintegration', 'VSO', 'culturale', 'TPR', 'organise', 'EDGI')\n",
      "=====================================\n",
      "gender direction 57.\n",
      " most_similar male: ('Sketchbook', 'Binders', 'Sketchbooks', 'Manuscript', 'Entries', 'Prose', 'Templates', 'Bookbinding', 'Inventories', 'Handal', 'Pages', 'Zine', 'Everyday', 'Conscientious', 'Calendars', 'Jernej', 'Page', 'Blurb.com', 'Critiques', 'Binder')\n",
      "; most_similar female: ('Afrobeats', 'Nigerian', 'landmine', 'Ukranian', 'Cameroonian', 'kidnapping', 'Congolese', 'scaling', 'grappling', 'collision', 'abduction', 'dodging', 'explosion', 'Brazillian', 'grapples', 'survivor', 'Ukrainian', 'homing', 'dual', 'shapeshifting')\n",
      "=====================================\n",
      "gender direction 58.\n",
      " most_similar male: ('AskMen', 'Mitochondrion', 'apostasy', 'Hashemi', 'Jackets', 'uterus', 'Women', 'sexism', 'misogyny', 'Ghaddar', 'imports', 'womens', 'Polarize', 'andropause', 'equivalents', 'famine', 'breadwinner', 'post-1970s', 'huh', 'women')\n",
      "; most_similar female: ('preschoolers', 'BES', 'PreK', 'preschool', 'identified', 'SEP', 'Briarwood', 'PreK-3', 'planted', 'MAR', 'Montessori', 'disabled', 'Playworld', 'belonging', 'NCSS', 'kindergarten', '0022', '0280', 'ILD', 'Educare')\n",
      "=====================================\n",
      "gender direction 59.\n",
      " most_similar male: ('Chafee', 'Lasell', 'Babson', 'CFMA', 'Applicability', 'Subsection', 'Heuristics', 'loopholes', 'Shepard', 'Predoctoral', 'G.S.', 'Paragraph', 'Greylock', 'Incivility', 'Shepherds', 'Arbitrator', 'ECE', 'MANETs', 'Quarters', 'Sentence')\n",
      "; most_similar female: ('FreeOnes', 'nudes', 'Garters', 'striptease', 'Playboy', 'softcore', 'ladyboy', 'Bayamon', 'Dansk', 'pictorials', 'GGurls', 'Kata', 'annexation', 'arousing', 'porno', 'Nudes', 'babe', 'FHM', 'pornographic', 'Kmart')\n",
      "=====================================\n",
      "gender direction 60.\n",
      " most_similar male: ('whipped', 'Moore', 'Calvin', 'Stringfellow', 'Gray', 'Lacroix', 'unquestionably', 'Switzer', 'Heade', 'Ingels', 'Mayhew', 'Katt', 'cornerstones', 'Hobbes', 'Hendricks', 'Mugler', 'screenshots', 'Mapp', 'McCray', 'stripes')\n",
      "; most_similar female: ('4th-8th', '4th-6th', 'Seatown', 'ureter', '10th-12th', 'chondrocytes', 'Raise', 'Inguinal', 'Carpal', 'Fermoy', 'carpel', 'Gorkem', 'Iraqi', 'microtia', 'Cartilage', 'DePuy', 'Zilvinas', 'cartilage', 'Rachna', 'orthopaedic')\n",
      "=====================================\n",
      "gender direction 61.\n",
      " most_similar male: ('choline', 'Tau', 'UAlbany', 'Estrich', 'Brahma', 'Tok', 'theologically', 'QSM', 'xoJane', 'theorize', 'Gok', 'Apatow', 'ecofeminism', 'exegesis', 'Misogyny', 'eggs', 'Har', 'chant', 'Kra', 'Ecofeminism')\n",
      "; most_similar female: ('Children', 'children', 'braces', 'Activity', 'marionettes', 'kids', 'Activities', 'kid', 'kiddos', 'Kids', 'Interests', 'Adults', 'childrens', 'Braces', 'impairments', 'LOVED', 'child', 'strabismus', 'rates', 'minimum')\n",
      "=====================================\n",
      "gender direction 62.\n",
      " most_similar male: ('utilizing', 'Merriman', 'Traumatology', 'incorporates', 'utilizes', 'AAOS', 'cardiothoracic', 'knee', 'VATS', 'spinal', 'thoracic', 'echos', 'rearfoot', 'clavicle', 'articular', 'Employing', 'patellofemoral', 'UAA', 'dislocated', 'preexisting')\n",
      "; most_similar female: ('Bus', 'Stonehenge', 'Nasik', 'bus', 'Siddhi', 'Hezbollah', 'Harihareshwar', 'Gangapur', 'Poona', 'Kingfisher', 'Satguru', 'P1', 'Porsche', '004', 'Haridwar', 'Holi', 'Hitech', 'Kalam', 'Belur', 'Atheists')\n",
      "=====================================\n",
      "gender direction 63.\n",
      " most_similar male: ('Gird', 'forecasting', 'Sanghi', 'OND', 'Ahya', 'GBM', 'someday', 'EEE', 'sis', 'A380', 'engraftment', 'B.E', 'Gab', 'IBRAHIM', 'Fratello', 'Isetan', 'Spencers', 'plotting', 'Shifa', 'Defy')\n",
      "; most_similar female: ('contested', 'Symbolism', 'marginalised', 'devalued', 'Comprehensive', 'Engaging', 'Critically', 'Chaplin', 'Importance', 'unrepresented', 'Meanings', 'Cultural', 'grappled', 'minoritized', 'Communities', 'disempowered', 'Problematizing', 'stigmatized', 'Reframing', 'Ethnography')\n",
      "=====================================\n",
      "gender direction 64.\n",
      " most_similar male: ('depraved', 'uncivilized', 'barbarism', 'dandyism', 'civilized', 'barbaric', 'Abdellah', 'atrocity', 'decadence', 'heinous', 'Sade', 'abomination', 'southerner', 'Moslem', 'squalid', 'disgusting', 'Timbuktu', 'mutilated', 'Libertines', 'stigmatized')\n",
      "; most_similar female: ('Busines', 'teambuilding', 'Business', 'Collider', 'clients', 'Agility', 'Pros', 'pipeline', 'strategy', 'solopreneurs', 'Dragon', 'Teamwork', 'hurdles', 'skillsets', 'chances', 'queries', 'Propeller', 'business', 'strategies', 'workstations')\n",
      "=====================================\n",
      "gender direction 65.\n",
      " most_similar male: ('Khar', 'Firmly', 'Mangal', 'Aashirwad', 'Kuber', 'Essel', 'Shahi', 'tenement', 'Deal', 'mogul', 'victimisation', 'Kashi', 'Atta', 'Ansari', 'Panchsheel', 'Jayshree', 'Ambika', 'Khurshid', 'Poddar', 'Listed')\n",
      "; most_similar female: ('Liaisons', 'Wartburg', 'Overtime', 'downtime', 'UGA', 'Chattahoochee', 'clinicals', 'CLCE', 'uniforms', 'exam', 'Cerner', 'Coding', 'CCHS', 'Knoxville', 'Corrections', 'evolves', 'Revolutions', 'labors', 'Benefis', 'Portneuf')\n",
      "=====================================\n",
      "gender direction 66.\n",
      " most_similar male: ('This', 'Biannual', 'Such', 'That', 'AMA', 'Panchali', 'Ruchira', 'Generally', 'It', 'ISCM', 'Certain', 'One', 'Another', 'Having', 'VONA', 'Aosdána', 'Biopics', 'lifetime', 'The', 'MELUS')\n",
      "; most_similar female: ('overtones', 'silliness', 'profitably', 'continued', 'absurdity', 'resumed', 'misadventures', 'difficulties', 'aggressively', 'supersymmetric', 'Carlinville', 'begin', 'excels', 'Qs', 'familiar', 'singled', 'resorting', 'strangeness', 'commonplaces', 'resorted')\n",
      "=====================================\n",
      "gender direction 67.\n",
      " most_similar male: ('ColorLines', 'Kondabolu', 'undo', 'MLJ', 'peep', '0170', 'Transgress', 'Colorlines', 'diff', 'disproportionality', 'Theol', 'Mic', 'Gest', 'Tuebingen', 'Deutsche', 'injustices', 'Hochschule', 'Tanaz', 'gap', '0066')\n",
      "; most_similar female: ('occupants', 'perishable', 'warm', 'possessions', 'warmer', 'blankets', 'transient', 'responders', 'dependable', 'Oliviero', 'Mustang', 'Statesville', 'nonspecific', 'Warm', 'Fernandina', 'linens', 'Dependable', 'porous', 'Ormond', 'Kennealy')\n",
      "=====================================\n",
      "gender direction 68.\n",
      " most_similar male: ('copyright', 'copyrights', 'copyleft', 'chemokines', 'dementias', 'guilds', 'ribosome', 'copyrighted', 'Faeries', 'Sigcomm', 'URBZ', 'pathogenic', 'KONAMI', 'Legalities', 'chemotaxis', 'mafias', 'Crytek', 'goblins', 'Giants', 'AIAS')\n",
      "; most_similar female: ('hs', 'pediatrician', '12-years', 'Worth', 'Children', 'Coroner', 'child', 'recommended', 'Diagnostician', 'P.E.', 'Pediatrician', 'willingness', 'Jett', 'Breckinridge', 'tire', 'physician', 'Tilden', 'Holabird', 'Northshore', 'Snorkel')\n",
      "=====================================\n",
      "gender direction 69.\n",
      " most_similar male: ('Sociedad', 'Torcuato', 'lexical', 'Quechua', 'Akademi', 'psycholinguistic', 'Musa', 'Azerbaijani', 'Emiri', 'Iñupiat', 'Maulana', 'GLS', 'Sangeet', 'Vygotsky', 'Republika', 'Sahrawi', 'Cristóbal', 'IALA', 'Uzbekistan', 'Circulo')\n",
      "; most_similar female: ('repaired', 'Lyme', 'Doylestown', 'enough', 'Weld', 'civilly', 'repair', 'fix', 'Herkimer', 'DHMC', 'Sherbourne', 'EDAC', 'Fishkill', 'Paltz', 'Bullett', 'Fries', 'inexpensively', 'Cozen', 'Ober', 'reconfigure')\n",
      "=====================================\n",
      "gender direction 70.\n",
      " most_similar male: ('IDG', 'Piaget', 'undisputed', 'elevates', 'catwalk', 'outlets', 'disputed', 'ADN', 'neuroanatomical', 'Huffington', 'Enquirer', 'MoMa', 'claims', 'MLB.com', 'Radar', 'catwalks', 'tuatara', 'celeb', 'elevated', 'mainstream')\n",
      "; most_similar female: ('Compiled', 'environnment', 'sands', 'Theme', 'Coins', 'Winds', 'Cleansing', 'Runes', 'Used', 'Wishing', 'Horseman', 'percussions', 'Seaside', 'Conducting', 'Rider', 'Liberating', 'Crusade', 'Bountiful', 'Prosperous', 'Kindler')\n",
      "=====================================\n",
      "gender direction 71.\n",
      " most_similar male: ('estimations', 'Suono', 'BBC3', 'Yankers', 'Bomba', 'interference', 'genus', 'linguists', 'PCS', 'Mosca', 'hermaphrodite', 'Milli', 'Armonica', 'ranking', 'unscientific', 'Grub', 'spin', 'Scienza', 'Squad', 'MIMO')\n",
      "; most_similar female: ('Monadnock', 'Timeline', 'fairmont', 'Amiel', 'Wiser', 'overnight', 'AJN', 'lade', 'Pilgrim', 'Sendai', 'Staying', 'Fairmont', 'mindful', 'vigilant', 'Badertscher', 'cookies', 'Lurene', 'Carmel', 'awhile', 'HIE')\n",
      "=====================================\n",
      "gender direction 72.\n",
      " most_similar male: ('Karr', 'MacIntyre', 'Shrout', 'Laird', 'Heit', 'Bluedorn', 'MacLean', 'Empirically', 'Braden', 'Rossman', 'Melnyk', 'Aitken', 'Tatman', 'Shefrin', 'Kornfield', 'Kendall', 'Russell', 'Laughing', 'Cowen', 'Krupnick')\n",
      "; most_similar female: ('sweltering', 'humid', 'stopover', 'showpiece', 'spiders', 'knockout', 'migrant', 'pulsating', 'bays', 'torrid', 'humidity', 'dusky', 'fixture', 'ingress', 'hotbed', 'injection', 'pivotal', 'afterhour', 'spider', 'teeming')\n",
      "=====================================\n",
      "gender direction 73.\n",
      " most_similar male: ('optioned', 'SMOKE', 'unmarketable', 'jobbing', 'ASPECT', 'emblazoned', 'Backlash', 'REGION', 'MuscleMag', 'veto', 'Relapse', 'flexes', 'noticeable', 'Hellraiser', 'occasional', 'début', 'Kerrang', 'flung', 'birthmark', 'pisses')\n",
      "; most_similar female: ('coastal', 'riverine', 'Vital', 'Nurul', 'Nasrul', 'calming', 'analytics', 'sophisticated', 'upazila', 'mangrove', 'Erudite', 'modernizing', 'cardiothoracic', 'defense', 'Pantai', 'Tarik', 'Anam', 'wetlands', 'mangroves', 'Boko')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 74.\n",
      " most_similar male: ('entice', 'not', 'Peirce', 'Durgin', 'teaser', 'Courant', 'vagueness', 'adjectives', 'descriptive', 'Rendell', 'persuade', 'ought', 'Propositions', 'Culliton', 'description', 'EPPA', 'Lussier', 'coin', 'might', 'propositions')\n",
      "; most_similar female: ('Zhanna', 'choreographing', 'logistical', 'Eurosport', 'YMT', 'choreographs', 'Rostov', 'Baku', 'aerials', 'upmost', 'spectating', 'Shanghainese', 'epicentre', 'parental', 'hora', 'T1D', 'bespoke', 'orchestrates', 'Dynastic', 'interruptions')\n",
      "=====================================\n",
      "gender direction 75.\n",
      " most_similar male: ('mythological', 'Maratha', 'ideological', 'nationalistic', 'historical', 'nationalist', 'militarized', 'microbiological', 'heritage', 'historiographic', 'bureaucratic', 'Ossetian', 'scientific', 'mythical', 'Bicentenary', 'Odissi', 'missionary', 'Afrikaner', 'Kuka', 'delegations')\n",
      "; most_similar female: ('admit', 'Bornstein', 'LL', 'Mcgee', 'Kaz', 'Biddle', 'Attell', 'Wiz', 'Hebb', 'Reggie', 'Golson', 'McNally', 'Ullman', 'Burrell', 'Macklin', 'Ivey', 'Darnell', 'Holzman', 'Spades', 'Sacks')\n",
      "=====================================\n",
      "gender direction 76.\n",
      " most_similar male: ('Mazzaglia', 'Plale', 'Crosbie', 'Gallaway', 'cents', 'indictment', 'Heifner', 'twenties', 'Wardrobe', 'Keg', 'Flinn', 'lb', 'barge', 'Geddings', 'storesThe', 'comment', 'Reckless', 'Shields', 'ANY', 'Boon')\n",
      "; most_similar female: ('CGS', 'Endocrinology', 'Shaare', 'HIAS', 'Kedar', 'Openhouse', 'SFVAMC', 'KCRW', 'Outpatient', 'institutionalization', 'asylum', 'Terman', 'Shalem', 'Yeshurun', 'Rivka', 'CARECEN', 'cardiology', 'Schwartzman', 'Asylum', 'Perlin')\n",
      "=====================================\n",
      "gender direction 77.\n",
      " most_similar male: ('Nines', 'Dirksen', 'humiliation', '2441', 'disfigurement', 'Lehnert', 'embarrassment', 'Notes', 'Pietrangelo', 'legs', 'meningioma', '2147', 'Luebbert', 'memories', '4144', 'Boram', 'Yusaf', 'Rhinos', 'Aamer', 'iceberg')\n",
      "; most_similar female: ('scrutinizing', 'UPV', 'scrutinize', 'Perform', 'Regulate', 'naturopathic', 'naturopath', 'MovNat', 'Inverse', 'lauded', 'conducting', 'Undertaking', 'surprised', 'parapsychologist', 'Conducting', 'prover', 'Proactive', 'scrutinized', 'correlating', 'Prove')\n",
      "=====================================\n",
      "gender direction 78.\n",
      " most_similar male: ('VxWorks', 'WRL', 'Uludağ', 'panoramas', 'HiRISE', 'Norikura', 'Keweenaw', 'panorama', 'Stegner', 'Yangtze', 'Alexanderplatz', 'PanoTools', 'Sharp', 'windsurfing', 'daylighting', 'stitching', 'AOFAS', 'CME', 'Dolin', 'hydroelectric')\n",
      "; most_similar female: ('Idols', 'Politiques', 'normalities', 'crushes', 'Valedictorian', 'princes', 'Mentality', 'precepts', 'outcasts', 'mantras', 'buddies', 'vows', 'idols', 'misfits', 'acts', 'Ideals', 'pranks', 'brothers', 'Ideologies', 'kings')\n",
      "=====================================\n",
      "gender direction 79.\n",
      " most_similar male: ('Korsakov', 'shipbuilding', 'Tula', 'Grandin', 'Georgi', 'neuroprotection', 'Calva', 'Vor', 'Perm', 'Nizhny', 'neuroprotective', 'arts', 'soy', 'NMS', 'gar', 'Canman', 'Bast', 'novo', 'Crucible', 'polyaromatic')\n",
      "; most_similar female: ('privacy', 'Knesset', 'smile', 'Privacy', 'Ridges', 'knob', 'millimeter', 'wholeness', 'consent', 'Erlbaum', 'unbearable', 'smiles', 'Dharamsala', 'obstacle', 'apart', 'Hearing', 'continuously', 'Smiling', 'Dharamshala', 'Osho')\n",
      "=====================================\n",
      "gender direction 80.\n",
      " most_similar male: ('Atol', 'thieves', 'jewelery', 'shielding', 'theft', 'dwarfs', 'Hoppé', 'Bielik', 'Stieg', 'generators', 'jewellery', 'Hugo', 'thefts', 'Ekstrand', 'jewllery', 'Koscielniak', 'masks', 'Oppermann', 'crooks', 'Steffan')\n",
      "; most_similar female: ('Refresh', 'Commanding', 'readmission', 'YLD', 'reminiscing', 'Rotation', 'Nittany', 'GSAPP', 'Transitioning', 'Dining', 'Forward', 'commanding', 'converged', 'Transgress', 'Admissions', 'Moderation', 'Pivot', 'Readmissions', 'Command', 'Healthy')\n",
      "=====================================\n",
      "gender direction 81.\n",
      " most_similar male: ('Outsiders', 'unsurprisingly', 'Retailers', 'assortment', 'Bigger', 'ACEP', 'lineup', 'fanfare', 'Issue', 'traditionally', 'pervade', 'Hoops', 'retailers', 'fewer', 'Freshmen', 'lotta', 'Princesses', 'Mavericks', 'wraparound', 'seemingly')\n",
      "; most_similar female: ('crucis', 'Bacillus', 'artemisinin', 'Basil', 'cerevisiae', 'Mathan', 'Edapally', 'whistleblower', 'whistleblowing', 'Hanbury', 'Perumbavoor', 'Shoba', 'multiferroic', 'Dhanya', 'Sreenivasan', 'SAPA', 'Soham', 'Katherine', 'Huy', 'testimonio')\n",
      "=====================================\n",
      "gender direction 82.\n",
      " most_similar male: ('Sabah', 'Huay', 'Sarawak', 'Khai', 'Tok', 'Topaz', 'Dian', 'Yue', 'Fen', 'ranger', 'Amulet', 'Lao', 'Kedah', 'Haji', 'Kalimantan', 'Tabin', 'Idris', 'Joss', 'Shou', 'Shang')\n",
      "; most_similar female: ('shootings', 'Victimology', 'Obsessive', 'Futurism', 'Nihilism', 'Shootings', 'femicide', 'fixation', 'casein', 'Fascists', 'gynecomastia', 'molestation', 'Interventions', 'Lacanian', 'PCI', 'infill', 'Compulsive', 'Psychopathy', 'Postmodern', 'autodidact')\n",
      "=====================================\n",
      "gender direction 83.\n",
      " most_similar male: ('epitomized', 'modeled', 'seasons', 'furthered', 'variation', 'explored', 'variations', 'phenology', 'demonstrated', 'facilitated', 'reinterpretation', 'explore', 'characterized', 'NCHA', 'summarized', 'season', 'identified', 'characterize', 'host', 'oversaw')\n",
      "; most_similar female: ('e-', 'pos', 'photosensitive', '0', 'Godrej', 'Akansha', 'Thio', 'HCL', 'Narmada', 'Veldt', 'l', 'Crawshaw', 'Lesen', 'orang', 'Akshaya', '10.0', 'Arnab', 'Akruti', 'Vats', 'Amitav')\n",
      "=====================================\n",
      "gender direction 84.\n",
      " most_similar male: ('MPC', 'Starfighter', 'Cranberry', 'pilots', 'touched', 'glide', 'Landings', 'landings', 'Pilots', 'Eke', 'Drosophila', 'SPB', 'AJK', 'fondled', 'Aikins', 'Neonatal', 'PAs', 'FCS', 'gliders', 'nurses')\n",
      "; most_similar female: ('skepticism', 'Zócalo', 'disagreement', 'Corbis', 'edX', 'cultura', 'mistrust', 'Postmodernity', 'distrust', 'cocina', 'originates', 'belief', 'Environmentalism', 'agrees', 'indemnification', 'clause', 'environmentalism', 'Belief', 'wealthy', 'presuppositions')\n",
      "=====================================\n",
      "gender direction 85.\n",
      " most_similar male: ('widow', 'Nicu', 'gure', 'heroic', '90th', 'prince', 'RAF', 'Molar', 'child-', 'tenth', 'Hardest', 'Hedi', 'fresco', 'molar', 'Heimlich', 'infant', 'twentieth', 'Airman', 'selfless', '0.7')\n",
      "; most_similar female: ('EVP', 'TRIP', 'expands', 'BUILD', 'Bioenergy', 'energy', 'SHIFT', 'generated', 'soundscapes', 'INNS', 'REC', 'MAKE', 'Trek', 'mapped', 'unleashes', 'Geographics', 'Quantum', 'contracts', 'generates', 'MIND')\n",
      "=====================================\n",
      "gender direction 86.\n",
      " most_similar male: ('Handwritten', 'Mornington', 'alphabet', 'alphabets', 'handwritten', 'Phrases', 'letters', 'Hawkesbury', 'Complicated', 'Solitaire', 'Typing', 'sums', 'Entertaining', 'Valentines', 'scribbling', 'Ab', 'variations', 'verses', 'Manly', 'emoji')\n",
      "; most_similar female: ('FLACSO', 'UHM', 'MESA', 'generalist', 'spatialization', 'Schimmelpfennig', 'rehabilitates', 'interdisciplinary', 'coordinates', 'colloquium', 'restores', 'phenomenologist', 'MSSW', 'Commonweal', 'NAU', 'Codirector', 'gathers', 'UABC', 'SDI', 'EHESS')\n",
      "=====================================\n",
      "gender direction 87.\n",
      " most_similar male: ('grown', 'aggregated', 'Velvet', 'cultivars', 'Charcoal', 'matured', 'Ash', 'gotten', 'beetle', 'pest', 'beetles', 'darker', 'Pug', 'grained', 'mite', 'Bark', 'Marl', 'ash', 'Beetles', 'mature')\n",
      "; most_similar female: ('emerita', 'RAICES', 'Iker', '\"The', 'speechwriter', 'Migrantes', 'veteran', 'Macri', 'veterans', 'CARECEN', 'lector', 'Sojourners', 'newfound', 'pedagogue', 'sainted', 'Agustín', 'LIRS', 'embattled', 'reassignment', 'reinsertion')\n",
      "=====================================\n",
      "gender direction 88.\n",
      " most_similar male: ('Amongst', 'cobalt', 'aqua', 'Buchenwald', 'dyes', 'Plaisir', 'Among', 'Arkansan', 'pistachio', 'Graduating', 'abalone', 'scrolls', 'Brevis', 'augmentative', 'polka', 'alphabetical', 'din', 'Theresienstadt', 'conjugation', 'Etoiles')\n",
      "; most_similar female: ('sanctioning', '7420', 'partnering', 'don’t', 'consider', 'Gateway', 'rethink', 'reconsider', 'Saldana', 'partnership', 'reinterpreting', 'Northlake', 'Watts', 'accusing', 'Partners', 'Windermere', 'reexamine', 'Wellmont', 'ProMedica', 'upfront')\n",
      "=====================================\n",
      "gender direction 89.\n",
      " most_similar male: ('conveniently', 'contemporaneously', 'paperless', 'effectively', 'Orwellian', 'unopposed', 'Zeitgeist', 'occurring', 'circulate', 'obtain', 'robotically', 'Conveniently', 'erasure', 'alternatively', 'TDSB', 'damning', 'Korydallos', 'vacate', 'Nightline', 'simultaneously')\n",
      "; most_similar female: ('lakeside', 'neuroendocrinology', 'Bassmaster', 'Sig', 'camaraderie', 'luthier', 'picnic', 'waterfowl', 'AKC', 'boating', 'trad', 'motorcycling', 'SUT', 'wheeler', 'motorsports', 'Mountaineering', '4-wheeling', 'Bavarian', 'bluegrass', 'family')\n",
      "=====================================\n",
      "gender direction 90.\n",
      " most_similar male: ('Fahmi', 'SOAS', 'Ishai', 'Yunan', 'Rachman', 'Shenkar', 'Icelandic', 'depths', 'BookFest', 'thee', 'Sadia', 'Barkat', 'Badawi', 'Sallam', 'Gan', 'Yiddishland', 'internationalization', 'Aku', 'Demain', 'Arab')\n",
      "; most_similar female: ('timeshare', 'Hepatitis', 'HCV', 'Homefront', 'hepatitis', 'DSG', 'rubella', 'HTLV', 'cytomegalovirus', 'Escalera', 'immaculate', 'Heizer', 'HSV', 'Telfair', 'sniper', 'HSV-2', 'Fluor', 'HCA', 'mesothelioma', 'Hartmayer')\n",
      "=====================================\n",
      "gender direction 91.\n",
      " most_similar male: ('homelessness', 'consolidations', 'metamorphosis', 'Allegiance', 'bullying', 'Leaps', 'WLRN', 'reshape', 'Menino', 'reinvention', 'transformation', 'redistricting', 'reform', 'reshaped', 'reshapes', 'Homelessness', 'jowl', 'Newhart', 'Philanthropies', 'leaps')\n",
      "; most_similar female: ('instructing', 'impregnated', 'Aymar', 'DF', 'camping', 'B.V.', 'visiting', 'Savasta', 'PF', 'Haemostasis', 'applicable', 'responsible', 'Découverte', 'seducing', 'für', 'C.G', 'guarding', 'Shobo', 'Camping', 'LifeForce')\n",
      "=====================================\n",
      "gender direction 92.\n",
      " most_similar male: ('jumpsuit', 'wallpaper', 'aqua', 'Estella', 'Snakeskin', 'Lenor', 'longing', 'Anthropologie', 'blouse', 'Barbarella', 'retro', 'maxi', 'saloon', 'Voyeur', 'strapless', 'frills', 'costar', 'love', 'tunic', 'neon')\n",
      "; most_similar female: ('FRONTLINE', 'Churchill', 'DOCUMENTARY', 'Documentary', 'Shatz', 'Grodzinski', 'historiography', 'WhoWhatWhy', 'ICH', 'sovereign', 'historiographies', 'RUSI', 'SPOTLIGHT', 'Frankfurter', 'Gatekeepers', 'MRC', 'OZY', 'Katzenstein', 'Rosenfeld', 'Kelner')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 93.\n",
      " most_similar male: ('shoulders', 'weakened', 'modifier', 'competed', 'compete', 'shoulder', 'grappled', 'braces', 'bows', 'decorate', 'weaker', 'qualifiers', 'solidify', 'Braces', 'PRODUCTS', 'top', 'icing', 'piping', 'grabs', 'squeezed')\n",
      "; most_similar female: ('reincarnation', 'Airpark', 'Ctr', 'madness', 'painless', 'Tun', 'eugenics', 'resort', 'millenials', 'martinis', 'Deming', 'Spira', 'Tread', 'folly', 'Th', 'fusion', 'Interval', 'Reynolda', 'paradoxes', 'quaint')\n",
      "=====================================\n",
      "gender direction 94.\n",
      " most_similar male: ('cup', 'kissed', 'Shade', 'kiss', 'adoration', 'Sailor', 'Lalah', 'stroll', 'infatuation', 'download', 'escort', 'Prince', 'Nerds', 'shake', 'Ishq', 'Alto', 'Morning', 'sake', 'Sake', 'Nylon')\n",
      "; most_similar female: ('ACLA', 'RFE', 'extracorporeal', 'ILSI', 'NAACL', 'AstraZeneca', 'Paleontological', 'Mallinckrodt', 'EMNLP', 'Medstat', 'CJA', 'ILP', 'Iterative', 'IPPE', 'AHFMR', 'CDCI', 'nongovernmental', 'CIGNA', 'LINGUIST', 'CRNI')\n",
      "=====================================\n",
      "gender direction 95.\n",
      " most_similar male: ('reporter', 'Lasswell', 'columnist', 'ambushes', 'insurgents', 'guerrillas', 'snipers', 'Laird', 'covertly', 'Irregulars', 'contributor', 'subversive', 'Flanagin', 'editor', 'columnists', 'intrepid', 'ink', 'UWEX', 'obscure', 'ambush')\n",
      "; most_similar female: ('Bugatti', 'orofacial', 'shoulders', 'MB', 'Cycladic', 'Bardot', 'EMST', 'Chiron', 'soreness', 'motions', 'Rearing', 'Motions', 'torque', 'Heals', 'dystocia', 'Porsche', 'pain', '600k', 'Karting', 'Transgenerational')\n",
      "=====================================\n",
      "gender direction 96.\n",
      " most_similar male: ('shortages', 'cosmic', 'shortage', 'unexplained', 'slowdown', 'inexplicable', 'ECCC', 'Lightning', 'enclave', 'finish', 'suffered', 'Lucky', 'cluster', 'Zodiac', 'experiencing', 'pieced', 'LepreCon', 'glitch', 'sweater', 'pavilion')\n",
      "; most_similar female: ('Ranges', 'Ultrasonography', 'ESEM', 'OpenLayers', 'AOI', 'ABPP', 'Depth', 'Agar', 'provably', 'DKI', 'VASER', 'AI2', 'caudatum', 'Instants', 'Paramecium', 'elegans', 'Transanal', 'Blacher', 'reflectometry', 'LESS')\n",
      "=====================================\n",
      "gender direction 97.\n",
      " most_similar male: ('roasting', 'Interregional', 'NMT', 'Dierks', 'Bulk', 'harvesting', 'lymphadenopathy', 'bulk', 'transplanting', 'cultivation', 'CO2', 'DWI', 'DeCordova', 'epidural', 'stimulation', 'interregional', 'intrathecal', 'transnasal', 'FCI', 'Biochar')\n",
      "; most_similar female: ('Prophecy', 'castle', 'Sectarianism', 'Aaj', 'Shahrzad', 'kingdom', 'Hazar', 'Shades', 'sworn', 'Foscari', 'dynasty', 'Synopsys', 'ruins', 'Naagin', 'torn', 'Loyalty', 'Number', 'Malahide', 'Tide', 'Qubool')\n",
      "=====================================\n",
      "gender direction 98.\n",
      " most_similar male: ('Internationaux', 'des', 'Mille', 'éditions', 'Des', 'Héros', 'Nouvelle', 'Ritz', 'Seuil', 'historique', 'Figues', 'flags', 'Etoiles', 'Nemours', 'Cues', 'Sequins', 'décoratifs', 'Souper', 'Au', 'accrued')\n",
      "; most_similar female: ('K-20', 'IT', 'sane', '587', 'interoperability', 'ARUP', 'standardize', 'screwball', 'SOME', 'OF', 'practicality', 'standardization', '0595', 'TO', 'PG', 'consensual', '3566', '787', 'OUT', '0694')\n",
      "=====================================\n",
      "gender direction 99.\n",
      " most_similar male: ('Cela', 'Lancet', 'last.fm', 'Asterism', 'detract', 'Kritya', 'Aosdána', 'Permafrost', 'Juste', 'Societas', 'cru', 'Gilius', 'icicles', 'eFiction', 'Exaudi', 'lowlands', 'Requiem', 'Klinikum', 'Ploughshares', 'Vita')\n",
      "; most_similar female: ('K-12', 'classroom', 'Toddler', 'nanny', 'instructional', 'Kids', 'schooling', 'Top', 'laminate', 'sitter', 'education', 'prekindergarten', 'Whitson', 'Teacher', 'kids', 'TEACHER', 'daycare', 'Chalkboard', 'Parents', 'preK')\n",
      "=====================================\n",
      "gender direction 100.\n",
      " most_similar male: ('sigma', 'Studien', 'Husserl', 'Parmenides', 'ab', 'magna', 'Epistemological', 'Lateral', 'venia', 'nos', 'docendi', 'Angle', 'DEXA', 'Extremity', 'Sigma', 'lateral', 'Gadamer', 'Hegel', 'Ser', 'Noven')\n",
      "; most_similar female: ('entertainer', 'lifestyle', 'consumption', 'vacationing', 'writer', 'rainforests', 'wordsmith', 'raconteur', 'newsman', 'florist', 'pollution', 'TODAY.com', 'Lifestyles', 'pollutant', 'Journatic', 'Madang', 'Rotarian', 'vaping', 'lifestyles', 'gossip')\n",
      "=====================================\n",
      "gender direction 101.\n",
      " most_similar male: ('Feaga', 'Farivar', 'Jurden', 'Horrocks', 'woodwork', 'Moylan', 'Csaky', 'inbreeding', 'loyalism', 'Bowyer', 'Swilley', 'Broadhead', 'Waddell', 'Partin', 'attracts', 'interstellar', 'Barwick', 'PMTrump', 'cholangitis', 'Vrana')\n",
      "; most_similar female: ('methodologies', 'supplier', 'modalities', 'authenticate', 'Standard', 'franchised', 'technologies', 'Presses', 'Centric', 'anew', 'Agencia', 'APress', 'STANDARD', 'GHD', 'validated', 'Supplier', 'Eastern', 'OCM', 'BPM', 'Presently')\n",
      "=====================================\n",
      "gender direction 102.\n",
      " most_similar male: ('Cenegenics', 'adulthood', 'Olympics', 'academy', 'IAE', 'Boz', 'intertemporal', 'storefront', 'Academy', 'widowhood', 'promotion', 'Paralympics', 'Phelps', 'NCQA', 'Viator', 'Menopause', 'adolescence', 'unmanaged', 'bandwagon', 'Grasshopper')\n",
      "; most_similar female: ('Charged', 'Arrest', 'Mesna', 'rinse', 'Needed', 'Nd', 'Meth', 'Convicted', 'washed', 'ore', 'n', 'syringe', 'în', 'fue', 'Florida1', 'cleaned', 'wiping', 'Arrestee', 'wash', 'Laundering')\n",
      "=====================================\n",
      "gender direction 103.\n",
      " most_similar male: ('Below', 'Warning', 'lengths', '10-Year', 'Above', '7-year', 'perils', 'BAC', 'dystopias', '5-year', 'above', 'NWSA', 'dangers', 'DEF', '25-year', 'Passengers', 'predicting', '10-year', 'below', 'UOWD')\n",
      "; most_similar female: ('Senat', 'Chatter', 'dello', '·', 'Pancake', 'informally', 'mojo', 'tele', 'Small', 'Boutique', 'Vall', 'Bayard', 'refocused', 'Dudy', 'Tele', 'Longuet', 'informal', 'Arbogast', 'Mon', 'Kermit')\n",
      "=====================================\n",
      "gender direction 104.\n",
      " most_similar male: ('Cressy', 'Groene', 'Gregoire', 'pitting', 'hardening', 'knot', 'Nationals', 'pitted', 'Stier', 'pits', 'Dufour', 'Sevigny', 'DeVos', 'bulging', 'Fournier', 'Pulled', 'PHOTOS', 'Kolbe', 'VSBA', 'Mises')\n",
      "; most_similar female: ('assistive', 'sedentary', 'Coltrane', 'downside', '77030', 'Ornette', 'bebop', 'R01', 'Sobh', '61801', 'cognitive-', 'neurorehabilitation', 'ergonomic', 'Nasrullah', 'paraplegic', 'underserved', 'accelerometry', 'upside', 'ambulatory', 'UCB')\n",
      "=====================================\n",
      "gender direction 105.\n",
      " most_similar male: ('altitude', 'topographic', 'potty', 'glacier', 'woodland', 'subalpine', 'crater', 'glaciers', 'glaciated', 'Garmin', 'snowpack', 'Polar', 'Fitbit', 'meadow', 'carb', 'U.S.-led', 'vegetation', 'Snowdon', 'elliptical', 'forecasters')\n",
      "; most_similar female: ('Karpal', 'demands', 'essence', 'Onstage', 'spirit', 'Sabai', 'DeVita', 'expressions', 'Jayan', 'theatrics', 'Playwright', 'Danley', 'pathos', 'Sincerity', 'Silat', 'epitome', 'expression', 'Jax', 'alive', 'playwright')\n",
      "=====================================\n",
      "gender direction 106.\n",
      " most_similar male: ('campfire', 'bonfire', 'alight', 'Tiene', 'shag', 'Farewell', 'fiddle', 'Primates', 'pants', 'rockin', 'shirtless', 'chop', 'underwear', 'flannel', 'escribir', 'willy', 'baggy', 'pee', 'Thriller', 'scent')\n",
      "; most_similar female: ('Commerce', 'Mercatus', 'FNB', 'Corporation', 'Marchenko', 'Kharkov', 'Yaroslav', 'Kiron', 'Blum', 'Sigel', 'Starkman', 'Thaler', 'DDA', 'CEC', 'JDC', 'voter', 'KMC', 'IFAS', 'TVM', 'Maltz')\n",
      "=====================================\n",
      "gender direction 107.\n",
      " most_similar male: ('Liceo', 'Primaria', 'Bonum', 'Colegio', 'liceo', 'Cinco', 'Tétouan', 'Admiral', 'Sabancı', 'Fondo', 'Gulbenkian', 'Shorenstein', 'acromegaly', 'Infantil', '1516', 'Stomatology', 'toreros', 'prognosis', 'leukemia', 'divorced')\n",
      "; most_similar female: ('Anywhere', 'Everywhere', 'raves', 'Bump', 'bump', 'stream', 'Requests', 'Devotees', 'Ventured', 'logs', 'Raves', 'Recommendations', 'butters', 'streams', 'Massive', 'speedups', 'snowpack', 'suggestions', 'Utube', 'Deeper')\n",
      "=====================================\n",
      "gender direction 108.\n",
      " most_similar male: ('Bringing', 'Rumor', 'rediscovers', 'fetched', 'contention', 'Stays', 'revived', 'stokes', 'benchmark', 'tumbled', 'GALLERY', 'bringing', 'Feature', 'Reviving', 'Refreshing', 'News', 'tumbling', 'Quite', 'alive', 'gone')\n",
      "; most_similar female: ('NJMS', 'Preprofessional', 'Muhimbili', 'Oculoplastics', 'Sokoine', 'Trabalhadores', 'sonography', 'dotting', '01089', 'ARDMS', 'S.U.N.Y.', 'InterAmerican', 'PUCRS', 'USCIS', 'NYCDOE', 'Sonography', '02115', 'UCEDD', 'Javits', '02114')\n",
      "=====================================\n",
      "gender direction 109.\n",
      " most_similar male: ('probationary', 'unreasonable', 'reasonable', 'qualify', 'prohibitive', 'subsidize', 'unfair', 'justify', 'fairer', 'gyms', 'rid', 'ranks', 'Disciplined', 'preferable', 'rank', 'accustomed', 'tiered', 'normal', 'memberships', 'fair')\n",
      "; most_similar female: ('Depuy', 'Duralde', 'Karlheinz', 'Lore', 'Gallenberger', 'Lado', 'Calva', 'Bayona', 'Morell', 'Ramón', 'McNeel', 'folklore', 'produced', 'Himmat', 'Paldi', 'Rickman', 'Gérard', 'Kirschstein', 'woodcarver', 'Brochet')\n",
      "=====================================\n",
      "gender direction 110.\n",
      " most_similar male: ('structure-', 'Preacher', 'buckle', 'Cites', 'Scarface', 'Xzibit', 'mLearning', 'Poker', 'theÂ', 'ACCU', '85-year', 'Reparations', 'Grudge', 'Kidnapped', 'Executions', 'W.Va', 'Hakes', 'Heckler', 'Nazi', 'bloodline')\n",
      "; most_similar female: ('Sagar', 'vernal', 'Ganymede', 'COMMUNICATION', 'Saptarshi', 'Slava', 'oneiric', 'Krishan', 'Vasant', 'moon', 'Gagan', 'Kripa', 'Vitebsk', 'Shveta', 'sublimation', 'Neeti', 'Anu', 'Metamorphoses', 'Jivan', 'Kenora')\n",
      "=====================================\n",
      "gender direction 111.\n",
      " most_similar male: ('Leawood', 'covenant', 'HASSELL', 'walkers', 'equating', 'zombies', 'sorghum', 'L.L.M.', 'conveyancer', 'Mahayana', 'RapidResponse', 'restatements', 'Shaara', 'MDes', 'mindless', 'Courtyard', 'grit', 'CADRE', 'NRP', 'pointillism')\n",
      "; most_similar female: ('Bald', 'XXX', 'Stenn', 'cam', 'webcam', 'FreeOnes', 'Compilation', 'Vixon', 'webcams', 'Cheeks', 'Tapes', 'pornstar', 'Deshon', 'Booty', 'ModelMayhem', 'Hotties', 'Playboy', 'Cam', 'Playgirl', 'Starr')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 112.\n",
      " most_similar male: ('securitized', 'numerics', 'topologies', 'topology', 'topological', 'syllabic', 'hygiene', 'securitizations', 'notated', 'insulated', 'multicenter', 'SCNM', 'superconducting', 'loops', 'maintained', 'essentialized', 'modulated', 'collateralized', 'DownBeat', 'tamed')\n",
      "; most_similar female: ('usher', 'Marketing', 'Labor', 'scraping', 'Ministry', 'Gribble', 'Capitalist', 'Rafat', 'Shuck', 'Winstanley', 'scouring', 'poking', 'graveyard', 'dawn', 'Victim', 'Brogden', 'Scraping', 'Deese', 'pit', 'Usher')\n",
      "=====================================\n",
      "gender direction 113.\n",
      " most_similar male: ('GAP', 'DSM', 'Outback', 'Voyagers', 'MPI', 'Numero', 'GEO', 'TSI', 'ION', 'exposing', '8410', 'Rorvik', 'Eggert', 'OnEarth', 'scanning', 'metadata', 'loader', 'offending', 'scanned', 'Vagabond')\n",
      "; most_similar female: ('Jimi', 'denies', 'B.Ed', 'lawful', 'Hanafi', 'faculties', 'Faculties', 'Singhania', 'Hendrix', 'Subha', 'M.M', 'Truvy', 'Bisexual', 'say', 'B.D.S', 'M.ed', 'Piercings', 'recollection', 'pupil', 'Boitnott')\n",
      "=====================================\n",
      "gender direction 114.\n",
      " most_similar male: ('perturb', 'regression', 'Mantle', 'amplification', 'reproducible', 'Stochastic', 'progenitors', 'Reproducing', 'Kidner', 'purify', 'ethnoarchaeology', 'Progeny', 'variants', 'generative', 'bless', 'emulation', 'hybridisation', 'dilute', 'Gush', 'variant')\n",
      "; most_similar female: ('nut', 'Compliance', 'Backpacker', '2013-', 'UNI', 'Priorities', 'traineeship', 'obligations', 'COOKIE', 'priorities', 'Dietitian', 'FISMA', 'PROFILE', 'taxes', 'Resolutions', 'FEDERAL', 'Budget', 'compliance', 'Worksite', 'reimbursements')\n",
      "=====================================\n",
      "gender direction 115.\n",
      " most_similar male: ('rehearsing', 'frantic', 'afternoon', 'tuck', 'Shortly', 'chore', 'nervous', 'habit', 'Cue', 'hurried', 'prologue', 'Zoot', 'beastly', 'finish', 'frantically', 'Lunaris', 'Queue', 'dreaded', 'Mallinger', 'Queuing')\n",
      "; most_similar female: ('ministries', 'churches', 'NTUA', 'Churches', 'regionalization', 'MRE', 'decentralization', 'Geodesy', 'MEG', 'ACN', 'TEC', 'DYS', 'dioceses', 'Regionalization', 'Steubenville', 'denominational', 'CREC', 'groundwater', 'generalizability', 'Ministries')\n",
      "=====================================\n",
      "gender direction 116.\n",
      " most_similar male: ('Defensive', 'Conestoga', 'Grossmont', 'EHow', 'Bears', 'Claiborne', 'Evangelista', 'polarity', 'Ledyard', 'Diggs', 'Lyndsie', 'Offense', 'Bollier', 'Nichelson', 'Thiel', 'Ayinde', 'offense', 'Tice', 'Dighton', 'Kalyn')\n",
      "; most_similar female: ('SPR', 'glazed', 'nan', 'begs', 'core', 'HMC', 'infill', 'Supercomputing', 'Enhancements', 'Endemic', 'Kielder', 'biomedically', 'abstraction', 'oncologists', 'Gaeilge', 'Owler', 'parallelization', 'shaders', 'Aged', 'sip')\n",
      "=====================================\n",
      "gender direction 117.\n",
      " most_similar male: ('3686', 'Alder', 'Blazer', '4064', 'Millican', 'occasion', 'Nephi', 'reads', 'fervour', 'disappoint', 'nan', 'votes', 'Temuco', 'interpretation', 'Lafferty', 'Duff', 'Carrico', 'nothing', 'Vizenor', 'Smithville')\n",
      "; most_similar female: ('wealthiest', 'externalizing', 'Externalization', 'Disappearing', 'richest', 'Benefits', 'Ruining', 'Decline', 'downloadable', 'Rethinking', 'Maximizing', 'earners', 'reconceptualizing', 'Strides', 'Profits', 'Interdependent', 'Effects', 'maximizing', 'benefits', 'Minimized')\n",
      "=====================================\n",
      "gender direction 118.\n",
      " most_similar male: ('swollen', 'debt', 'benign', 'recovers', 'Ganges', 'warlike', 'cognates', 'Metaphilosophy', 'debts', 'Bharata', 'recuperates', 'Douro', 'doubtful', 'securitized', 'entangled', 'withdraws', 'Persianate', 'republicanism', 'reorganize', 'degraded')\n",
      "; most_similar female: ('Keys', 'Mims', 'Wiggs', 'classroom', 'FOUR', 'Leichman', 'Pointer', 'Crayton', 'SIX', 'Sherling', 'Row', 'Nevermind', 'KIRA', 'WARD', 'keys', 'Angello', 'row', 'Parkview', 'Jonell', 'Moshi')\n",
      "=====================================\n",
      "gender direction 119.\n",
      " most_similar male: ('watched', 'Kalamazoo', 'Rapids', 'Kalida', 'Benda', 'Saginaw', 'deciding', 'faced', 'Openness', 'Dowagiac', 'Allegan', 'Chauncey', 'Farmington', 'Troy', 'unfold', 'Marquette', 'MI', 'WDIV', 'MAKE', 'Jalen')\n",
      "; most_similar female: ('pointillism', 'Belotero', 'immunohistochemistry', 'Juvederm', 'Radiesse', 'brushwork', 'monochromes', 'anatomical', 'Byun', 'Smilex', 'Joo', 'Volbella', 'museological', 'Precise', 'Shevchenko', 'MCZ', 'Choo', 'Ishida', 'Kybella', 'Swelling')\n",
      "=====================================\n",
      "gender direction 120.\n",
      " most_similar male: ('TSC2', 'terpenoid', 'mysteriously', 'Slack', 'remote', 'JVM', 'conf', 'headaches', 'middleware', 'Headaches', 'bullied', 'RxJS', 'multiplatform', 'Confederated', 'streaming', 'binds', 'Piscataway', 'inexplicably', 'Fios', 'integrations')\n",
      "; most_similar female: ('Strauss', 'Eisenstein', 'Ethical', 'ethically', 'Artistic', 'Bucharest', 'Braun', 'Domestic', 'Urban', 'Museology', 'Ettinger', 'responsibly', 'Oster', 'Oberholzer', 'Disegno', 'Metropolis', 'Curatorship', 'portray', 'Ethically', 'Singer')\n",
      "=====================================\n",
      "gender direction 121.\n",
      " most_similar male: ('actionable', 'achieving', 'Graduating', 'attaining', 'graduating', 'earning', 'garnering', 'achievable', 'constructive', 'stabilizing', '4228', 'brightest', 'implementers', 'implementable', 'MCP', 'CPIP', 'feedback', 'smartest', 'sabotaging', 'rejoining')\n",
      "; most_similar female: ('hymnals', 'Passenger', 'bilingually', 'Englishes', 'Trapp', 'Houten', 'Sharan', 'Ahlstrom', 'Weiter', 'Dindy', 'van', 'Wyk', 'vocabulary', 'Dass', 'Proulx', 'B.V.', 'Vistar', 'Luso', 'Homero', 'Uden')\n",
      "=====================================\n",
      "gender direction 122.\n",
      " most_similar male: ('DF', 'parametric', 'SSG', 'EOS', 'Eos', 'ROC', 'TOR', 'TAI', 'LX', 'LIV', 'RES', 'Ruina', 'MCF', 'MF', 'DSG', 'DFC', 'BEG', 'SM', 'GK', 'APG')\n",
      "; most_similar female: ('Educationally', 'selfhood', 'psychologies', 'portrayals', 'antidotes', 'undermined', 'enactments', 'authentically', 'confrontations', 'injurious', 'Positively', 'transpersonal', 'Americanization', 'educationally', 'selves', 'detrimental', 'Mäori', 'harmful', 'undesirable', 'Māori')\n",
      "=====================================\n",
      "gender direction 123.\n",
      " most_similar male: ('Vision', 'Syntha', 'Adagio', 'Harmony', 'K7', 'Solace', 'potentials', 'Bandy', 'intentions', 'Concerto', 'Astell', 'Gleneagles', 'Revelation', 'irons', 'lifes', 'twin', 'No.2', 'GH4', 'Swope', '4AD')\n",
      "; most_similar female: ('Smithsonian', 'Testudo', 'Polynesian', 'Wendy', 'mammal', 'SCBWI', 'crocodile', 'Capitals', 'penguins', 'D.C.', 'je', 'PowerPoint', 'Kissing', 'penguin', 'Penguins', 'WTOP', 'Disneyland', 'Cafeteria', 'doughnut', 'Jafar')\n",
      "=====================================\n",
      "gender direction 124.\n",
      " most_similar male: ('internalize', 'suffer', 'perish', 'TIRR', 'Inflammation', 'Immunodeficiency', 'Thrombophilia', 'Immunobiology', 'Retardation', 'sit', 'Sperm', 'Ischemia', 'perpetuate', 'Syndrom', 'Psychophysiology', 'jumble', 'Supremacist', 'plastered', 'Thrombosis', 'Tumor')\n",
      "; most_similar female: ('Cudahy', 'Kildare', 'discreet', 'Las', 'Meath', 'Tipperary', 'Cesar', 'Broadbent', 'g.', 'stepson', 'informal', 'involving', 'Clare', 'Daly', 'Calabasas', 'Hacienda', 'Fiore', 'attending', 'Richie', 'Wicklow')\n",
      "=====================================\n",
      "gender direction 125.\n",
      " most_similar male: ('worldwide', 'adventurers', 'globally', 'arises', 'trachomatis', 'Discontinuity', 'rift', 'Keenly', 'Globally', 'motorcyclists', 'Ahmadabad', 'Mpumalanga', 'Gauteng', 'polarized', 'Maharastra', '%', 'region', 'rifts', 'Raigad', 'wherein')\n",
      "; most_similar female: ('Adelphi', 'iSchool', 'Widener', 'REU', 'ODU', 'ORNL', 'Towson', 'EdD', 'SU', 'UMBC', 'NYIT', '4-volume', 'SJSU', 'instructorship', 'ORAU', 'UMD', 'assistantship', 'JHU', 'Michener', 'UConn')\n",
      "=====================================\n",
      "gender direction 126.\n",
      " most_similar male: ('Tiered', 'vertically', 'discipled', 'Align', 'concentric', 'approx', 'Kukatpally', '3-tier', 'horizontally', '2x', '946', 'catalyze', 'interconnects', '2223', 'catalyse', 'mentored', 'spaced', 'labelled', 'Ø', 'tiered')\n",
      "; most_similar female: ('contraband', 'rhinovirus', 'Harpold', 'smuggling', 'animating', 'NMU', 'foreignness', 'impersonations', 'objectionable', 'weariness', 'Orientalist', 'impersonation', 'Bribery', 'enforcing', 'customs', 'Bers', 'asthma', 'Wassenaar', 'Mounties', 'vacationing')\n",
      "=====================================\n",
      "gender direction 127.\n",
      " most_similar male: ('propeller', 'AIJ', '2035', 'propellers', 'impossible', '2025', 'computationally', 'Goncharov', 'Farooqui', 'Abidi', 'Shestakov', 'Mehndiratta', 'nonlinearities', 'microfluidics', 'nonconvex', 'Ruchita', 'metabarcoding', '2030', 'MFC', 'deuterium')\n",
      "; most_similar female: ('OTHER', 'wondering', 'sowing', 'referring', 'Ranting', 'homophobic', 'COURSE', 'sore', 'courses', 'turf', 'preacher', 'Heartland', 'concerned', 'horsewoman', 'teaching', 'bigot', 'See', 'unaccredited', 'Episcopalian', 'offended')\n",
      "=====================================\n",
      "gender direction 128.\n",
      " most_similar male: ('Mois', 'implode', 'CPN', 'Kettler', 'infect', 'Sitbon', 'CDMS', 'psychotherapeutic', 'Brayton', 'infecting', 'Nicoloff', 'Kadmon', 'psychoanalyst', 'Spaniol', 'estranged', 'Dyadic', 'psychopathological', 'Barkin', 'psychoses', 'CPNP')\n",
      "; most_similar female: ('Transportation', 'Talent', 'guesswork', 'Restaurants', 'Tip', 'section', 'Ambassadors', 'jug', 'Motoring', 'recommendations', 'tips', 'ambassadors', 'Musicians', 'Generosity', 'Abundance', 'Shaykh', 'itinerary', 'rule', 'wineries', 'Tell')\n",
      "=====================================\n",
      "gender direction 129.\n",
      " most_similar male: ('Wiens', 'Denice', 'Haag', 'Kolar', 'Jonelle', 'Leiva', 'Milam', 'Kosak', 'Huberty', 'Wiese', 'Tanja', 'Whitten', 'Tollefson', 'Darci', 'Ibarra', 'Killeen', 'Jann', 'Haggard', 'Davidsonville', 'Jeannie')\n",
      "; most_similar female: ('irritation', 'rivalry', 'Papyrus', 'prophylaxis', 'rivalries', 'impact-', 'rule-', 'religiously', 'Rivalry', 'Obstetric', 'hangover', 'continuous', 'enmity', 'Acute', 'prophylactic', 'Prophylaxis', 'Traction', 'Mild', 'PPD', 'preference')\n",
      "=====================================\n",
      "gender direction 130.\n",
      " most_similar male: ('cultures', 'PCI', 'differences', 'polarity', 'Colombians', 'Ideals', 'genders', 'discs', 'polarities', 'hours', 'NDAs', 'Quidditch', 'ideals', 'loyalties', 'commonality', 'uncommon', 'Initech', 'peaceful', 'commonalities', 'faiths')\n",
      "; most_similar female: ('LINE', 'Introducing', 'Communications', 'ala', 'Multimedia', 'ported', 'ann', 'arsenal', 'Telephony', 'Telecommunication', 'tempt', 'esque', 'iPhone', 'telecommunications', 'iphone', 'N8', 'veto', 'cf', 'leake', 'hurled')\n",
      "=====================================\n",
      "gender direction 131.\n",
      " most_similar male: ('VMs', 'VMware', 'Eleazar', 'filesystem', 'filesystems', 'HDFS', 'aggregates', 'EMFStore', 'ClearCase', 'QRadar', 'virtualization', 'NVM', 'vSphere', 'Win32', 'hyperconverged', 'VM', 'HCS', 'debugger', 'Crashlytics', 'MidoNet')\n",
      "; most_similar female: ('Challenging', 'Rediscover', 'Mornings', 'dare', 'exhilarating', 'flair', 'urban', 'mornings', 'invigorating', 'Towards', 'Contemporary', 'rediscover', 'challenging', 'contemporary', 'Rabat', 'novelty', 'Urban', 'Marrakesh', 'Someday', 'Copenhagen')\n",
      "=====================================\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "gender direction 132.\n",
      " most_similar male: ('Realists', 'australis', 'potted', 'Raphaels', 'Columba', 'Aristote', 'Boathouse', 'Seagull', 'borealis', 'neo', 'Fireweed', 'Swans', 'emanated', 'Herter', 'Augustine', 'literati', 'Expressionists', 'Basking', 'asemic', 'lighthouse')\n",
      "; most_similar female: ('multiple', 'assignments', 'specialisation', 'specialization', 'assigments', 'selective', 'coverages', 'seniority', 'Selective', 'specialist', 'Specialist', 'AIT', 'arbitrations', 'specialisations', 'job', 'Tuba', 'reservist', 'Tez', 'specializations', 'IBN7')\n",
      "=====================================\n",
      "gender direction 133.\n",
      " most_similar male: ('conventionally', 'andhas', 'well-', 'classically', 'ruff', 'francophone', 'lotta', 'asa', 'Kortright', 'sizeable', 'clubhouse', 'provincially', 'large-', 'dyke', 'nonformal', 'a', 'Rideau', 'civically', 'TCC', 'Lisgar')\n",
      "; most_similar female: ('Devices', 'Sorcery', 'saddened', 'heading', 'Demise', 'Elder', 'Intent', 'Falling', 'disclosure', 'closeness', 'heartbeats', 'passing', 'idolizes', 'discontinuing', 'Controller', 'Obsession', 'Orson', 'Monopolies', 'Presidency', 'Hardship')\n",
      "=====================================\n",
      "gender direction 134.\n",
      " most_similar male: ('Microbiologists', 'Ragtime', 'Niccolai', 'Kubernetes', 'Docker', 'Meyerson', 'Lerone', 'Narcisi', 'Novello', 'Chemists', 'Astronautical', 'Hortonworks', 'Garvey', 'Organists', 'Rhysling', 'Buttigieg', 'Alumnae', 'musicology', 'Aretha', 'Ayyar')\n",
      "; most_similar female: ('SZ', 'uncertainties', 'trembling', 'depiction', 'TM', 'stress', 'bearing', 'reliefs', 'outflow', 'numbness', 'flashes', 'uncertainty', 'springs', 'flickering', 'TS', 'animations', 'Ts', 'sublimation', 'displays', 'KL')\n",
      "=====================================\n",
      "gender direction 135.\n",
      " most_similar male: ('interiority', 'HAL', 'ATM', 'Grief', 'Cleansing', 'Ergonomic', 'PCOS', 'Disappeared', 'CPAP', 'PML', 'Enquiry', 'Personhood', 'personhood', 'CART', 'OI', 'Ondaatje', 'Intimacies', 'CARD', 'Activa', 'Facilitation')\n",
      "; most_similar female: ('freshmen', 'freshman', 'sophomores', 'chancellor', 'college', 'dean', 'grades', 'semesters', 'headmaster', 'classmates', 'deans', 'WCU', 'school', 'colleges', 'students', 'tom', 'graduate', 'roommates', 'varsity', 'graduates')\n",
      "=====================================\n",
      "gender direction 136.\n",
      " most_similar male: ('Thema', 'Approximately', 'foreigners', 'wie', 'Match', 'Mariel', 'Saidy', '�', 'Ich', 'Pairs', 'Euros', 'Following', 'Tag', 'Lucerne', 'Maxi', 'Simple', 'hier', 'Loose', 'Eder', 'Freestyle')\n",
      "; most_similar female: ('technoscience', 'medicolegal', 'zooming', 'docket', 'subspeciality', 'litigative', '98105', 'paleoclimate', 'io9', 'conferencing', 'reinventing', 'dc', '15224', 'redeveloping', 'Horsham', 'sci', 'neglecting', 'otosclerosis', 'VR', 'rheumatology')\n",
      "=====================================\n",
      "gender direction 137.\n",
      " most_similar male: ('Lohmann', 'INRS', 'Heuer', 'CIRANO', 'Bernick', 'Hofmann', 'Hauger', 'Rohling', 'OTREC', 'Goller', 'Eawag', 'Lalive', 'FNRS', 'Foth', 'EUROGRAPHICS', 'Kallmeyer', 'flexography', 'Hasselberg', 'IHDP', 'Kling')\n",
      "; most_similar female: ('Far', 'dreaded', 'deemed', 'Sword', 'deems', 'weathers', 'sickle', 'Anvil', 'biblical', 'Warrior', 'Tyranny', 'tolerable', 'sacred', 'heath', 'truths', 'encountered', 'Cairn', 'mighty', 'Mornings', 'none')\n",
      "=====================================\n",
      "gender direction 138.\n",
      " most_similar male: ('syndromes', 'Gamewell', '-omics', 'farce', 'Taymor', 'braking', 'cytogenetics', 'MCC', 'Aster', 'assays', 'Sibley', 'bells', 'redundant', 'farcical', 'occlusive', 'STEMI', 'misusing', 'Heights', 'jeopardy', 'yield')\n",
      "; most_similar female: ('contributor', 'URL', 'tanned', 'models.com', 'Rusher', 'Juxtapoz', '1413', 'Rippy', '15203', 'Academia.edu', 'XLR8R', 'academia.edu', 'DLD', 'Influencers', 'SSRN', 'http', '4577', '8796', 'Jelen', 'surfs')\n",
      "=====================================\n",
      "gender direction 139.\n",
      " most_similar male: ('imagination', 'Hypochondria', 'Reveal', 'Depicting', 'Regard', 'Bartók', 'fascination', 'resides', 'Heracles', 'Revealing', 'ailment', 'nature', 'Geographical', 'Fairies', 'persistence', 'inner', 'Ravel', 'mythical', 'tranquility', 'fixated')\n",
      "; most_similar female: ('BMT', 'LSC', 'grading', 'IPS', 'grad', 'WMU', 'CU', 'seminary', 'MDX', 'IMRT', 'TSX', 'flotation', 'CGS', 'BCU', 'cadets', 'Vantage', 'grade', 'ELPS', 'UHV', 'shootouts')\n",
      "=====================================\n",
      "gender direction 140.\n",
      " most_similar male: ('Concertino', 'Affinities', 'chances', 'stats', 'Kindle', 'Defender', 'IMDb', 'Transcendent', 'Civilisation', 'Iolani', 'Artifact', 'Notables', '200th', 'historicity', 'IMDB', 'Masterpieces', 'Civilization', 'category', 'Literati', 'Quintet')\n",
      "; most_similar female: ('CALM', 'calm', 'applicator', 'sage', 'hazel', 'lavender', 'Nicosia', 'Tangier', 'beck', 'counselors', 'supervisors', 'spill', 'calmness', 'RCRA', 'hotline', 'biomonitoring', 'consultants', 'gray', 'regulators', 'Willner')\n",
      "=====================================\n",
      "gender direction 141.\n",
      " most_similar male: ('9400', '757', '9550', '8007', '1070', '1061', '1144', '7080', 'FEES', '1150', '4600', '1250', '9850', '1045', '980', '7000', '5250', '4200', '990', '8950')\n",
      "; most_similar female: ('intertwine', 'meld', 'intermingle', 'harmonize', 'melds', 'counterbalance', 'aligns', 'resonates', 'detective', 'interlinking', 'coexist', 'sleuth', 'personified', 'imbues', 'emanate', 'pinpoints', 'spotlights', 'juxtapose', 'align', 'bureaus')\n",
      "=====================================\n",
      "gender direction 142.\n",
      " most_similar male: ('imprint', 'chancery', 'digest', 'unsettled', 'salivary', 'whether', 'Biopsy', 'rereleased', 'bullet', 'superannuation', 'elastomer', 'gland', 'biopsy', 'release', 'roost', 'reissue', 'Doubt', 'seminal', 'environs', 'unsettling')\n",
      "; most_similar female: ('Teller', 'Peoples', 'Dornsife', 'Basset', 'Online', 'AJS', 'Bonanza', 'Wayuu', 'Poker', 'Morongo', 'Maidu', 'Fortney', 'Biker', 'Agoura', 'Eckley', 'BBW', 'Hippie', 'Iroquois', 'Javan', 'Lovers')\n",
      "=====================================\n",
      "gender direction 143.\n",
      " most_similar male: ('breakers', 'reached', 'Ameriprise', 'Protection', 'DSL', 'Ngoepe', 'Sash', 'Comission', 'Hallandale', 'consulation', 'Breakers', 'Pretoria', 'zoned', 'Counselor', 'Appt', 'Rang', 'Nedbank', 'advise', 'Barrier', 'Fikile')\n",
      "; most_similar female: ('egos', 'wrangles', '\\x97', 'backstage', 'musicians', 'creativities', 'wrangling', 'restless', 'biopharma', 'enriches', 'collaborating', 'hyperactive', 'juggles', 'overactive', 'dicks', 'collaboration', 'teamwork', 'insatiable', 'cocks', 'colossal')\n",
      "=====================================\n",
      "gender direction 144.\n",
      " most_similar male: ('church', 'pharmacy', 'coffee', 'latte', 'milk', 'Finnegan', 'Presti', 'Nicks', 'league', 'wifi', 'Collinsworth', 'Starbucks', 'concussions', 'candlelight', 'Conformation', 'meatballs', 'NFL', 'lattes', 'surfed', 'AFTRA')\n",
      "; most_similar female: ('Shahdara', '1541', '1546', 'Adugodi', '1483', 'Banashankari', '1185', 'MRTA', '7780', '14203', '1276', '14215', '1591', '1255', 'BAU', '1386', '10400', '1231', 'Domlur', '1617')\n",
      "=====================================\n",
      "gender direction 145.\n",
      " most_similar male: ('coils', 'Webs', 'coil', 'rods', 'webs', 'exhibit', 'rugs', 'loops', 'impulsively', 'nanotube', 'ducts', 'transgression', 'rug', 'exhibits', 'spirals', 'fibers', 'reticulum', 'cascades', 'exhibited', 'Zanjan')\n",
      "; most_similar female: ('MPS', 'iSeries', 'KODAK', 'CEREC', 'OpenEmbedded', 'Socialists', 'COA', 'Austerity', 'Convenience', 'relying', 'Relying', 'Bolstering', 'OPS', 'SAD', 'ELITE', 'Ricoh', 'GPC', 'Tories', 'OPC', 'electing')\n",
      "=====================================\n",
      "gender direction 146.\n",
      " most_similar male: ('Marathwada', 'Sangh', 'Orissa', 'Odisha', 'rainfall', 'clothed', 'Ujjain', 'fed', 'Bihar', 'Berhampur', 'Jharkhand', 'Utkal', 'Vicenza', 'Gorakhpur', 'Pontificial', 'Chattisgarh', 'Todi', 'Arunachal', 'Pontifical', 'Govt')\n",
      "; most_similar female: ('Litigator', 'Cruising', 'blepharoplasty', 'Kravis', 'Hellraiser', 'Painkiller', 'Zofia', 'multisport', 'Epix', 'Eye', 'ILW.COM', 'POZ', 'Dentons', 'LeClairRyan', 'genderqueer', 'Beholder', 'Jinx', 'Swix', 'Cozen', 'gel')\n",
      "=====================================\n",
      "gender direction 147.\n",
      " most_similar male: ('Shrek', 'Lancome', '50.000', 'prize', 'lemmings', 'Harrods', 'Huntsman', 'Fairytale', 'Frozen', 'giveaway', 'wishes', 'Disney', 'Pixar', 'Lancôme', 'refused', 'beens', 'price', 'Apple', '5000', 'Miliband')\n",
      "; most_similar female: ('Started', 'Conceived', 'Incorporating', 'breeze', 'Formed', 'aegis', '90.1', 'Propagation', '90.7', 'Recorded', 'Maddoux', 'Separating', 'Getting', 'Deriving', 'Adherence', 'Implementing', 'ASCA', 'breezy', '88.5', 'Baselios')\n",
      "=====================================\n",
      "gender direction 148.\n",
      " most_similar male: ('3121', 'biorefinery', 'B.S.E.E.', 'Hilfinger', '3090', '2801', 'Kanbar', 'coinventor', '2980', 'NREL', 'Bellcore', 'Fachhochschule', '3045', 'Outfest', '2901', 'BAVC', 'CITRIS', '2701', '2360', 'CEEE')\n",
      "; most_similar female: ('explanations', 'descriptions', 'helpful', 'explanation', 'heartbreaking', 'spoiling', 'understandable', 'exaggeration', 'complicating', 'unhelpful', 'colouring', 'drastically', 'tremendously', 'mood', 'Heartbreaking', 'greatly', 'Easter', 'distressing', 'upsetting', 'prayers')\n",
      "=====================================\n",
      "gender direction 149.\n",
      " most_similar male: ('30-year', '20-year', '10-year', 'MEDEX', 'trialled', 'Wanderings', 'Scandinavia', '30-day', 'commemorate', 'Arbour', '25-year', 'Maritimes', 'Saltspring', '15-year', 'Pangnirtung', 'NorthShore', '5-year', 'laminates', '40-year', 'laminate')\n",
      "; most_similar female: ('Hayek', 'optional', 'Barbi', 'Libertarian', 'Poonam', 'Laxmi', 'Fed', 'Paz', 'libertarian', 'extraneous', 'Wataru', 'Salaz', 'Krit', 'F.A.', 'Shinji', 'Ram', 'Kavya', 'Gakuen', 'Twombly', 'fork')\n",
      "=====================================\n"
     ]
    }
   ],
   "source": [
    "i = 0\n",
    "for w in Ws[:150]:\n",
    "    \n",
    "    #sims = word2vec.similar_by_vector(-w.squeeze(), topn = 60, restrict_vocab=None)\n",
    "    most_similar_f, _ =  list(zip(*word2vec_bios.similar_by_vector(w.squeeze(), topn = 20, restrict_vocab=None)))\n",
    "    most_similar_m, _ =  list(zip(*word2vec_bios.similar_by_vector(-w.squeeze(), topn = 20, restrict_vocab=None)))\n",
    "    print(\"gender direction {}.\\n most_similar male: {}\\n; most_similar female: {}\".format(i, most_similar_m, most_similar_f))\n",
    "    print(\"=====================================\")\n",
    "    i += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'g': 'f',\n",
       " 'p': 'nurse',\n",
       " 'text': 'Esther E Brown is a Nurse Practitioner Specialist in Salem, Oregon. She graduated with honors in 2002. Having more than 15 years of diverse experiences, especially in NURSE PRACTITIONER, Esther E Brown affiliates with Salem Hospital, and cooperates with other doctors and specialists in medical group Salem Clinic, Pc. Call Esther E Brown on phone number (503) 399-2424 for more information and advises or to book an appointment.',\n",
       " 'start': 67,\n",
       " 'hard_text': 'She graduated with honors in 2002. Having more than 15 years of diverse experiences, especially in NURSE PRACTITIONER, Esther E Brown affiliates with Salem Hospital, and cooperates with other doctors and specialists in medical group Salem Clinic, Pc. Call Esther E Brown on phone number (503) 399-2424 for more information and advises or to book an appointment.',\n",
       " 'text_without_gender': '_ graduated with honors in 2002. Having more than 15 years of diverse experiences, especially in NURSE PRACTITIONER, _ _ _ affiliates with Salem Hospital, and cooperates with other doctors and specialists in medical group Salem Clinic, Pc. Call _ _ _ on phone number (503) 399-2424 for more information and advises or to book an appointment.',\n",
       " 'hard_text_tokenized': 'She graduated with honors in 2002 . Having more than 15 years of diverse experiences , especially in NURSE PRACTITIONER , Esther E Brown affiliates with Salem Hospital , and cooperates with other doctors and specialists in medical group Salem Clinic , Pc . Call Esther E Brown on phone number ( 503 ) 399 - 2424 for more information and advises or to book an appointment .'}"
      ]
     },
     "execution_count": 243,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nurse_train = [r for r in train if r[\"p\"] == \"nurse\"]\n",
    "nurse_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [],
   "source": [
    "profs_train_str = np.array([i2p[y] for y in Y_train[:]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((255682, 300), (255682,))"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape, profs_train_str.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 332,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accountant 1.4361589787313929 % 0.6947167755991286 0.6636710239651417\n",
      "architect 2.570380394396164 % 0.615642118076689 0.5348447961046866\n",
      "attorney 8.288420772678561 % 0.8696677991694979 0.8259720649301623\n",
      "chiropractor 0.6601950860834944 % 0.47334123222748814 0.44016587677725116\n",
      "comedian 0.7110394943719152 % 0.7442244224422442 0.7046204620462047\n",
      "composer 1.422861210409806 % 0.7660802638812534 0.7218251786695987\n",
      "dentist 3.6799618275827006 % 0.8686364119460092 0.8567329152938675\n",
      "dietitian 1.01063039244061 % 0.7430340557275542 0.6385448916408669\n",
      "dj 0.37663973216730157 % 0.6542056074766355 0.6105919003115264\n",
      "filmmaker 1.7815098442596662 % 0.7369923161361142 0.706476399560922\n",
      "interior_designer 0.3715552913384595 % 0.5452631578947369 0.43894736842105264\n",
      "journalist 5.070751949687502 % 0.7170844581565754 0.6748168145005785\n",
      "model 1.9066653108157792 % 0.7577435897435898 0.683076923076923\n",
      "nurse 4.822396570740216 % 0.7705596107055961 0.7046228710462287\n",
      "painter 1.9657230465969444 % 0.7600477516912058 0.7278153601273378\n",
      "paralegal 0.44899523627005417 % 0.27439024390243905 0.2029616724738676\n",
      "pastor 0.6429862094320288 % 0.5395377128953771 0.4811435523114355\n",
      "personal_trainer 0.3633419638457146 % 0.6027987082884823 0.573735199138859\n",
      "photographer 6.183071158704954 % 0.8666582326522867 0.8209247896767664\n",
      "physician 9.8035841396735 % 0.8291310939120722 0.7955796696720657\n",
      "poet 1.7819009550926541 % 0.7047848990342406 0.6202809482001755\n",
      "professor 30.025969759310396 % 0.8593609566112204 0.8587487462713785\n",
      "psychologist 4.6503078042255614 % 0.6747687132043734 0.6119428090832633\n",
      "rapper 0.35708419051790896 % 0.7053669222343921 0.6232201533406353\n",
      "software_engineer 1.7549143076164926 % 0.6469801649208825 0.5910407844885224\n",
      "surgeon 3.373722045353212 % 0.6291444470206353 0.5503130071875725\n",
      "teacher 4.117223738863119 % 0.5994110382825116 0.47829391089579176\n",
      "yoga_teacher 0.4220085887938924 % 0.6200185356811863 0.5987025023169601\n"
     ]
    }
   ],
   "source": [
    "for p in p2i.keys():\n",
    "    x_train_nurse = X_train[profs_train_str == p]\n",
    "    y_train_nurse = Y_train[profs_train_str == p]\n",
    "    print(p,len(x_train_nurse)/len(X_train)*100,\"%\", clf_original.score(x_train_nurse, y_train_nurse), clf.score(P.dot(x_train_nurse.T).T, y_train_nurse))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((4556, 300), (4556,))"
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train_nurse.shape, y_train_nurse.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6248902546093064"
      ]
     },
     "execution_count": 267,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 268,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7047848990342406"
      ]
     },
     "execution_count": 268,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 306,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'professor': 0.45118956904580476,\n",
       " 'chiropractor': 0.26558891454965355,\n",
       " 'psychologist': 0.6223011751844766,\n",
       " 'architect': 0.23712053792148718,\n",
       " 'physician': 0.507688318423441,\n",
       " 'nurse': 0.9085446207369142,\n",
       " 'dentist': 0.35589474411216243,\n",
       " 'surgeon': 0.14857228961048746,\n",
       " 'rapper': 0.09665955934612651,\n",
       " 'model': 0.8283124500133298,\n",
       " 'photographer': 0.35721920736720936,\n",
       " 'composer': 0.16392857142857142,\n",
       " 'comedian': 0.21150410861021793,\n",
       " 'filmmaker': 0.3295762590954487,\n",
       " 'paralegal': 0.8483305036785512,\n",
       " 'journalist': 0.49488721804511276,\n",
       " 'personal_trainer': 0.45670391061452514,\n",
       " 'teacher': 0.603111879476414,\n",
       " 'painter': 0.4579886246122027,\n",
       " 'attorney': 0.38316925813475633,\n",
       " 'accountant': 0.36818825194621374,\n",
       " 'software_engineer': 0.1576889661164205,\n",
       " 'poet': 0.49080017115960634,\n",
       " 'dj': 0.1420875420875421,\n",
       " 'pastor': 0.24052132701421802,\n",
       " 'yoga_teacher': 0.8454600120264583,\n",
       " 'dietitian': 0.9273504273504274,\n",
       " 'interior_designer': 0.8086124401913876}"
      ]
     },
     "execution_count": 306,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "{'professor': 0.45118956904580476, 'chiropractor': 0.26558891454965355, 'psychologist': 0.6223011751844766, 'architect': 0.23712053792148718, 'physician': 0.507688318423441, 'nurse': 0.9085446207369142, 'dentist': 0.35589474411216243, 'surgeon': 0.14857228961048746, 'rapper': 0.09665955934612651, 'model': 0.8283124500133298, 'photographer': 0.35721920736720936, 'composer': 0.16392857142857142, 'comedian': 0.21150410861021793, 'filmmaker': 0.3295762590954487, 'paralegal': 0.8483305036785512, 'journalist': 0.49488721804511276, 'personal_trainer': 0.45670391061452514, 'teacher': 0.603111879476414, 'painter': 0.4579886246122027, 'attorney': 0.38316925813475633, 'accountant': 0.36818825194621374, 'software_engineer': 0.1576889661164205, 'poet': 0.49080017115960634, 'dj': 0.1420875420875421, 'pastor': 0.24052132701421802, 'yoga_teacher': 0.8454600120264583, 'dietitian': 0.9273504273504274, 'interior_designer': 0.8086124401913876}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}