Assignment 04 - Gendered Directions in Embeddings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "58LjICoY5YsV"
   },
   "source": [
    "# Assignment 4: Gendered Directions in Embeddings\n",
    "\n",
    "*version 2023.1, details/bugs might be fixed within the first week of publication, you will be notified if this happens*\n",
    "\n",
    "Most of those interested in computational linguistics have probably read about model bias. What is often less clear is what bias means in terms of data and modelling. In this assignment, you will reproduce results from a famous paper on the technical side of this problem. The discussion on if bias has real-world implications and the details of the meaning of the word \"bias\" will be picked up again later.\n",
    "\n",
    "The following will be based on the paper *Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings*. In this paper, the authors start by finding word analogies and investigate semantic closeness. Your task is to reproduce some of their results. You are given code for the analysis, but the reading and representing of the data is up to you to implement.\n",
    "\n",
    "\n",
    "## Submission\n",
    " \n",
    "Please submit your code as a notebook through studium. You should include the following:\n",
    " \n",
    "1. A working implementation of your embedding class, with the given code cells untouched or with additional experiments, reproducing your principal results when run. Please ensure that outputs are stored in the notebook (preferably by rerunning your notebook as the last thing you do before submitting).\n",
    "2. Comment *your* code properly, especially for longer or opaque functionality. Please try to write self documenting code (e.g. by choosing descriptive variable names, refactoring to isolate functionality, and by minimizing code duplication).\n",
    "4. Comments on what you thought was hard in the assignment, what you think was educational, what took most time, and which parts might be unnecessarily tricky.\n",
    "5. As the submission is anonymous, **all personal information must be removed**.\n",
    "6. Apart from sections and titles, please remove all unnecessary text and code from the notebook you hand in. Keep only that which strengthens the case for you having fulfilled the listed requirements.\n",
    "\n",
    "\n",
    "## Requirements for grade G\n",
    "\n",
    "To achieve a pass (G) on this assignment, you must finish the `Embedding` class without serious errors.\n",
    "\n",
    "\n",
    "### Additional requirements for grade VG\n",
    "\n",
    "For a pass with distinction grade (VG), you should make a serious attempt at extending the experiments to an additional language. The experiments should not be as thorough, and can be limited to finding a smaller set of professions and directional pairs. As the course is given in English, you can not (for legal reasons) be required to know any other languages. However, you can use the dictionaries from [MUSE](https://github.com/facebookresearch/MUSE#ground-truth-bilingual-dictionaries) (or other source) to find rough translations.\n",
    "\n",
    "The language aligned embeddings, as described in Joulin et al. (2018), can be found [on the fasttext website](https://fasttext.cc/docs/en/aligned-vectors.html). This means that the word vectors can be used for translations using nearest neighbours. Also, the vectors can be run through PCA so that the closeness of words can be visualized.\n",
    "\n",
    "In addition to your notebook, please submit 1-3 slides with a short manuscript (< 500 words) presenting your results.\n",
    "\n",
    "\n",
    "### Plagiarism\n",
    " \n",
    "In code assignments, plagiarism is a tricky concept. A clean cut way would be to demand that you write all the code yourself, from memory, with only the assigned literature as help. This is not how code is developed professionally. It is common to copy and share. However, since this is a learning exercise, you must implement everything on your own, but please look at the course repo, Stack Overflow etc. Moreover, discuss with course mates and TAs to find inspiration and solutions. Code that is *obviously* copied (with minor modifications) will be considered as plagiarized. As a part of the examination, you might be asked to explain any particular part of the functionality in your implementation.\n",
    "\n",
    "\n",
    "### References\n",
    "\n",
    "Bolukbasi et al., (2016) \"[*Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings*](https://arxiv.org/pdf/1607.06520.pdf),\" arXiv\n",
    "\n",
    "Joulin et al. (2018) \"[*Loss in Translation: Learning Bilingual Word Mapping with a Retrieval Criterion*](https://arxiv.org/abs/1804.07745),\" arXiv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IV-wK45Rymev"
   },
   "source": [
    "Let's download the embedding data file. You can find more languages and manuals [here](https://fasttext.cc/docs/en/crawl-vectors.html)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Fo2kQHnR0mU8",
    "outputId": "914b5b8f-56e0-43dc-f82d-55ba9a251251"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "--2022-05-04 07:48:10--  https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.en.300.vec.gz\n",
      "Resolving dl.fbaipublicfiles.com (dl.fbaipublicfiles.com)... 104.22.74.142, 172.67.9.4, 104.22.75.142, ...\n",
      "Connecting to dl.fbaipublicfiles.com (dl.fbaipublicfiles.com)|104.22.74.142|:443... connected.\n",
      "HTTP request sent, awaiting response... 200 OK\n",
      "Length: 1325960915 (1.2G) [binary/octet-stream]\n",
      "Saving to: ‘cc.en.300.vec.gz’\n",
      "\n",
      "cc.en.300.vec.gz    100%[===================>]   1.23G  39.8MB/s    in 33s     \n",
      "\n",
      "2022-05-04 07:48:43 (38.7 MB/s) - ‘cc.en.300.vec.gz’ saved [1325960915/1325960915]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "!wget -N https://dl.fbaipublicfiles.com/fasttext/vectors-crawl/cc.en.300.vec.gz"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "bCziUxV_yq6y"
   },
   "source": [
    "The format is fairly simple. It has one token per line, which are space separated lists. For fasttext, the tokens are usually ordered by frequency in the training data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f8LTAWcm0pMd",
    "outputId": "c13276ae-c0cb-4477-9b1b-8cb3050b9267"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2000000 300\n",
      ", 0.1250 -0.1079 0.0245 -0.2529 0.1057 -0.0184 0.1177 -0.0701 -0.0401 -0.0080 0.0772 -0.0226 0.0893 -0.0487 -0.0897 -0.0835 0.0200 0.0273 -0.0194 0.0964 0.0875 0.0098 0.0453 0.0155 0.1462 0.0225 0.0448 0.0137 0.0570 0.1764 -0.1072 -0.0826 0.0173 0.1090 0.0207 -0.1271 0.2445 0.0375 -0.0209 -0.0445 0.0540 0.1282 0.0437 0.0588 0.0984 0.0539 0.0004 0.1290 0.0242 -0.0120 -0.0480 0.0346 -0.0664 -0.0330 -0.0625 -0.0708 -0.0579 0.1738 0.4448 0.0370 -0.1001 -0.0032 0.0359 -0.0685 -0.0361 0.0070 0.1316 -0.0945 -0.0610 0.0178 -0.0763 -0.0192 0.0033 0.0056 0.1878 -0.0754 -0.0095 0.0446 -0.0588 0.0244 -0.0251 -0.0493 0.0308 -0.0359 -0.1884 -0.0988 0.1887 0.0459 -0.0816 -0.1524 -0.0375 -0.0692 0.0427 -0.0471 -0.0086 -0.2190 -0.0064 0.0877 -0.0074 -0.1400 -0.0156 0.0161 0.1040 -0.1445 -0.0719 -0.0144 -0.0293 -0.0126 0.0619 -0.0373 -0.1471 -0.2552 -0.0685 0.2892 -0.0275 0.0436 0.0311 0.0249 0.0142 0.0403 0.1729 0.0023 -0.0255 -0.0212 0.0701 -0.0727 0.0279 0.1151 -0.0394 -0.0962 -0.0598 -0.0459 -0.0326 -0.2317 0.0945 0.0110 -0.2511 0.1087 -0.0699 0.0359 0.0208 -0.0536 0.0478 0.0178 0.0095 0.0354 -0.7726 -0.0790 0.0472 0.0584 -0.1013 0.0448 -0.1202 0.0376 0.0510 -0.0616 0.4321 0.0179 0.0263 0.0271 0.0473 -0.0951 -0.2261 0.0261 0.0262 -0.0235 -0.0369 -0.1655 -0.0697 0.0122 -0.0303 0.0427 0.0787 -0.0360 0.0206 -0.0068 0.1257 0.0447 -0.0776 -0.1122 -0.0291 0.4654 0.1010 0.4440 0.0095 0.1312 0.0766 0.0873 -0.0878 -0.0296 0.0046 0.0416 -0.0134 0.0571 -0.0109 -0.0655 0.0082 -0.0563 -0.0830 -0.0550 -0.0688 0.0091 -0.0677 -0.1001 0.0200 -0.0979 0.1134 -0.0188 0.0136 0.0782 0.0207 0.0133 -0.0492 -0.0139 0.0123 0.0360 0.1249 0.0503 0.0015 0.1246 -0.0897 -0.0121 -0.0182 0.2245 -0.0313 -0.1596 0.0073 -0.0772 -0.0830 -0.0716 0.0112 0.0218 0.1245 -0.0361 0.0312 0.0652 0.0560 0.0670 0.0709 -0.0248 0.0449 -0.1296 0.1408 -0.0359 1.1585 0.0027 0.0185 0.0549 0.1367 0.2742 -0.0472 -0.0414 -0.0548 -0.0911 -0.0015 -0.1613 0.0179 -0.0040 0.0610 0.0559 0.1138 0.2978 -0.1511 -0.0079 -0.0838 0.0296 -0.1041 0.1627 -0.0670 0.0504 -0.0420 -0.0020 0.1840 0.0596 0.0448 0.0989 -0.2157 -0.0117 0.2142 -0.1672 -0.0444 0.2045 -0.4620 -0.0482 0.0688 -0.0304 0.0478 0.1583 0.0920 0.0949 0.0650 -0.0398 -0.1376 -0.0436 0.0578 0.0188 0.0148 0.2305 -0.0696 -0.0215\n",
      "the -0.0517 0.0740 -0.0131 0.0447 -0.0343 0.0212 0.0069 -0.0163 -0.0181 -0.0020 -0.1021 0.0059 0.0257 -0.0026 -0.0586 -0.0378 0.0163 0.0146 -0.0088 -0.0176 -0.0085 -0.0078 -0.0183 0.0088 0.0013 -0.0938 0.0139 0.0149 -0.0394 -0.0294 0.0094 -0.0252 -0.0104 -0.2213 -0.0229 -0.0089 -0.0322 0.0822 0.0021 0.0282 0.0072 -0.0091 -0.0352 -0.0178 -0.0706 0.0630 -0.0092 -0.0223 -0.0056 0.0515 -0.0307 0.0436 -0.0110 -0.0555 0.0089 -0.0673 0.0105 0.0574 0.0099 -0.0283 0.0470 0.0053 0.0030 0.0007 0.0443 0.0069 -0.0334 0.0091 -0.0076 0.0066 0.0917 0.0311 0.0543 0.0282 -0.0200 -0.0334 0.0053 0.0364 0.2249 0.0928 -0.0123 0.0086 -0.0599 0.0676 0.0402 0.0012 0.0464 -0.0437 0.0059 0.0917 -0.0412 -0.0151 -0.0231 0.0095 0.0588 0.0279 0.0647 -0.0568 -0.0130 0.0474 0.0354 -0.0121 -0.0077 -0.1306 0.0134 -0.0506 0.0111 0.0119 -0.0221 0.0394 0.0221 0.0245 0.0039 0.1146 0.0228 -0.0468 -0.0459 -0.0189 0.0076 -0.0302 -0.0348 -0.0289 -0.0399 0.0245 -0.0102 0.0578 -0.0388 -0.0117 -0.0304 0.2466 -0.0111 0.0356 0.0046 0.2094 -0.1022 0.0336 0.0688 -0.0708 0.0269 -0.0423 0.0077 -0.0267 0.0072 0.0035 0.0351 -0.0063 -0.4462 0.0105 -0.0120 -0.0449 -0.1696 0.0504 0.0930 -0.0044 -0.0041 0.0322 0.2026 0.0613 -0.0295 0.0228 -0.0190 0.0173 0.1480 -0.0175 -0.0125 0.0687 0.0333 -0.0303 0.0428 0.0051 0.0228 0.0104 0.0731 0.0079 -0.0051 0.0543 -0.0325 0.0512 0.0288 -0.0586 -0.0000 0.0493 0.0166 -0.0143 0.0359 0.0543 -0.0005 -0.0589 0.0162 -0.0222 -0.0199 0.0235 -0.0678 0.0179 0.0033 0.0114 0.0473 -0.0443 0.0323 0.0195 -0.0647 0.3388 0.0699 -0.0215 -0.0244 -0.0034 -0.0034 -0.0622 0.0123 0.0375 -0.0197 0.0241 -0.0877 0.0201 -0.0061 -0.0256 -0.0191 -0.0264 0.0190 -0.0422 0.0251 0.0825 -0.0096 0.1288 0.0621 0.0538 0.0189 0.0422 0.1803 -0.0010 -0.0328 -0.0559 -0.0157 0.0490 0.0352 -0.0417 0.0159 -0.0766 -0.0657 0.0497 0.0102 0.1471 -0.0711 -0.1469 0.4737 -0.0169 -0.0051 0.0159 0.0550 -0.0634 -0.0210 0.0122 0.0269 0.0060 0.0664 0.0106 -0.0705 -0.0207 -0.0784 -0.0291 -0.0283 -0.1568 -0.0393 0.0050 0.0204 -0.0026 0.0436 0.0279 -0.0393 0.0367 -0.0042 -0.0156 -0.0733 -0.1637 0.0652 -0.0062 -0.0650 -0.1984 -0.0411 -0.1534 0.0020 0.0133 -0.2364 -0.0528 -0.0042 -0.0447 0.0112 -0.0332 -0.0550 0.0013 0.0169 -0.0439 -0.0578 0.0223 -0.0777 -0.0432 -0.0251 0.2370 0.0004 -0.0042\n",
      ". 0.0342 -0.0801 0.1162 -0.3968 -0.0147 -0.0533 0.0606 -0.1052 0.0005 -0.0360 0.0257 0.0177 0.0285 0.0037 -0.0419 0.2374 0.0073 -0.0303 -0.0578 -0.0616 0.0065 0.0055 0.0620 -0.0036 -0.1077 -0.1046 0.0654 -0.0007 0.0235 0.0449 0.0013 -0.0496 -0.0030 0.1332 0.0314 -0.0152 0.0773 -0.3238 -0.0081 0.0108 -0.0478 0.1037 -0.0894 0.1494 0.5013 -0.1842 -0.0259 0.0780 -0.0296 0.0597 0.0438 -0.0477 -0.0359 -0.0604 0.0376 -0.0455 -0.1325 -0.0595 -0.0999 -0.0824 -0.0963 -0.0855 -0.0244 0.5080 -0.0271 -0.0886 -0.0160 -0.0503 -0.0295 -0.0177 0.3846 0.1046 0.1692 -0.0120 0.0465 -0.0801 0.0126 0.0522 0.1096 0.2034 0.0128 0.0064 -0.0638 0.0208 0.1247 0.0043 0.0863 0.1134 0.0314 0.0870 0.0581 0.0139 0.1129 0.0039 -0.1928 -0.1918 -0.2264 0.0319 -0.0108 -0.0572 0.0661 0.0072 -0.0243 0.0817 0.0124 -0.0023 -0.0671 0.0637 0.0683 -0.0160 0.0638 0.0179 0.0930 -0.1994 -0.0062 -0.0831 -0.0390 0.0223 0.0752 -0.0473 -0.0331 0.0411 -0.0264 0.0050 0.0732 0.0127 -0.0120 -0.1518 -0.0482 0.0112 -0.0677 -0.0567 0.0096 -0.4102 -0.1054 0.0108 -0.4377 0.0009 -0.0180 0.0575 0.0029 -0.0109 0.0439 -0.0377 0.0886 -0.0430 -0.9954 -0.0102 0.0165 -0.0913 0.1988 0.1050 -0.0605 0.0899 0.0143 0.0235 0.5098 0.1648 -0.0803 0.0982 -0.0146 -0.0691 -0.2319 0.0812 0.0033 0.0185 0.0190 0.0252 0.0410 -0.0093 -0.0462 -0.0149 0.1279 0.0029 -0.0192 -0.0110 -0.0070 -0.0739 0.0538 -0.0325 0.0112 -0.3188 0.0089 0.7444 0.0718 0.1976 0.0864 -0.1032 -0.0250 0.0158 -0.0967 -0.0028 -0.0473 0.0746 0.0153 0.0036 -0.0198 -0.0598 0.0702 -0.0945 -0.0444 -0.5588 -0.0280 -0.0655 0.1006 0.0174 -0.2034 -0.0155 0.0551 -0.0476 0.0795 0.0621 -0.0732 -0.0093 -0.0274 -0.0583 0.0597 -0.0671 0.0376 -0.1652 0.0615 0.0988 0.0516 -0.1770 -0.0328 0.0048 0.0501 -0.0052 -0.1288 -0.0688 -0.1061 -0.0701 -0.0147 -0.0525 0.0487 0.1475 0.0576 -0.0084 -0.0227 -0.0459 -0.0869 -0.1352 0.0435 -0.1126 1.1697 -0.0520 0.0239 0.0460 -0.0097 0.2175 -0.0613 -0.0294 -0.0274 -0.1675 -0.0110 0.0282 -0.1193 -0.0581 0.1425 -0.0218 0.0801 -0.1664 -0.0608 0.0108 0.2657 0.0531 0.0112 -0.1190 -0.1190 -0.1326 -0.1026 0.0041 0.4929 0.3728 0.1085 0.1950 -0.3836 -0.1544 0.2013 0.0457 0.0052 0.0232 0.2838 -0.1311 0.0553 -0.0563 0.0974 -0.0534 -0.0549 0.0354 0.0415 -0.0418 -0.1440 -0.0659 -0.0491 -0.0467 -0.0042 0.5423 -0.0624 0.0900\n",
      "and 0.0082 -0.0899 0.0265 -0.0086 -0.0609 0.0068 0.0652 0.0106 -0.0475 -0.0076 -0.0023 0.0009 -0.0067 -0.0226 -0.0066 -0.0725 0.0203 0.0215 -0.0507 0.0392 0.0533 -0.0045 -0.0019 0.0557 0.0146 -0.0571 0.0005 -0.0071 0.0149 -0.0528 0.0285 -0.0496 0.0222 0.0586 0.0213 0.0079 0.0187 0.0919 -0.0278 0.0239 0.0426 -0.0099 -0.0223 0.0856 -0.0448 -0.0142 -0.0001 -0.1002 -0.0124 0.0595 0.0236 0.0156 0.0194 0.0086 0.0152 0.0416 0.0198 -0.0075 -0.0308 -0.0211 0.0274 -0.0417 -0.0356 -0.0217 0.0138 0.0272 0.0291 0.0162 0.0223 0.0312 -0.0075 0.0103 0.0294 0.0126 0.0729 -0.0124 -0.0444 0.0331 -0.0976 -0.0259 0.0053 0.0176 -0.0195 0.0300 -0.0319 0.0281 0.0548 -0.0153 -0.0437 -0.0541 0.0112 0.0216 -0.0360 -0.0145 -0.0036 0.0196 0.0122 -0.0571 -0.0270 0.0638 0.0429 -0.0607 -0.0319 0.0032 -0.0125 -0.0713 -0.0304 -0.0122 0.0212 -0.0269 0.0082 -0.1155 0.0008 0.0896 -0.0165 -0.0128 -0.0411 -0.0135 0.0154 -0.0281 0.0618 -0.0037 -0.0811 0.0207 0.0202 0.0587 -0.0165 0.1276 -0.0317 -0.0039 -0.0439 0.0147 -0.0587 -0.0186 0.0126 -0.0054 -0.0426 0.0606 -0.0428 -0.0296 -0.0103 -0.0180 0.0103 0.0300 0.0206 0.0247 -0.3690 -0.0039 0.0272 -0.0251 -0.0401 0.0254 -0.0440 -0.0038 -0.0002 0.0138 0.0608 0.0326 -0.0870 -0.0047 -0.0584 -0.0243 -0.0934 -0.0029 -0.0029 0.0198 0.0149 -0.0409 0.0181 -0.0210 -0.0221 0.0168 0.0311 0.0334 -0.0504 0.0164 -0.0313 0.1001 0.0283 -0.0844 0.0168 0.0222 -0.0625 -0.0627 0.0308 0.0266 -0.0627 -0.0390 0.0234 -0.0504 -0.0445 0.0474 -0.0452 0.0411 0.0174 -0.0617 0.0638 -0.0659 -0.0149 0.0685 -0.0729 -0.0177 0.0046 -0.0138 -0.0161 -0.0184 -0.0989 0.0070 0.0543 0.1277 0.0356 -0.0027 0.0858 0.0111 -0.0020 0.0273 0.0501 -0.0387 0.0323 0.0417 0.0524 0.0277 -0.0144 0.0433 0.0427 -0.0107 0.0395 -0.0518 -0.0333 -0.0117 -0.0287 -0.0242 0.0370 0.0450 0.0316 -0.0527 0.0119 0.0500 0.0142 0.0323 0.0257 -0.0020 0.0564 -0.2396 0.6197 0.0086 0.0366 -0.0422 0.0743 0.0188 -0.0432 -0.0123 -0.0217 -0.0457 0.0127 -0.0475 -0.1059 0.0110 -0.0014 -0.0093 0.0387 0.0941 -0.0442 0.0169 -0.0713 0.0749 0.0029 0.0932 -0.1163 0.0106 -0.0150 0.0558 -0.0585 -0.0191 0.0228 0.0304 0.0189 -0.0319 -0.0564 0.0581 0.0060 0.0722 -0.1537 0.0284 0.0112 0.0418 0.0354 -0.0962 -0.0008 0.0197 0.0607 -0.0355 -0.0238 -0.0278 -0.0260 -0.0381 -0.0211 -0.0116 -0.0411 0.0363\n",
      "to 0.0047 0.0281 -0.0296 -0.0108 -0.0620 -0.0532 -0.0980 0.0970 -0.0864 -0.0425 0.0272 -0.0376 -0.0428 0.0374 -0.0040 -0.0354 0.0064 0.0182 -0.0584 0.6645 0.0644 0.0707 -0.0402 -0.1074 -0.2283 -0.0792 0.0025 -0.0092 -0.2078 0.1934 -0.0301 -0.1259 0.0444 0.0713 0.0433 -0.0415 -0.6236 0.3654 -0.0123 -0.0423 0.0435 0.0502 -0.1302 0.0649 0.2670 0.1210 0.0346 0.1850 0.0521 0.0283 0.0776 -0.1018 -0.0567 -0.0728 -0.0254 -0.0705 -0.1293 0.1515 -0.1173 -0.1377 -0.0204 -0.0217 0.0070 -0.5928 -0.0595 -0.0666 0.0318 -0.0878 -0.0167 -0.0460 0.0700 0.0213 0.0985 -0.0205 0.0085 -0.0939 0.0774 0.1188 -0.1067 -0.2018 0.0288 -0.0399 -0.0219 -0.0134 0.4109 -0.0234 0.1017 0.2881 -0.0449 -0.3368 -0.0315 -0.0809 0.0940 -0.0518 -0.0475 -0.5167 -0.1999 0.0226 -0.0811 -0.0437 -0.0081 0.1105 0.0517 -0.1964 -0.0126 0.0381 0.1052 0.0019 0.0569 0.0195 -0.0397 0.0607 0.0438 -0.0737 -0.0172 -0.0208 0.0498 0.1633 -0.0268 -0.0278 0.0018 -0.0246 -0.2153 -0.0124 0.0213 -0.0624 -0.0041 -0.2018 0.0505 0.0619 -0.0408 -0.0456 0.1332 0.3132 -0.0459 0.0016 -0.3367 -0.0274 0.0390 0.0620 -0.0640 -0.0869 -0.0004 0.0439 0.4356 0.0317 -0.9923 -0.0041 -0.0609 0.0086 -0.0195 0.0984 0.3098 0.0713 -0.0617 0.0414 0.2753 -0.0233 -0.0001 0.1100 0.0523 0.0180 0.3354 0.0563 0.0776 -0.0596 -0.0898 0.0156 -0.1101 0.0454 -0.0145 -0.0607 0.0578 -0.0799 0.0001 -0.0533 0.0791 0.0198 0.0549 0.1604 -0.0379 0.3503 -0.1227 -0.0384 0.0086 0.1561 -0.2621 -0.1502 0.0508 0.1018 -0.0441 0.0975 0.0229 -0.0156 0.0132 0.0223 0.0090 -0.0066 0.0932 -0.1725 0.0339 0.6197 0.0383 -0.1625 0.1081 -0.0064 -0.2402 0.0024 -0.0789 -0.1186 -0.0651 -0.0011 0.1207 0.0766 -0.0091 -0.0295 -0.1350 0.0065 0.0444 -0.0736 -0.0202 0.0356 -0.0056 -0.0090 0.1048 -0.1593 0.0499 -0.0075 -0.2655 -0.1139 0.0383 0.0871 -0.5801 -0.1249 0.0065 -0.0604 0.0299 0.0610 -0.0484 0.0209 -0.0498 0.3544 0.1185 -0.1023 1.1741 0.0099 0.0551 0.0513 0.0347 -0.5451 0.1302 -0.2305 0.0088 -0.1085 -0.0125 0.1309 -0.1173 -0.0451 0.1080 0.0820 -0.0348 -0.0139 -0.0015 -0.0310 0.2747 0.0404 0.0027 0.0841 -0.0140 -0.0191 -0.1035 -0.0725 -0.1953 0.2172 -0.0067 0.2449 0.1352 0.2282 0.2828 0.6275 -0.1176 -0.1204 -0.3137 0.0473 0.0263 -0.0255 -0.0015 -0.0473 -0.0656 0.0381 0.3740 0.0261 -0.2208 -0.0312 0.0375 -0.0164 -0.0011 0.3256 -0.0509 0.0072\n",
      "of -0.0001 -0.1877 -0.0711 -0.4632 0.0002 0.0115 -0.0588 0.0574 -0.0275 -0.0036 -0.0098 -0.0414 0.0935 0.0541 -0.1960 -0.2909 -0.0632 0.0466 -0.0160 0.0308 -0.0679 0.0927 -0.0136 -0.0143 -0.2465 -0.1221 -0.0306 -0.0195 -0.0239 0.1938 0.0857 -0.0231 0.1146 -0.1669 -0.0564 -0.0697 0.4156 0.0970 -0.0382 -0.0540 -0.0327 0.1035 0.1476 -0.0286 -0.0977 -0.5861 -0.0348 -0.1596 0.0368 0.0350 -0.1274 -0.0766 -0.2187 -0.1020 0.0292 -0.0611 -0.1248 -0.1226 -0.1561 0.0026 -0.0567 -0.0128 0.0199 -0.2058 0.0172 -0.0017 0.0816 0.0123 0.0371 0.0352 0.6317 0.1009 0.0842 0.0339 0.0174 -0.1590 -0.0236 0.0065 0.4384 -0.1326 0.0838 -0.0216 -0.0125 -0.0029 0.1592 0.0248 0.3087 -0.7341 -0.1604 0.1809 -0.0747 0.0004 0.0340 -0.0629 0.2521 -0.1315 -0.1092 -0.0868 -0.0197 0.0131 0.1246 0.0272 0.0188 -0.2150 0.0209 0.0378 -0.0092 0.0391 0.0826 -0.0169 -0.0893 0.1233 -0.0342 0.5112 0.0278 0.0369 0.0182 0.0390 0.0222 -0.0440 0.0504 -0.1305 -0.0322 -0.0399 0.0364 -0.0060 -0.0363 0.1907 -0.0493 0.4924 0.0278 -0.0126 0.1377 0.1620 0.0777 0.0488 0.3074 -0.0699 0.0511 -0.0823 -0.0869 -0.0885 -0.0049 0.0623 0.3402 -0.1010 -0.7793 0.0548 -0.0364 -0.0255 -0.2537 0.1285 0.2463 0.0976 0.0179 0.0757 0.3209 -0.2619 0.0336 0.0767 0.0296 -0.0676 0.2080 0.0227 0.0692 0.0442 0.0338 0.1349 -0.0172 -0.1376 -0.0530 0.0074 0.1089 0.0009 -0.0356 0.0108 0.0209 0.0932 0.0039 0.0393 0.0138 0.2375 -0.1512 0.0935 -0.0149 0.2441 0.1910 -0.1638 0.0034 0.0412 0.0553 0.0555 -0.0829 0.0516 0.0580 -0.0115 -0.0250 -0.0066 0.0592 0.0241 -0.0390 0.6112 0.0601 -0.0083 0.0457 0.0616 0.0048 0.0812 0.0495 0.1826 -0.1636 0.0685 -0.1536 -0.0510 -0.0751 -0.0082 0.1437 -0.0251 0.0897 0.2579 -0.0355 0.0881 -0.0166 -0.2616 -0.0774 0.2639 0.0426 0.0113 0.0280 0.0270 0.0230 0.0102 -0.1128 -0.0435 0.0866 -0.0221 0.0277 -0.2634 -0.0695 0.0444 0.0419 0.2172 -0.1482 0.0125 1.4608 -0.0253 0.0681 0.0866 -0.0364 0.1117 -0.1593 0.0709 0.0980 -0.0395 0.0646 -0.1876 0.0072 0.0047 -0.3355 0.0212 -0.0011 -0.1406 -0.0946 -0.0412 0.2241 0.0659 0.0379 -0.0737 0.1541 -0.0353 -0.0094 -0.0470 0.1624 -0.2474 0.1695 -0.1705 0.3022 -0.4618 0.1029 0.1895 0.0998 0.0124 -0.1222 -0.0837 0.0732 -0.0552 0.0545 0.1123 -0.0209 0.0668 -0.4615 -0.0033 -0.1103 0.2095 -0.1291 -0.0253 -0.0001 0.0285 -0.1384 0.0181\n",
      "a 0.0876 -0.4959 -0.0499 -0.0937 -0.0472 -0.0211 0.2624 0.0269 -0.0900 -0.0363 -0.0306 -0.0142 0.0337 0.1001 -0.1523 0.6262 -0.0115 -0.0063 0.0011 0.0125 -0.0592 0.0788 -0.0150 -0.0954 0.0027 0.0076 -0.0123 -0.0037 0.0022 0.2066 0.0584 0.0186 0.0608 -0.2230 0.0379 -0.0436 0.0293 -0.3732 0.1711 -0.0031 -0.1176 0.0889 -0.0107 0.0289 0.1135 0.3175 -0.0360 0.1177 -0.0032 -0.0286 0.0827 0.0294 -0.1301 -0.0617 -0.0188 0.1010 -0.0382 -0.2412 -0.2218 -0.0646 -0.0414 -0.0542 0.0312 -0.4290 -0.1143 -0.1295 -0.0414 -0.0602 0.0798 -0.0576 -0.0207 0.0654 -0.0058 -0.0901 -0.0713 -0.0966 0.1082 0.0701 0.4568 0.0263 -0.0104 -0.0207 -0.0029 0.0185 0.4614 -0.0357 0.1848 0.0604 0.0982 -0.1659 -0.0496 -0.1049 0.0558 -0.0604 -0.2724 0.2926 -0.1992 0.0047 -0.0613 -0.0197 0.0297 0.1301 0.0787 -0.0087 -0.0238 0.1296 0.0073 0.0485 -0.0556 0.0309 -0.0801 0.0640 -0.0663 -0.1790 0.0063 -0.0717 -0.0422 0.1437 -0.0923 -0.0329 0.0838 0.0847 0.0479 -0.0159 -0.0198 -0.0455 0.0566 0.3171 -0.0132 -0.0021 -0.0030 -0.1055 0.2173 0.3917 -0.0060 0.0089 -0.1015 0.0867 0.1599 -0.0567 0.0566 -0.0222 0.0343 0.0722 -0.1135 -0.0054 -0.8177 0.0457 -0.0387 0.0425 -1.0323 0.0132 -0.4213 0.1022 -0.0329 -0.0337 0.1715 -0.3718 0.0154 0.1216 0.0959 -0.0552 0.3990 0.0265 0.1393 -0.0912 0.0671 0.0573 -0.0525 -0.0711 0.0154 -0.0703 0.0895 -0.1358 -0.0460 -0.0456 0.0366 -0.0285 0.0584 0.0479 0.0916 0.3069 -0.1600 0.0581 -0.0455 0.1881 0.2715 -0.0432 -0.0345 -0.0180 -0.0139 -0.0344 0.1960 -0.0279 0.0761 0.0084 -0.0915 -0.0209 0.0457 -0.2600 -0.0404 1.2806 0.0451 0.0136 0.0885 0.0409 0.2782 -0.0724 0.0089 -0.2271 -0.1389 -0.0332 -0.3227 0.0117 -0.0545 0.0011 0.0816 0.0499 0.1140 0.2573 0.0516 0.0677 0.0472 0.0712 0.0811 -0.1531 0.0244 -0.0545 0.0177 0.0133 -0.0322 0.0894 -0.0334 -0.1128 -0.0228 0.2817 -0.0184 -0.0121 -0.1439 0.0245 0.0586 0.0980 0.0907 -0.0582 1.2555 -0.1311 -0.0550 0.1610 -0.1878 0.0115 0.0014 0.1759 0.0245 -0.1354 -0.0282 0.1128 -0.0907 0.0487 -0.1194 0.1444 -0.1511 0.0101 -0.0533 -0.1374 -0.0266 0.0477 0.1119 -0.1755 0.0768 -0.3507 -0.1941 -0.0894 0.1250 0.2225 -0.0788 -0.2031 -0.1970 0.0817 0.1293 -0.1875 0.1125 0.0538 -0.2714 -0.1384 0.0443 -0.0590 -0.0373 -0.0474 -0.0847 0.0545 0.4287 0.1403 -0.1608 0.1409 0.0109 -0.0062 -0.0242 0.5349 -0.0558 -0.0166\n",
      "</s> 0.0731 -0.2430 -0.0353 -0.3631 0.0380 0.0580 -0.1924 0.0283 -0.0737 0.0127 0.0917 -0.0233 0.0858 -0.0136 0.0569 0.1742 -0.0814 -0.0681 -0.0338 -0.1833 0.0175 0.0802 0.1190 -0.0762 -0.3242 -0.0395 -0.0796 0.0193 -0.0352 0.7771 -0.0255 -0.0422 0.0609 0.0792 -0.1056 -0.0170 0.2134 -0.3751 -0.0071 0.0697 0.0306 0.1578 -0.1587 0.0049 0.5308 -0.1463 0.0383 0.2731 0.1562 -0.0714 -0.1342 -0.0814 0.0615 0.0212 -0.0447 -0.0964 -0.2225 -0.0074 0.1077 0.0110 0.0169 0.0839 0.0349 0.5137 -0.0066 0.1021 0.1689 -0.0946 0.0339 -0.0342 0.4111 -0.0312 0.0602 -0.0502 0.1304 0.0027 0.0546 -0.1327 0.2900 0.1126 0.0759 -0.0372 0.2213 -0.0459 0.2853 -0.0216 -0.0895 0.1296 -0.0923 -0.0105 -0.0161 -0.1782 -0.2344 -0.0089 -0.2624 -0.0525 -0.3413 0.1510 0.0429 -0.1404 -0.0088 0.1659 0.1526 0.2601 -0.1245 0.1331 -0.0375 -0.0007 0.0764 0.0974 -0.0150 0.0942 -0.0188 -0.2265 0.0439 0.1505 0.1908 0.0461 -0.0173 -0.0496 0.1151 -0.1019 -0.0696 -0.0789 -0.1233 -0.1995 0.0457 -0.0318 -0.1206 -0.1645 0.0443 -0.0386 0.1336 -0.2531 0.0672 -0.1441 -0.4399 -0.1664 -0.0456 0.0037 -0.0939 -0.0639 -0.1184 0.0349 0.2886 -0.1034 -1.0094 -0.0419 -0.0683 -0.0083 0.0201 -0.0888 0.3062 0.1137 -0.0503 -0.0650 0.8131 0.0939 0.1026 0.0829 0.1135 -0.0774 -0.1822 0.0910 0.0850 -0.1230 0.0744 -0.0762 -0.1740 -0.0824 0.0926 -0.0573 -0.0370 -0.0519 -0.0065 -0.1886 0.0833 -0.1467 -0.0947 0.2036 -0.0790 0.8405 0.1763 0.6171 0.0773 0.0801 0.1462 0.0633 0.0065 -0.0058 0.0619 -0.0880 -0.0116 -0.0744 0.0365 -0.0335 0.1250 0.0718 -0.0050 -0.5228 -0.0419 -0.1213 -0.0524 -0.0024 0.1224 -0.0658 -0.5442 -0.1019 -0.0860 -0.4417 0.0790 0.1641 -0.2158 0.0772 0.0493 0.0037 -0.1223 0.1075 0.0419 -0.2768 -0.1969 -0.0005 -0.0065 -0.2081 -0.2939 -0.1594 -0.0341 -0.0719 -0.2962 -0.0822 0.1190 0.1834 0.2438 -0.1116 -0.1770 0.3082 -0.0419 0.0102 0.0299 -0.1144 -0.0023 -0.3917 0.4806 -2.2121 1.4759 -0.1241 -0.0648 0.0309 0.1354 0.1987 0.0993 0.0934 -0.1105 -0.1429 -0.0583 -0.0105 0.3661 0.1222 -0.0797 0.1509 0.0325 0.1190 -0.4640 0.0027 0.0642 -0.1214 -0.0172 0.0086 -0.2630 -0.0439 0.1087 -0.1628 0.4457 0.4626 -0.1086 0.3086 -0.1952 0.0850 0.3832 0.0482 0.1605 -0.0066 0.0749 0.0166 0.1219 -0.0253 0.0106 -0.0688 0.0705 0.0807 -0.2426 -0.0385 -0.2364 -0.1510 0.0645 0.1142 0.1037 0.4093 -0.0932 -0.0461\n",
      "in -0.0140 -0.2522 0.0715 -0.0246 -0.0637 -0.0364 0.1619 0.0268 -0.0475 0.0192 0.0102 0.0243 0.0691 -0.0131 0.0100 0.0384 0.0049 -0.0646 0.0062 -0.2644 -0.0436 0.0482 0.0965 0.0303 -0.0419 -0.0725 -0.0309 0.0358 0.0183 0.1035 0.1145 -0.0688 -0.0053 -0.2819 -0.2296 -0.0920 0.0032 0.7827 0.0972 -0.0344 -0.0813 0.1042 -0.0097 -0.0404 -0.1460 -0.0859 -0.0571 0.1671 -0.0321 -0.0204 -0.0316 0.0623 -0.1638 -0.0375 -0.0512 -0.2756 -0.2067 -0.2015 0.0878 -0.0350 -0.0385 -0.0004 0.0474 -0.1300 0.0052 -0.0484 0.0840 -0.1381 -0.0642 0.0140 -0.1044 -0.0456 0.0877 -0.0229 -0.0419 -0.1083 0.0768 -0.0058 0.0683 -0.3820 -0.0489 -0.0137 0.0042 0.0274 0.0018 -0.0764 0.0188 -0.0126 -0.0266 -0.0022 -0.0397 -0.0326 0.1004 0.0008 -0.3710 0.0143 -0.5057 -0.0563 -0.0324 -0.0150 0.0679 -0.0075 -0.0516 -0.1919 0.0231 0.0041 -0.0303 0.0057 0.0732 0.0722 0.0289 -0.1766 0.0102 0.5299 0.0130 0.0332 -0.0445 0.0497 0.0220 -0.0794 -0.0549 -0.0055 0.0152 -0.0195 -0.1831 -0.0861 0.0327 -0.1276 -0.0176 0.2193 -0.0152 0.0204 0.0681 -0.0003 -0.3027 -0.0579 -0.1060 0.1507 0.0009 0.2192 -0.0802 -0.0026 -0.0305 -0.0336 0.2453 0.0414 -0.8213 0.0120 -0.0473 0.0539 -0.4324 0.0062 0.2391 -0.0053 0.0046 0.0013 0.5302 -0.2481 0.0565 0.0618 0.0533 -0.0440 -0.2049 0.0077 0.0326 -0.0322 -0.0085 -0.1772 -0.0106 -0.1253 -0.0131 0.0267 0.0493 -0.0791 0.0806 0.0444 -0.0212 0.3405 0.0054 -0.0862 0.1064 0.1316 -0.0495 0.2879 -0.0545 0.2327 -0.0770 -0.0906 0.1125 0.1008 -0.1095 0.0341 -0.1275 0.0054 0.0216 -0.0083 -0.0621 0.0767 -0.1278 -0.0720 0.0083 0.6660 -0.0168 0.0465 0.0063 -0.0075 -0.0267 -0.0080 0.0482 -0.2656 0.1732 0.1060 0.0242 0.1088 -0.0683 -0.0114 0.1227 -0.0706 -0.1164 0.1825 -0.0261 0.1076 0.0820 -0.0716 -0.2241 0.0683 0.0585 0.0379 0.2511 -0.0656 0.0367 0.0144 -0.1207 0.0538 0.0330 0.1530 0.0071 -0.1196 0.0067 -0.0452 -0.0267 0.6865 0.0394 -0.0609 1.3707 -0.0411 0.0020 0.1064 0.1183 -0.4046 0.0442 0.1037 0.0727 -0.0822 0.0172 -0.0795 -0.0801 -0.0283 -0.1014 0.1521 0.0364 -0.1821 -0.1327 -0.1123 -0.0203 -0.0533 -0.0602 0.0791 0.3133 0.3044 -0.0951 -0.1197 0.1718 0.2629 -0.1348 -0.3196 0.1980 -0.4672 0.1230 -0.5156 0.0471 -0.2120 -0.1742 -0.1205 0.0024 -0.0434 -0.0344 -0.0141 -0.0519 0.0879 0.1052 0.0814 -0.0323 0.0973 -0.0105 0.0093 -0.0196 0.1370 0.0045 0.0329\n",
      "is -0.0978 -0.2083 -0.1037 -0.0160 -0.2402 -0.0449 0.0030 0.0990 -0.0404 -0.0504 -0.0924 0.0404 0.0830 -0.0256 -0.1788 0.3079 0.0285 -0.0808 -0.0433 -0.0420 -0.0159 0.1147 -0.0551 -0.0339 0.1651 -0.0201 0.0150 0.0626 -0.2056 0.3048 -0.0958 0.0060 -0.0651 0.0364 -0.0158 -0.0182 0.2086 -0.4443 0.1909 -0.1083 -0.1451 -0.0209 -0.0556 -0.0567 0.2259 -0.0547 -0.0839 0.1608 0.0939 0.0275 -0.0333 -0.2950 -0.0821 -0.0940 0.0043 -0.1230 -0.0602 -0.2598 -0.0817 0.0540 0.0116 -0.1107 -0.0486 -0.2980 -0.0263 0.0161 -0.0585 -0.0038 0.0380 0.0368 0.1951 -0.0937 -0.0006 0.0138 -0.0347 -0.1009 0.1686 0.0524 -0.3772 0.0031 0.0810 -0.0148 -0.0810 0.0772 -0.2656 -0.1271 0.2901 -0.1384 0.1152 -0.0215 -0.0714 -0.1168 0.1923 -0.1195 -0.0792 0.5574 -0.5475 0.0762 -0.0829 -0.0512 -0.0020 0.0862 0.1134 -0.0250 -0.0977 0.1127 0.0918 0.0310 0.0080 0.1225 -0.0017 0.1592 0.0369 -0.1098 0.0430 -0.0647 0.0321 0.1120 -0.0324 -0.1663 0.0000 0.1169 0.1254 0.0028 -0.0112 -0.0579 -0.0729 -0.0516 -0.0569 0.1858 0.0152 -0.1072 0.4025 -0.0794 0.0138 -0.0194 -0.5649 -0.1108 0.1086 0.1318 0.0355 0.0341 -0.1214 0.0629 -0.2418 -0.0021 -0.9340 -0.0566 -0.0238 -0.1142 -0.3444 -0.0380 0.1139 0.0745 0.0223 0.0569 0.2188 -0.0531 -0.0681 0.0225 0.0527 -0.0691 0.1215 0.0252 0.0545 -0.0111 -0.0229 -0.1493 -0.0350 0.0023 -0.0854 -0.0422 0.0735 -0.0132 0.0042 -0.0982 -0.0048 0.0791 -0.1055 0.0483 -0.0325 0.4715 -0.1421 -0.0189 0.1107 0.1312 -0.2176 -0.0576 0.0234 0.1053 -0.1001 -0.2986 -0.0767 0.0415 0.1016 0.0414 -0.0121 -0.1159 0.0325 0.0039 -0.0477 0.2933 0.0364 0.0420 0.1523 0.0761 0.0179 -0.0349 -0.1144 -0.0165 0.0105 0.0833 -0.3966 0.1685 -0.0388 -0.0719 0.0187 -0.0131 0.0672 0.0156 0.1755 -0.0502 -0.0014 0.1856 0.0880 -0.1462 0.1679 0.0651 0.0887 0.0032 0.0350 0.0006 0.3070 -0.0647 -0.0047 -0.2834 0.0249 -0.1193 -0.1071 0.0167 0.0274 -0.0180 0.2381 0.0085 1.3330 -0.1601 0.0897 0.1164 -0.2484 0.1466 -0.0237 -0.0212 -0.0854 -0.2000 -0.0135 -0.0526 -0.0919 -0.1161 -0.1572 -0.0364 -0.0179 0.6769 -0.1027 -0.0913 -0.1047 -0.1482 0.0630 -0.0930 0.2727 -0.2298 0.0120 -0.1923 0.0820 -0.0229 -0.0670 -0.2394 -0.4955 -0.1646 -0.0076 -0.2122 0.0807 0.0939 -0.7335 -0.2539 0.0710 -0.0294 0.1805 -0.0726 -0.0172 0.0195 0.0594 0.0362 -0.1824 0.1408 -0.0868 0.0243 0.0660 -0.2282 -0.2244 0.0890\n",
      ": 0.1520 0.1851 -0.0915 -0.4386 0.0684 0.0303 -0.2930 0.0955 0.2615 -0.0498 -0.0039 -0.0551 0.0090 -0.0570 0.0449 -0.4179 -0.0935 -0.2237 0.0405 0.1532 -0.0269 -0.0148 0.0992 -0.2204 0.7670 0.0280 -0.1019 0.1751 -0.0644 0.3343 -0.1382 0.0470 0.0367 0.1819 0.0158 0.0884 0.1500 -0.1227 0.0290 -0.1311 -0.0210 0.0090 0.0909 -0.1408 0.5060 0.0965 -0.0392 -0.0635 0.0618 -0.0576 -0.1112 -0.1349 0.1029 0.0562 -0.0630 0.0708 -0.3590 0.1202 0.1229 0.1538 -0.0429 -0.0046 0.0409 0.5766 -0.0216 -0.0260 0.0358 0.1788 -0.0352 0.0370 0.3611 -0.1466 -0.1116 -0.0680 -0.0105 0.0869 0.1860 -0.0310 0.2641 -0.1662 0.1304 -0.0471 0.0741 0.0218 0.2294 -0.1960 -0.4922 -0.0364 -0.0381 -0.1393 -0.0131 -0.0702 0.3853 -0.0945 -0.0810 -0.1175 -0.0680 0.1685 0.1982 0.0537 0.0155 0.0438 0.1582 0.1062 -0.1922 -0.0190 0.0374 0.0249 -0.0202 -0.0605 0.0091 -0.0113 0.0379 -0.2055 0.1788 0.1592 0.1490 0.0318 -0.0837 0.0903 0.0626 0.0246 -0.1269 0.0422 -0.0348 0.0053 -0.0462 -0.1426 -0.0432 0.1103 0.1425 -0.0452 0.0765 -0.0630 0.4051 -0.1054 -0.3370 -0.2450 -0.2054 -0.0609 -0.1640 0.0556 -0.0798 0.0716 0.2087 -0.1769 -1.0498 -0.1150 0.0320 -0.0239 -0.3379 0.0478 -0.2771 -0.0087 -0.0047 -0.0263 0.4287 -0.0337 -0.0497 -0.0322 -0.0730 0.0120 0.0143 0.0455 -0.0223 -0.1317 -0.0076 -0.0491 0.0467 0.0610 -0.0035 -0.0071 -0.0478 -0.0374 0.0365 -0.0856 0.0585 -0.3594 -0.1104 0.2408 -0.0410 0.3899 0.0311 0.7779 -0.0029 0.0743 0.0991 0.2239 0.0939 0.0178 0.0243 0.0341 -0.0436 -0.1935 0.0889 0.0159 -0.1209 0.1003 -0.1229 -0.2013 -0.2963 -0.0657 0.0055 0.1958 0.1631 -0.0654 0.3708 0.0307 0.0897 -0.2486 -0.4385 0.0642 0.0502 0.0887 -0.0523 0.1300 -0.1803 -0.0876 -0.1684 -0.1154 -0.0802 0.0565 0.0382 -0.1689 -0.0449 -0.1741 0.0014 -0.0397 0.1221 -0.1089 0.1871 0.2371 0.4241 -0.1014 -0.1234 0.1433 -0.0794 -0.0143 0.0224 -0.0245 -0.0951 -0.3457 0.2254 0.3384 1.1417 -0.0018 0.1206 0.1256 0.0675 -0.0367 0.0582 0.4003 0.0501 -0.0209 -0.0036 0.0595 0.1038 -0.0043 0.0210 0.0239 0.1949 0.2109 -0.2935 0.1378 0.2973 0.0401 0.0827 -0.0980 0.1169 0.3763 0.0692 -0.1560 0.5115 0.1317 -0.4118 0.2811 0.1768 0.1111 0.2839 0.0984 -0.1727 -0.0261 0.0556 0.0713 0.1222 -0.0554 0.0278 0.2176 0.2008 -0.1166 -0.0197 -0.0950 -0.2453 0.1115 0.0568 0.0599 0.1660 0.2164 -0.0346 -0.1456\n",
      "I 0.1005 -0.1526 0.0444 -0.0668 0.0463 -0.1002 -0.3285 0.0127 -0.0449 -0.0292 0.0761 -0.1642 0.1231 -0.0661 -0.0123 0.3127 -0.0406 0.0029 -0.1124 0.0940 0.1951 0.0827 0.0042 -0.2475 -0.0327 -0.1690 -0.1027 0.0269 0.0670 0.5836 -0.0241 -0.1423 -0.1059 0.1424 0.0784 -0.1687 0.2657 -0.0331 -0.0726 -0.0862 -0.0177 0.2105 -0.0302 0.3947 0.3859 0.0912 0.0665 0.3924 -0.1110 -0.1178 0.0337 0.0416 -0.1206 0.0337 -0.1139 -0.0546 -0.2124 0.1006 -0.2229 0.0330 -0.0182 -0.0595 0.0442 -0.1952 -0.1789 0.0327 -0.2021 -0.1305 0.0086 -0.1089 -0.0263 0.0316 0.0144 -0.1678 0.0530 -0.1659 0.0457 -0.0866 -0.3472 0.1207 -0.1024 -0.1734 0.0219 0.0937 0.1014 -0.1120 -0.1019 -0.0312 -0.0344 0.0308 -0.0743 0.0188 0.7645 -0.1433 0.0055 -0.2545 -0.1283 0.1593 0.0137 -0.1239 0.1429 0.0973 0.0093 -0.4845 0.0916 -0.0875 0.0556 0.1236 0.0346 0.0065 0.0694 -0.2709 0.0231 0.2466 -0.0595 -0.1200 -0.1220 0.1440 0.1068 0.0572 0.0504 0.0662 0.0295 0.0028 -0.0191 0.4397 0.0736 -0.1598 0.0310 -0.0154 -0.0629 -0.0095 0.0735 -0.1619 0.2568 -0.2515 -0.5408 0.0074 -0.0946 0.0573 -0.0796 -0.0090 -0.0344 0.0790 0.3527 0.0866 -0.6917 -0.2032 0.0488 0.2144 0.0284 0.0673 0.0890 0.1381 0.0896 -0.0101 0.2654 0.1775 0.0985 0.2773 0.0340 0.1046 0.0303 0.0194 0.1328 -0.1764 0.0878 -0.0657 0.0559 0.0934 0.0572 -0.0816 -0.0354 -0.0744 0.1421 -0.1359 0.1603 -0.2699 0.1916 0.1363 0.1930 0.4175 0.0180 0.3501 0.1414 0.4974 0.3241 -0.0016 -0.0011 0.0771 -0.0412 0.0018 0.0058 -0.2695 -0.0219 -0.2563 -0.2134 -0.0537 0.0839 0.0934 -0.2283 -0.0936 -0.0331 -0.1140 -0.0550 0.0030 0.2360 -0.1259 0.1765 -0.0000 0.0664 0.1741 -0.3637 0.0308 -0.0139 0.0489 0.0918 0.0851 0.1021 0.0380 0.0772 0.1537 0.0403 0.0531 -0.0483 -0.2816 -0.0840 -0.0883 -0.1046 -0.1555 0.1389 0.1754 -0.0261 -0.2408 -0.0021 -0.0416 -0.0851 0.0361 0.0664 -0.1364 -0.0558 -0.2484 -0.0435 -0.0038 1.2010 0.0417 -0.0404 0.0301 -0.0398 0.3214 0.1626 -0.0192 -0.1118 -0.2500 0.0294 0.2138 0.1345 -0.0177 0.0060 0.2732 0.0932 0.2956 -0.0342 0.0560 0.0105 -0.0438 -0.1758 0.0315 -0.2645 -0.3122 -0.1178 -0.0228 0.1255 0.1257 0.1983 -0.0143 -0.1180 0.2400 0.1021 -0.0490 -0.1513 -0.0104 -0.8734 0.0057 0.1652 -0.0429 0.0654 -0.3860 0.0067 0.0549 0.0820 -0.0930 -0.3220 -0.0947 0.1407 0.0617 0.0043 0.6786 -0.1639 0.0854\n",
      "for 0.0164 -0.0646 -0.0208 0.0630 -0.0401 0.0282 0.1083 -0.0249 0.0119 -0.0002 -0.0297 0.0213 -0.0081 0.0038 -0.0148 -0.0748 0.0255 0.0230 -0.0338 0.1943 -0.0562 0.0140 -0.0438 -0.1480 -0.1865 -0.0391 -0.0176 0.0253 -0.0056 -0.0648 -0.0964 -0.0003 -0.0727 -0.1439 -0.0235 0.0264 0.0581 0.0381 0.0586 0.0567 -0.0867 -0.0143 0.0713 0.0273 -0.0062 -0.0670 -0.0492 -0.1341 0.0179 -0.0069 0.0434 0.0660 -0.0409 -0.0289 0.0417 0.0347 0.0610 -0.1492 -0.1740 0.0113 0.0245 -0.0223 -0.0783 -0.0345 -0.0837 -0.0258 -0.0073 -0.0079 0.0592 -0.0099 0.0829 0.0036 0.0212 -0.0572 0.0144 0.0039 0.0492 0.0802 0.1475 -0.0935 0.0262 0.0473 -0.0142 0.0081 0.0938 0.0373 0.1557 0.0701 -0.0163 0.1426 -0.0085 0.0303 -0.0241 0.0367 -0.0662 0.0180 0.0872 -0.0115 0.0254 0.0322 0.0048 -0.0420 -0.0016 0.0412 -0.0046 -0.0029 -0.0456 -0.0051 -0.0005 -0.0097 0.0641 0.1212 -0.0107 0.2008 0.0414 -0.0805 -0.0039 -0.0075 0.0122 -0.0234 0.0550 0.0281 -0.0552 0.0044 -0.0386 0.0839 -0.0206 0.1980 0.0056 0.0114 -0.0286 -0.0221 -0.0299 0.0455 0.0434 -0.0249 -0.0426 0.0092 -0.0195 0.0004 0.0078 0.0182 0.0466 -0.0168 -0.0753 0.0707 -0.3519 -0.0144 0.0328 0.0074 -0.1088 0.0320 0.0216 0.0017 -0.0677 -0.0252 0.2098 0.0288 -0.0505 0.0172 -0.0358 0.0888 0.0256 0.0110 -0.0603 0.0127 0.0267 -0.1252 -0.0212 0.0367 -0.0109 -0.0114 0.0236 -0.0244 -0.0051 0.0292 -0.0937 -0.1515 -0.0043 -0.0180 0.0868 -0.0676 0.0286 -0.0345 0.0283 -0.0576 -0.2236 -0.0985 0.0725 -0.0590 -0.0143 0.0128 0.0753 -0.0073 0.0039 0.0236 0.0309 0.0405 0.0380 0.2817 -0.1004 0.1951 -0.0691 -0.0515 -0.0298 0.0212 -0.0615 -0.1500 0.0786 0.2217 -0.0907 0.0133 0.1682 0.0239 0.0981 -0.0060 -0.0280 -0.0213 0.0004 0.1464 0.0941 -0.0566 0.0371 0.0408 0.0152 0.0175 -0.0059 0.0031 -0.0755 0.0215 -0.0137 -0.0526 0.0528 0.0730 0.0054 0.1281 0.0267 -0.0230 -0.0133 0.0329 -0.0345 0.1613 0.0274 -0.1763 0.5313 0.0004 0.0256 -0.0107 0.0085 -0.1757 -0.0680 0.0183 -0.0066 -0.0450 0.0040 0.0234 -0.0623 0.0131 0.1963 0.0068 0.0470 -0.0430 -0.0003 0.0363 -0.2440 -0.0156 0.0005 0.0004 0.2251 -0.1247 -0.0123 -0.0172 -0.0611 -0.0918 -0.1271 -0.1214 -0.0638 -0.2065 -0.0581 0.3505 -0.1634 0.0543 -0.0910 0.0067 -0.0172 0.0043 0.0790 -0.1132 -0.0420 0.0069 0.0562 0.0680 -0.0290 -0.0891 0.0618 0.0123 -0.0284 -0.0355 0.0179 0.0385\n",
      "that -0.0315 0.0328 0.0396 0.0911 -0.0911 -0.0352 -0.0009 -0.0402 -0.0375 -0.0239 -0.0125 0.0045 0.0145 -0.0080 -0.0294 -0.0192 0.0126 0.0463 -0.0257 0.1033 0.0392 -0.0065 -0.0089 0.0179 -0.0921 -0.0201 0.0198 -0.0048 -0.0173 -0.0234 -0.0065 -0.0106 0.0028 -0.0993 0.0146 0.0067 0.0254 -0.0242 0.0094 -0.0082 0.0140 0.0077 -0.0032 0.0542 0.0773 0.0402 0.0013 0.0522 -0.0035 0.0111 0.0128 -0.0112 -0.0490 -0.0164 -0.0039 -0.0084 -0.0130 -0.0585 -0.0480 -0.0087 -0.0482 -0.0363 -0.0266 0.0277 -0.0203 0.0066 -0.0403 -0.0069 -0.0195 -0.0188 -0.0943 -0.0571 0.0370 -0.0348 -0.0289 -0.0395 0.0172 0.0177 0.0805 -0.0071 -0.0280 -0.0054 -0.0705 0.0345 -0.0923 0.0000 0.0163 -0.0949 0.0641 0.0096 -0.0046 -0.0086 0.0471 -0.0187 0.1522 -0.0013 -0.0123 -0.0136 -0.0249 -0.0005 0.0246 0.0117 0.0194 -0.1126 0.0168 0.0036 -0.0046 -0.0321 -0.0395 0.0106 0.0192 -0.0207 -0.0151 0.0664 -0.0012 -0.0548 -0.0505 -0.0080 0.0191 -0.0250 0.0201 0.0277 0.0279 0.0037 0.0373 0.0679 -0.0029 -0.0465 0.0028 0.0510 -0.0254 -0.0112 0.0116 -0.0288 0.0326 0.0077 -0.0900 -0.0442 0.0141 -0.0040 -0.0312 0.0035 0.0262 0.0126 -0.0934 -0.0024 -0.2439 0.0279 -0.0205 -0.0065 0.0055 0.0102 0.0183 0.0126 0.0049 0.0199 0.1181 -0.0261 0.0167 0.0451 -0.0022 -0.0002 0.0798 -0.0292 0.0073 -0.0389 0.0062 -0.0249 0.0657 0.0224 -0.0087 -0.0241 0.0580 -0.0262 0.0143 0.0165 -0.0015 -0.0403 -0.0016 -0.0028 -0.0200 0.0862 -0.0448 0.0178 0.0111 0.0378 -0.0090 -0.0223 -0.0337 -0.0100 -0.0160 0.0416 -0.0347 0.0351 0.0236 0.0114 -0.0056 -0.0036 0.0078 0.0503 -0.0559 0.0084 0.0199 -0.0255 -0.0327 0.0134 -0.1264 0.0058 -0.0385 0.0297 0.0403 0.0297 -0.0978 -0.0255 0.0173 -0.0412 0.1068 0.0198 0.0085 0.0441 0.0618 -0.0154 -0.0172 0.1022 0.0107 -0.0465 0.0136 0.0410 0.0754 -0.0447 -0.0158 -0.0445 0.0662 0.0056 0.0227 -0.2409 0.0068 0.0477 -0.0261 0.0243 0.0089 0.0537 -0.0492 -0.1550 0.3057 0.0086 0.0093 -0.0281 0.0176 0.0303 0.0181 -0.0115 -0.0260 0.0080 0.0081 0.0461 -0.0605 -0.0384 -0.0158 0.0096 -0.0263 -0.0249 -0.0016 0.0114 -0.1063 -0.0108 -0.0016 0.0086 0.0453 -0.1011 0.0052 -0.0268 -0.0285 -0.0585 0.0535 -0.0038 -0.1147 0.0050 -0.0794 -0.0027 -0.0450 -0.0160 -0.1989 -0.0327 0.0145 -0.0385 0.0102 -0.0405 -0.0412 -0.0121 0.0310 -0.0199 -0.0911 0.0213 0.0008 0.0077 -0.0035 0.1924 0.0240 0.0391\n",
      ") -0.1294 -0.0707 -0.1770 -0.3495 0.1993 -0.0204 -0.0645 -0.0485 -0.0556 -0.1879 0.1331 -0.1358 -0.0299 0.1571 0.0209 0.0561 -0.0295 0.0618 0.1310 -0.0462 0.0905 0.0470 0.1740 -0.0171 0.2008 -0.2794 0.0645 0.0543 -0.0179 0.1114 0.0348 0.0127 -0.0442 0.0666 -0.1650 0.1543 -0.0306 -0.3294 0.0938 -0.0386 0.2285 -0.0119 -0.0418 -0.0579 0.3339 -0.2317 -0.0088 0.1306 -0.1558 -0.0166 0.0738 -0.0554 0.0648 0.0484 -0.0663 -0.0555 -0.0937 -0.0161 0.0223 0.0302 0.0078 -0.0268 -0.0380 -0.0793 -0.0108 0.1014 0.0694 0.0359 -0.1114 0.0095 0.1606 -0.1377 -0.1415 0.1692 -0.1499 0.0659 -0.0341 -0.2062 -0.1006 0.1730 0.0643 0.1092 -0.1726 -0.0238 -0.1819 -0.2477 -0.0979 0.1714 -0.1653 -0.2287 -0.1402 0.0175 0.0655 -0.0482 -0.1623 -0.2397 -0.0674 -0.1481 0.2533 -0.1451 -0.1315 0.0293 -0.0560 0.1375 0.1364 0.0011 -0.0985 0.2298 -0.1992 0.0142 -0.1114 0.0320 -0.0046 -0.5553 0.0618 0.2435 -0.1011 0.1261 -0.0333 -0.0480 -0.0345 0.0306 0.0923 -0.0400 -0.0207 0.0589 0.1141 0.0860 -0.1277 -0.1928 0.0491 0.1091 0.0041 -0.4097 -0.1576 0.1890 -0.3534 0.0265 -0.0266 -0.0473 -0.0441 0.0276 -0.0602 0.0535 -0.0715 -0.0416 -0.8557 -0.1591 -0.0245 0.1619 0.1819 -0.0177 -0.0138 -0.0492 -0.0172 0.0801 0.5441 0.0934 -0.0232 -0.1086 -0.0283 -0.0053 0.1014 0.0793 -0.0101 0.0135 -0.1616 -0.0878 -0.0320 0.0609 0.0127 -0.0384 0.0862 -0.2751 -0.0271 0.1261 -0.1229 0.1296 -0.1008 -0.1402 0.0195 0.3986 0.2592 0.5463 0.0497 0.2489 -0.1316 0.2526 0.1172 -0.0581 -0.0759 -0.1524 0.0110 -0.1625 0.1177 -0.1402 0.1600 0.0876 0.0721 0.0049 -0.1626 -0.3656 -0.1388 0.1224 0.0133 -0.2050 -0.0546 -0.0622 0.1677 -0.2233 0.1447 -0.0367 -0.0938 -0.0761 0.0260 -0.0772 -0.0087 0.2498 0.0455 -0.1581 -0.1740 -0.1686 0.1030 -0.0830 0.1224 -0.1634 0.1093 0.1011 -0.0683 0.0373 0.1240 0.0580 -0.0158 -0.0657 -0.1788 -0.0038 0.0219 -0.1074 0.0286 -0.1103 0.2348 -0.2109 0.1144 -0.1663 1.1676 0.0586 -0.0434 0.0772 -0.0249 0.3144 0.3256 -0.0734 -0.0005 -0.1325 0.0472 -0.0223 0.2340 -0.0740 -0.0852 0.1226 -0.0268 0.0051 -0.4937 0.0936 0.2085 -0.0851 -0.1078 0.1068 -0.0511 0.1410 0.2127 -0.1263 0.4264 0.2534 0.0424 0.0722 -0.0767 0.1935 0.3299 0.0428 0.2237 -0.0300 -0.0987 -0.1538 -0.1079 -0.0033 0.0359 0.0070 0.0645 0.0917 -0.1394 0.2283 0.2056 -0.1032 0.0763 0.0853 0.0675 0.2903 -0.3215 -0.1485\n",
      "\" 0.1037 -0.1643 -0.0056 -0.1876 -0.7213 0.0092 -0.2550 -0.0379 -0.0599 -0.0475 -0.1430 -0.0229 0.0841 0.0474 0.2982 0.1792 -0.3178 -0.0987 0.0166 0.0085 0.1027 0.0063 0.0054 0.2031 0.1078 0.1671 0.0433 0.1157 -0.0756 0.2992 -0.1194 0.1094 0.0194 0.2029 -0.1528 0.0008 0.0370 -0.0135 -0.0223 -0.0783 -0.0860 -0.1309 0.0410 0.1672 0.3231 -0.0706 -0.1643 0.2100 0.0723 0.0526 -0.2421 0.1197 0.1848 -0.0709 0.0013 0.1197 -0.3189 -0.0023 -0.0081 -0.0836 0.1671 -0.0224 -0.1761 0.2477 -0.0900 0.0483 -0.2923 -0.0646 0.0698 -0.1731 0.2257 -0.0482 -0.2583 0.0339 -0.0758 -0.0825 0.0353 0.0927 -0.0423 -0.0263 0.1638 -0.0724 0.0087 0.1809 -0.0969 -0.0775 -0.0818 0.0189 -0.0188 -0.1780 -0.0251 0.0437 -0.1148 -0.2080 0.0196 -0.1068 -0.2438 -0.0366 0.1458 0.0766 0.0776 0.0054 0.0781 0.1434 0.1945 0.0943 0.1301 0.1061 -0.0123 -0.1184 -0.1463 0.1211 0.0611 -0.0996 -0.0547 -0.0352 0.0429 0.0897 -0.0725 0.0926 0.0835 0.0468 0.1516 -0.1358 -0.0392 0.2930 -0.0457 -0.0793 -0.2064 -0.1310 0.1396 -0.1466 0.1240 -0.0986 -0.2456 0.0922 -0.2289 -0.1192 0.0634 -0.0210 0.1373 -0.0705 -0.1308 -0.0771 0.0246 -0.1944 -0.8830 0.0439 0.1435 -0.1719 -0.1654 0.0449 -0.0850 -0.0887 -0.1148 -0.1745 0.2850 0.0029 0.1059 0.3099 0.0676 -0.0272 0.0600 -0.1776 0.1134 -0.1276 0.1747 -0.1441 -0.0196 0.0033 -0.0090 -0.0312 0.2433 -0.0115 0.1903 -0.0358 0.0563 0.0071 -0.0587 0.1300 -0.1227 -0.1328 -0.1326 0.2324 -0.1617 0.5448 0.1785 0.0890 -0.0886 0.2186 -0.0126 -0.2045 -0.0188 -0.0571 0.0388 -0.0225 0.0099 0.1973 0.0726 -0.1233 -0.3189 -0.0816 0.0365 0.3008 0.8099 -0.1178 -0.0411 -0.0922 0.1056 -0.0313 -0.2130 -0.0539 -0.0196 0.0499 -0.2320 -0.0730 -0.0803 0.0255 0.1401 -0.1046 -0.2058 -0.1043 0.0557 -0.2410 -0.1706 -0.1541 0.0415 -0.0634 -0.0704 0.1294 -0.0227 -0.2413 0.1780 0.0112 0.0439 -0.0591 0.0615 0.1400 -0.1883 -0.2210 0.0628 -0.0381 0.2845 0.0439 1.0331 0.2443 0.2122 0.2826 -0.1303 0.0975 -0.2550 -0.0176 -0.0347 -0.0981 0.0488 0.2088 0.1953 -0.0790 0.1349 0.0673 0.0074 -0.0432 -0.5623 -0.0032 0.1135 0.0404 0.1080 -0.2515 -0.1028 0.0216 0.0915 -0.3115 0.2312 0.1482 -0.0192 0.3239 -0.1103 -0.1399 0.0533 -0.0357 -0.0135 0.0985 -0.0832 -0.1566 0.1319 -0.0616 -0.0623 0.3486 -0.2015 -0.0771 0.0198 0.0023 0.1318 -0.1160 0.0468 -0.1154 0.0897 0.2778 0.1692 0.1266\n",
      "( -0.1807 -0.0722 -0.2748 -0.2896 0.1906 -0.0032 -0.2037 -0.0761 -0.0372 -0.1973 0.1609 -0.1085 -0.0480 0.1580 0.0240 0.0288 -0.0170 0.0905 0.1110 -0.1882 0.0814 0.0476 0.1854 -0.0408 0.1215 -0.2840 0.0533 0.0600 -0.0284 0.1787 0.0356 0.0101 -0.0614 0.2779 -0.1701 0.1729 -0.0835 0.1200 0.1363 -0.0592 0.2499 0.0145 -0.0218 -0.0183 0.4841 -0.0753 -0.0098 0.1441 -0.1416 -0.0291 0.0956 -0.0544 0.0434 0.0692 -0.0736 -0.0846 -0.0959 -0.1032 0.1335 0.0351 0.0004 -0.0403 -0.0310 0.0848 -0.0380 0.0956 0.0926 0.0375 -0.1108 -0.0265 0.2043 -0.1385 -0.1569 0.1177 -0.1569 0.1173 -0.0133 -0.2108 -0.2251 0.0465 0.0758 0.1162 -0.1475 -0.0276 -0.1966 -0.2403 -0.1333 0.0206 -0.1752 -0.0994 -0.1521 0.0035 0.1295 -0.0536 -0.2818 -0.0842 0.2333 -0.1334 0.2518 -0.1522 -0.1354 0.0548 -0.0480 0.0584 0.1877 0.0227 -0.0873 0.2316 -0.1600 0.0415 -0.0946 0.0897 -0.0061 0.3014 0.0666 0.2452 -0.1131 0.0961 -0.0689 -0.0427 0.0025 -0.0112 0.1323 -0.0682 -0.0717 -0.0055 0.1254 0.1095 -0.1413 -0.2590 0.0308 0.0870 0.0007 -0.2519 0.0436 0.1960 -0.3329 0.0109 -0.0107 -0.0231 -0.0538 0.0424 -0.0745 0.0628 0.1049 -0.0276 -0.6871 -0.1660 -0.0479 0.0900 -0.5195 -0.0231 -0.0925 -0.0618 -0.0212 0.0612 0.5308 0.1137 0.0099 -0.0544 0.0047 -0.0189 -0.0964 0.1134 0.0070 0.0321 -0.1390 -0.0527 -0.0802 0.0588 0.0039 -0.0162 0.1130 -0.2697 0.0023 0.1106 -0.1120 0.1004 -0.1052 -0.1034 -0.0128 -0.0221 0.2811 0.4153 0.0335 0.3179 0.1182 0.2864 0.1356 -0.0097 -0.0706 -0.1609 0.0731 -0.1323 0.1463 -0.1307 0.1828 0.0941 0.0752 -0.2505 -0.1396 -0.0126 -0.1606 0.1149 0.0247 -0.1957 0.0707 -0.0267 0.1908 -0.2644 -0.0846 -0.0332 -0.0032 -0.0640 0.0040 -0.0890 0.0241 0.2820 0.0191 0.0811 -0.2053 -0.2237 0.1088 0.0553 0.0646 -0.1395 0.1134 0.1229 -0.2117 0.0610 0.1285 0.0762 0.0587 -0.0773 -0.2060 0.1867 0.0364 -0.1332 0.0526 -0.1087 0.2284 -0.1820 0.1705 -0.0177 1.3595 0.0572 -0.0466 0.0513 -0.1148 0.2699 0.2649 -0.0663 -0.0200 -0.1380 0.0319 -0.0465 0.2491 -0.0713 -0.0987 0.1562 0.0036 0.3296 -0.5862 0.1071 0.1259 -0.0871 -0.1304 0.0864 0.0091 0.2876 0.2583 -0.1588 0.1114 0.4166 0.0198 0.0924 0.0050 -0.1099 0.2019 -0.0816 0.2279 0.0209 -0.1850 -0.1784 -0.1358 0.0044 0.0161 0.0328 0.0440 0.0861 -0.0143 0.2263 0.3275 -0.0869 0.0447 0.0846 0.0747 0.0215 -0.3351 -0.1350\n",
      "on -0.0568 0.5490 0.1312 0.1825 -0.0408 -0.0062 0.2756 -0.0198 0.0183 -0.0160 -0.0081 0.0360 -0.0153 -0.0049 -0.0417 0.0403 0.0401 -0.0267 -0.1116 -0.2787 0.1538 0.1320 0.1024 -0.1156 0.6145 -0.0615 -0.1039 0.0562 0.0147 0.1679 -0.0028 -0.0880 -0.0525 -0.4669 0.0498 0.0272 -0.1842 0.4104 -0.1249 0.0098 0.0071 0.0769 0.0762 0.0705 0.1268 -0.0701 -0.1009 0.3316 -0.0348 0.0534 -0.0575 0.0349 -0.0476 -0.0716 0.0405 -0.0107 -0.1956 -0.0848 -0.1107 -0.0128 0.0449 -0.0096 -0.1287 -0.2576 -0.0685 -0.0336 -0.0653 -0.1169 0.1110 0.0423 0.1279 0.1603 0.0823 -0.1279 0.0366 -0.0609 0.1506 0.0961 0.1148 -0.0928 -0.0009 -0.1419 -0.0296 0.0456 -0.0459 0.0121 0.0517 0.4279 0.0471 -0.1484 -0.1828 -0.0388 0.1439 -0.0827 -0.2111 -0.0660 0.0399 0.0133 0.0430 -0.0681 0.0925 0.0555 0.0801 0.3372 -0.0458 0.1663 0.0224 0.0461 0.0460 -0.0178 -0.3178 0.0543 0.0446 0.4307 -0.0349 0.0151 -0.0390 -0.0657 -0.0651 0.0720 0.3747 -0.0347 0.1499 0.0274 0.0745 -0.0155 0.0944 0.4920 -0.0243 0.1211 -0.0870 -0.1139 0.0743 -0.0295 -0.1761 0.0197 -0.2513 -0.0522 0.0263 0.1106 -0.1947 -0.0525 -0.0741 0.0199 0.3135 -0.4271 -0.9073 0.1220 -0.0476 -0.0084 -0.4623 0.0767 0.2873 0.0893 -0.0193 -0.0586 0.5378 0.5018 0.0828 0.0967 0.1514 0.0824 -0.1947 -0.0142 0.0557 0.0023 0.1833 0.0243 0.0389 -0.0191 0.0738 -0.0018 -0.0258 -0.0766 -0.0160 -0.1098 -0.0549 -0.4029 0.0500 -0.0458 0.1226 0.1168 0.3314 0.2458 -0.0342 0.1620 -0.3570 -0.0573 -0.1383 0.1578 -0.0674 -0.0416 0.0741 -0.0819 0.0324 -0.0527 -0.0868 -0.2240 0.1741 -0.2539 -0.0041 0.8346 0.0761 -0.1556 0.0701 -0.1192 -0.2436 0.1884 0.1327 0.0352 -0.2928 0.0570 0.1620 0.1682 0.1449 0.0066 -0.0615 0.1636 0.0748 -0.2150 -0.0021 0.0149 0.0544 0.0401 0.1679 0.1663 -0.0974 -0.0433 -0.0207 -0.0601 0.0427 0.0805 0.0262 0.0798 -0.0023 0.0356 0.0350 0.2931 -0.1686 -0.0033 -0.1507 0.5915 0.1437 -0.0168 1.3037 -0.1686 0.1783 0.2458 -0.1563 -0.2926 -0.0514 0.5490 0.0849 -0.0276 0.0040 0.0481 -0.1184 0.1071 -0.2086 -0.0495 -0.0042 -0.0645 -0.0876 -0.0686 0.1854 -0.1137 0.0602 0.1218 -0.2715 0.3126 0.0475 -0.0847 -0.0643 -0.0037 0.0508 -0.3443 0.4078 -0.2287 0.0779 -0.4201 0.1812 -0.0810 -0.2354 0.0234 0.0170 -0.0332 0.0292 -0.1043 -0.0735 0.1356 -0.1026 0.0028 -0.1814 -0.1754 -0.0805 0.0561 0.0843 -0.0334 -0.0634 -0.0605\n",
      "with 0.0149 -0.1152 -0.0248 0.0003 -0.0575 -0.0008 0.0758 -0.0099 -0.0632 -0.0068 -0.0139 0.0043 -0.1062 0.0298 -0.0319 -0.0808 0.0071 -0.0032 0.0014 -0.0511 -0.0650 -0.0017 -0.0039 0.0779 0.0762 -0.0441 0.0127 0.0101 0.0526 -0.0489 0.0359 0.0116 0.0642 0.0449 0.0241 0.0289 0.0527 0.0157 0.0003 0.0129 0.0270 0.0022 -0.0983 0.0424 -0.0151 -0.0084 0.0010 0.0477 -0.0123 0.0366 0.0687 0.0318 -0.0801 0.0075 0.0122 -0.0810 0.1157 0.0406 -0.1213 -0.0499 0.0488 -0.0109 -0.0050 0.0087 -0.0224 -0.0802 -0.0344 -0.0021 -0.0014 -0.0075 -0.0737 -0.0365 0.0180 -0.0863 -0.0309 -0.0071 0.0244 0.0524 0.1118 -0.0127 -0.0242 0.0116 0.0231 0.0199 0.0969 0.0123 0.0272 0.0769 0.0071 0.0126 0.0123 0.0211 -0.0156 0.0088 -0.0007 0.1047 0.0499 -0.0068 0.0063 0.0082 0.0296 0.0134 -0.0408 -0.0181 -0.0044 -0.0519 -0.0054 0.0323 -0.0007 0.0209 0.0499 -0.1140 -0.0355 0.0596 0.0369 -0.0451 -0.0033 -0.0105 0.0272 -0.0412 0.0574 0.0316 -0.0990 0.0110 0.0061 0.0497 -0.0034 -0.0420 -0.0188 -0.0307 -0.0118 -0.0003 -0.0707 0.0660 0.0323 -0.0198 -0.0731 -0.0722 0.0153 -0.1041 -0.0121 0.0287 0.0184 -0.0075 0.1018 0.0404 -0.2646 0.0375 0.0449 -0.0205 -0.0746 0.0058 0.0178 0.0077 0.0168 0.0377 0.1773 -0.0024 -0.0487 -0.0118 0.0220 -0.0062 0.0406 0.0023 0.0024 -0.0425 0.0364 0.0077 0.0846 -0.0574 0.0470 -0.0168 0.0169 -0.0024 -0.0980 0.0416 -0.0234 0.1231 0.0283 -0.0936 0.0003 -0.0518 -0.0629 -0.0131 0.0089 -0.0334 -0.0773 -0.0323 -0.0067 -0.0915 -0.0194 -0.0240 0.0152 0.0143 -0.0068 0.0114 -0.0192 0.0628 0.0092 -0.0022 -0.0454 0.1369 0.0497 -0.0245 -0.0335 -0.0048 0.0357 -0.0037 -0.0073 0.0738 -0.1363 0.0219 0.0839 0.0150 -0.0291 0.0244 0.0370 0.0003 -0.0268 -0.0164 0.1265 0.0196 -0.0018 0.1252 0.0345 -0.0438 0.0192 0.0096 -0.0485 -0.0116 -0.0133 0.0029 0.1089 0.0103 0.0020 0.0697 -0.0015 0.0190 -0.0027 -0.0049 0.0098 0.0661 0.0872 -0.1296 0.3899 0.0293 0.0666 -0.0009 0.0116 -0.1033 0.0212 0.0515 0.0242 -0.0276 -0.0053 0.0099 -0.0594 0.0013 0.0396 0.0267 0.0253 0.0032 -0.0131 -0.0246 0.1145 0.1059 -0.0028 0.0820 -0.1168 -0.0846 -0.0212 0.0102 -0.0995 -0.1574 0.0130 -0.0796 0.1759 -0.1546 -0.2049 -0.0214 0.0502 -0.0050 -0.0719 -0.0108 0.0056 0.0272 0.0070 -0.0764 -0.0104 -0.0217 0.0591 0.0225 0.0245 -0.0171 -0.0107 0.0074 -0.0318 -0.0264 -0.0309 0.0143\n"
     ]
    }
   ],
   "source": [
    "!gzip -cd cc.en.300.vec.gz | head -n 20"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pMCadRAP0aey"
   },
   "source": [
    "Standard imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "H3llUwbv3JS-"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from urllib.request import urlopen"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_ICpYr9Fy9x4"
   },
   "source": [
    "Now for the class you are to implement in this assignment. *Tip: You don't need to load all the vectors for the embedding to be usefull.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "g1-78pxMAa_Q",
    "outputId": "04bfffd7-455b-47e2-db31-fe889c33d728"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Parsing word vectors: 100%|██████████| 100000/100000 [00:07<00:00, 13957.29it/s]\n"
     ]
    }
   ],
   "source": [
    "from tqdm import tqdm\n",
    "import os.path\n",
    "import gzip\n",
    "from sklearn.metrics.pairwise import cosine_similarity, euclidean_distances\n",
    "\n",
    "class Embedding(object):\n",
    "  def __init__(self, filename):\n",
    "    \"\"\"Embedding loader with methods for comparisons and searching\n",
    "    \n",
    "    filename:\n",
    "      The path of the chosen fasttext embedding data file.\"\"\"\n",
    "    assert os.path.exists(filename)\n",
    "    pass\n",
    "\n",
    "  def __len__(self):\n",
    "    \"\"\"Gives the number of tokens in the embedding.\"\"\"\n",
    "    pass\n",
    "\n",
    "  def __getitem__(self, tokens):\n",
    "    \"\"\"Returns the vectors for a token or a list of tokens.\"\"\"\n",
    "    assert type(tokens) == str or type(tokens) == list\n",
    "    pass\n",
    "\n",
    "  def __contains__(self, token):\n",
    "    \"\"\"Allows a user query if a tokens is in the embedding.\"\"\"\n",
    "    pass\n",
    "\n",
    "  def projection(self, direction, vectors):\n",
    "    \"\"\"The cosine similarity between a reference direction and a matrix of word vectors.\n",
    "    \n",
    "    direction:\n",
    "      A vector\n",
    "    vectors:\n",
    "      A row matrix with word vectors\"\"\"\n",
    "    pass\n",
    "\n",
    "  def simple_analogy(self, pair, query, k=5):\n",
    "    \"\"\"Simple/Classic word analogy as: man - woman = king - queen\"\"\"\n",
    "    pass\n",
    "\n",
    "  def similarity(self, token1, token2):\n",
    "    \"\"\"Returns the cosine similarity between two tokens.\"\"\"\n",
    "    pass\n",
    "\n",
    "  def analogies(self, analogy_direction, tokens):\n",
    "    \"\"\"The more complex word analogy method\"\"\"\n",
    "    pass\n",
    "\n",
    "embedding = Embedding(\"cc.en.300.vec.gz\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rEuxXfXilufb"
   },
   "source": [
    "Some unit tests that should work when using an embedding for english. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "sREGXFXQ9rwv"
   },
   "outputs": [],
   "source": [
    "assert 'man' in embedding\n",
    "assert 'woman' in embedding\n",
    "assert 'kdjfsfhgkdsjfdh' not in embedding\n",
    "\n",
    "assert len(embedding) > 10000\n",
    "\n",
    "assert type(embedding['man']) is np.ndarray\n",
    "assert np.isclose(np.sum(embedding['man']-embedding['man']), 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EX_7UbfKeLzH"
   },
   "source": [
    "## Word similarity\n",
    "\n",
    "We should be able to meassure semantic similarity using cosine similarity.\n",
    "\n",
    "$cos(\\overrightarrow{cat}, \\overrightarrow{dog}) \\geq cos(\\overrightarrow{cat}, \\overrightarrow{cow})$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "LwrYqvyYMUQq"
   },
   "outputs": [],
   "source": [
    "assert embedding.similarity('cat', 'dog') >= embedding.similarity('cat', 'cow')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TKsJ7LfELybJ"
   },
   "source": [
    "## Word analogies\n",
    "\n",
    "One of the core tasks for showing any semantic properties in an embedding is finding word analogies. We can this of this as word a is to word b, as word c is to word d. Our task is to find word d, knowing a, b, and c.\n",
    "\n",
    "$\\overrightarrow{a} - \\overrightarrow{b} \\approx \\overrightarrow{c} - \\overrightarrow{d}$\n",
    "\n",
    "For gender, this becomes:\n",
    "\n",
    "$\\overrightarrow{man} - \\overrightarrow{woman} \\approx \\overrightarrow{king} - \\overrightarrow{x}$\n",
    "\n",
    "By rearranging we get:\n",
    "\n",
    "$\\overrightarrow{king} - \\overrightarrow{man} + \\overrightarrow{woman} \\approx \\overrightarrow{x}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "CMtTEzuvHDls",
    "outputId": "a5e89b8d-be4b-4047-f4a6-e1a288426713"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Matches (in order of similarity):\n",
      "1 king\n",
      "2 queen\n",
      "3 kings\n",
      "4 Queen\n",
      "5 royal\n"
     ]
    }
   ],
   "source": [
    "best_matches = embedding.simple_analogy(('man', 'woman'), 'king', k=5)\n",
    "print(\"Matches (in order of similarity):\")\n",
    "for i, token in enumerate(best_matches):\n",
    "  print(i+1, token)\n",
    "\n",
    "assert 'queen' in embedding.simple_analogy(('man', 'woman'), 'king', k=5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "yY3FS9q5x9qr"
   },
   "source": [
    "We can do this for capitols too.\n",
    "\n",
    "$\\overrightarrow{France} - \\overrightarrow{Japan} \\approx \\overrightarrow{Paris} - \\overrightarrow{x}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "xL3dPVqKx9wa",
    "outputId": "167a6396-ddd7-4020-e0b9-2bfd66f3e1e2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Matches (in order of similarity):\n",
      "1 japan\n",
      "2 tokyo\n",
      "3 paris\n",
      "4 japanese\n",
      "5 shanghai\n"
     ]
    }
   ],
   "source": [
    "best_matches = embedding.simple_analogy(('france', 'japan'), 'paris', k=5)\n",
    "print(\"Matches (in order of similarity):\")\n",
    "for i, token in enumerate(best_matches):\n",
    "  print(i+1, token)\n",
    "\n",
    "assert 'tokyo' in embedding.simple_analogy(('france', 'japan'), 'paris', k=5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X1EknjnwpI2H"
   },
   "source": [
    "Please try doing this for at least one other famous category (e.g. comparing adjectives)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XgPGM1cvzIIo"
   },
   "source": [
    "## Finding the genderedness of professions\n",
    "\n",
    "Bulokbasi (2016) has a repo for the paper. We can download their data for professions. This allows us to find the genderedness for our embedding and plot against their survey data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ABmxQ1Fu56fr",
    "outputId": "7a8279c4-caac-4f07-98bd-0dc57ac1cdc6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 320 professions\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "with urlopen('https://raw.githubusercontent.com/tolga-b/debiaswe/master/data/professions.json') as remote_file:\n",
    "  professions = {title: {'definitional': defi, 'stereotypical': ster} for title, defi, ster in json.load(remote_file)}\n",
    "print(\"Loaded %i professions\" % len(professions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1mbuzbDNpygo"
   },
   "source": [
    "We want to find the projection of vectors for different profession on a gender direction. This can be done by finding the cosine similarity as:\n",
    "\n",
    "$cos(\\overrightarrow{man} - \\overrightarrow{woman}, \\overrightarrow{butcher})$\n",
    "\n",
    "$cos(\\overrightarrow{man} - \\overrightarrow{woman}, \\overrightarrow{ballerina})$\n",
    "\n",
    "The projection should be non-negative as $|cos(\\cdot,\\cdot)|$. However, by letting the projection method of your embedding class give us the cosine similarity, we can see the direction of the projection by looking at its sign."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uXVOOb8UEOsG"
   },
   "outputs": [],
   "source": [
    "assert type(embedding[['man', 'woman', 'butcher']]) is np.ndarray\n",
    "assert embedding[['man', 'woman', 'butcher']].shape[0] == 3\n",
    "\n",
    "assert embedding.projection(embedding['man']-embedding['woman'], embedding['butcher']) > 0\n",
    "assert embedding.projection(embedding['man']-embedding['woman'], embedding['ballerina']) < 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hvWMsRHWrZ-G"
   },
   "source": [
    "We create a woman-man direction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "o7qVcJM5_h7Y"
   },
   "outputs": [],
   "source": [
    "direction = embedding['man']-embedding['woman']\n",
    "titles = [title for title in professions.keys() if title in embedding]\n",
    "projection = embedding.projection(direction, embedding[titles])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "I4ld8hekrdEm"
   },
   "source": [
    "The estimation of the gender direction can be improved by using several pairs. Again, we use the data from Bulokbasi (2016)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "unzkpmHOs6Wi",
    "outputId": "c413166d-9d3a-4cfd-9dae-0bcf0f42c797"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('woman', 'man'),\n",
       " ('girl', 'boy'),\n",
       " ('she', 'he'),\n",
       " ('mother', 'father'),\n",
       " ('daughter', 'son'),\n",
       " ('gal', 'guy'),\n",
       " ('female', 'male'),\n",
       " ('her', 'his'),\n",
       " ('herself', 'himself'),\n",
       " ('Mary', 'John')]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import json\n",
    "with urlopen('https://raw.githubusercontent.com/tolga-b/debiaswe/master/data/definitional_pairs.json') as remote_file:\n",
    "  definitional_pairs = [(m, f) for m, f in json.load(remote_file)]\n",
    "definitional_pairs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "sueBKKyFGduH"
   },
   "outputs": [],
   "source": [
    "direction2 = np.zeros(embedding['man'].shape)\n",
    "for female, male in definitional_pairs:\n",
    "  direction2 += embedding[male.lower()]\n",
    "  direction2 -= embedding[female.lower()]\n",
    "\n",
    "projection2 = embedding.projection(direction2, embedding[titles])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DrRCBj3dr6m_"
   },
   "source": [
    "We can plot the relation between genderedness in the data and human annotation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 381
    },
    "id": "qxgJnULiG0_Q",
    "outputId": "5a61f6f2-7236-4f22-a010-e7068b729b53"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "fig = plt.figure(figsize=(12, 5))\n",
    "ax = fig.subplots(1, 2)\n",
    "stereotypicality = [professions[title]['stereotypical'] for title in titles]\n",
    "sns.regplot(x=projection, y=stereotypicality, ax=ax[0])\n",
    "a = sns.regplot(x=projection2, y=stereotypicality, ax=ax[1])\n",
    "ax[0].set_xlabel(\"Projection (one pair)\")\n",
    "ax[1].set_xlabel(\"Projection (several pairs)\")\n",
    "ax[0].set_ylabel(\"Human labelled stereotypicality\")\n",
    "fig.tight_layout(pad=.5)\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 381
    },
    "id": "Rb4MO33THKEN",
    "outputId": "b3fd543b-e5f9-4b74-a6c5-62ec9c462eaf"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(12, 5))\n",
    "ax = fig.subplots(1, 2)\n",
    "definitionality = [professions[title]['definitional'] for title in titles]\n",
    "sns.regplot(x=projection, y=definitionality, ax=ax[0])\n",
    "a = sns.regplot(x=projection2, y=definitionality, ax=ax[1])\n",
    "ax[0].set_xlabel(\"Projection (one pair)\")\n",
    "ax[1].set_xlabel(\"Projection (several pairs)\")\n",
    "ax[0].set_ylabel(\"Human labelled definitional genderedness\")\n",
    "fig.tight_layout(pad=.5)\n",
    "fig.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "cjgbqFix9A6Z"
   },
   "source": [
    "## More analogies\n",
    "\n",
    "The paper defines a more advances procedure for analogies as:\n",
    "\n",
    "$S_{(a,b)}(x, y) = \\left\\{\\begin{matrix}\n",
    "cos(\\overrightarrow{a}-\\overrightarrow{b}, \\overrightarrow{x}-\\overrightarrow{y}) & \\parallel \\overrightarrow{x}-\\overrightarrow{y} \\parallel \\leq 1\\\\ \n",
    "0 &  \\text{otherwise}\n",
    "\\end{matrix}\\right.$\n",
    "\n",
    "This should return a list of tuples with word pairs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "HY3EsrrTvxFx",
    "outputId": "a6e2394a-b1cb-4da3-dbac-fb9fe9886ace"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('actor', 'actress')]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "assert len(embedding.analogies(embedding['man']-embedding['woman'], ['actor', 'actress'])) == 1\n",
    "assert type(embedding.analogies(embedding['man']-embedding['woman'], ['actor', 'actress'])) == list\n",
    "assert type(embedding.analogies(embedding['man']-embedding['woman'], ['actor', 'actress'])[0]) == tuple\n",
    "\n",
    "embedding.analogies(embedding['man']-embedding['woman'], ['actor', 'actress'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "VK9et5C29AYD",
    "outputId": "461cf0e7-6ae7-4cd2-9484-02da54362ce3"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('accountant', 'businesswoman'),\n",
       " ('acquaintance', 'businesswoman'),\n",
       " ('actor', 'actress'),\n",
       " ('administrator', 'actress'),\n",
       " ('adventurer', 'businesswoman'),\n",
       " ('advocate', 'actress'),\n",
       " ('alderman', 'businesswoman'),\n",
       " ('ambassador', 'actress'),\n",
       " ('analyst', 'businesswoman'),\n",
       " ('anthropologist', 'actress'),\n",
       " ('archaeologist', 'actress'),\n",
       " ('archbishop', 'businesswoman'),\n",
       " ('architect', 'businesswoman'),\n",
       " ('artist', 'actress'),\n",
       " ('artiste', 'actress'),\n",
       " ('assassin', 'businesswoman'),\n",
       " ('astronaut', 'businesswoman'),\n",
       " ('astronomer', 'actress'),\n",
       " ('athlete', 'businesswoman'),\n",
       " ('attorney', 'paralegal'),\n",
       " ('author', 'businesswoman'),\n",
       " ('ballerina', 'actress'),\n",
       " ('banker', 'businesswoman'),\n",
       " ('barber', 'hairdresser'),\n",
       " ('barrister', 'actress'),\n",
       " ('bartender', 'waitress'),\n",
       " ('biologist', 'businesswoman'),\n",
       " ('bishop', 'parishioner'),\n",
       " ('bodyguard', 'actress'),\n",
       " ('bookkeeper', 'homemaker'),\n",
       " ('broadcaster', 'actress'),\n",
       " ('broker', 'consultant'),\n",
       " ('bureaucrat', 'businesswoman'),\n",
       " ('businessman', 'businesswoman'),\n",
       " ('businesswoman', 'actress'),\n",
       " ('butcher', 'businesswoman'),\n",
       " ('butler', 'housekeeper'),\n",
       " ('cabbie', 'bartender'),\n",
       " ('cameraman', 'actress'),\n",
       " ('campaigner', 'actress'),\n",
       " ('captain', 'businesswoman'),\n",
       " ('cardiologist', 'actress'),\n",
       " ('caretaker', 'businesswoman'),\n",
       " ('carpenter', 'homemaker'),\n",
       " ('cartoonist', 'actress'),\n",
       " ('cellist', 'ballerina'),\n",
       " ('chancellor', 'businesswoman'),\n",
       " ('chaplain', 'homemaker'),\n",
       " ('character', 'actress'),\n",
       " ('chemist', 'businesswoman'),\n",
       " ('choreographer', 'actress'),\n",
       " ('cinematographer', 'actress'),\n",
       " ('citizen', 'businesswoman'),\n",
       " ('cleric', 'archbishop'),\n",
       " ('collector', 'businesswoman'),\n",
       " ('colonel', 'businesswoman'),\n",
       " ('columnist', 'actress'),\n",
       " ('comedian', 'actress'),\n",
       " ('commander', 'businesswoman'),\n",
       " ('commentator', 'actress'),\n",
       " ('commissioner', 'businesswoman'),\n",
       " ('composer', 'actress'),\n",
       " ('conductor', 'businesswoman'),\n",
       " ('confesses', 'businesswoman'),\n",
       " ('congressman', 'businesswoman'),\n",
       " ('constable', 'businesswoman'),\n",
       " ('consultant', 'actress'),\n",
       " ('correspondent', 'actress'),\n",
       " ('councilman', 'businesswoman'),\n",
       " ('councilor', 'businesswoman'),\n",
       " ('counselor', 'businesswoman'),\n",
       " ('critic', 'choreographer'),\n",
       " ('crooner', 'businesswoman'),\n",
       " ('crusader', 'businesswoman'),\n",
       " ('curator', 'businesswoman'),\n",
       " ('custodian', 'businesswoman'),\n",
       " ('dancer', 'actress'),\n",
       " ('dentist', 'receptionist'),\n",
       " ('deputy', 'businesswoman'),\n",
       " ('dermatologist', 'actress'),\n",
       " ('detective', 'actress'),\n",
       " ('diplomat', 'actress'),\n",
       " ('director', 'actress'),\n",
       " ('doctor', 'nurse'),\n",
       " ('drummer', 'vocalist'),\n",
       " ('economist', 'businesswoman'),\n",
       " ('editor', 'businesswoman'),\n",
       " ('educator', 'actress'),\n",
       " ('electrician', 'businesswoman'),\n",
       " ('employee', 'businesswoman'),\n",
       " ('entertainer', 'actress'),\n",
       " ('entrepreneur', 'businesswoman'),\n",
       " ('environmentalist', 'actress'),\n",
       " ('epidemiologist', 'actress'),\n",
       " ('evangelist', 'businesswoman'),\n",
       " ('farmer', 'businesswoman'),\n",
       " ('filmmaker', 'actress'),\n",
       " ('financier', 'businesswoman'),\n",
       " ('firefighter', 'businesswoman'),\n",
       " ('fireman', 'homemaker'),\n",
       " ('fisherman', 'businesswoman'),\n",
       " ('footballer', 'actress'),\n",
       " ('foreman', 'businesswoman'),\n",
       " ('gangster', 'businesswoman'),\n",
       " ('gardener', 'homemaker'),\n",
       " ('geologist', 'businesswoman'),\n",
       " ('goalkeeper', 'businesswoman'),\n",
       " ('guitarist', 'actress'),\n",
       " ('hairdresser', 'actress'),\n",
       " ('handyman', 'homemaker'),\n",
       " ('headmaster', 'businesswoman'),\n",
       " ('historian', 'actress'),\n",
       " ('hitman', 'assassin'),\n",
       " ('homemaker', 'actress'),\n",
       " ('housekeeper', 'actress'),\n",
       " ('housewife', 'actress'),\n",
       " ('illustrator', 'actress'),\n",
       " ('industrialist', 'businesswoman'),\n",
       " ('infielder', 'businesswoman'),\n",
       " ('inspector', 'businesswoman'),\n",
       " ('instructor', 'actress'),\n",
       " ('inventor', 'businesswoman'),\n",
       " ('investigator', 'actress'),\n",
       " ('janitor', 'homemaker'),\n",
       " ('jeweler', 'businesswoman'),\n",
       " ('journalist', 'actress'),\n",
       " ('judge', 'prosecutor'),\n",
       " ('jurist', 'barrister'),\n",
       " ('laborer', 'homemaker'),\n",
       " ('landlord', 'businesswoman'),\n",
       " ('lawmaker', 'businesswoman'),\n",
       " ('lawyer', 'paralegal'),\n",
       " ('lecturer', 'businesswoman'),\n",
       " ('legislator', 'businesswoman'),\n",
       " ('librarian', 'actress'),\n",
       " ('lieutenant', 'businesswoman'),\n",
       " ('lifeguard', 'businesswoman'),\n",
       " ('lyricist', 'actress'),\n",
       " ('maestro', 'businesswoman'),\n",
       " ('magician', 'businesswoman'),\n",
       " ('magistrate', 'businesswoman'),\n",
       " ('manager', 'businesswoman'),\n",
       " ('marksman', 'businesswoman'),\n",
       " ('marshal', 'businesswoman'),\n",
       " ('mathematician', 'businesswoman'),\n",
       " ('mechanic', 'businesswoman'),\n",
       " ('mediator', 'businesswoman'),\n",
       " ('midfielder', 'businesswoman'),\n",
       " ('minister', 'businesswoman'),\n",
       " ('missionary', 'businesswoman'),\n",
       " ('mobster', 'socialite'),\n",
       " ('musician', 'actress'),\n",
       " ('narrator', 'actress'),\n",
       " ('naturalist', 'businesswoman'),\n",
       " ('negotiator', 'businesswoman'),\n",
       " ('neurologist', 'businesswoman'),\n",
       " ('neurosurgeon', 'businesswoman'),\n",
       " ('novelist', 'actress'),\n",
       " ('observer', 'businesswoman'),\n",
       " ('officer', 'businesswoman'),\n",
       " ('organist', 'homemaker'),\n",
       " ('painter', 'businesswoman'),\n",
       " ('paralegal', 'homemaker'),\n",
       " ('parishioner', 'actress'),\n",
       " ('parliamentarian', 'businesswoman'),\n",
       " ('pastor', 'educator'),\n",
       " ('pathologist', 'actress'),\n",
       " ('pediatrician', 'actress'),\n",
       " ('performer', 'actress'),\n",
       " ('pharmacist', 'businesswoman'),\n",
       " ('philanthropist', 'businesswoman'),\n",
       " ('philosopher', 'businesswoman'),\n",
       " ('photographer', 'actress'),\n",
       " ('photojournalist', 'actress'),\n",
       " ('physician', 'actress'),\n",
       " ('physicist', 'actress'),\n",
       " ('pianist', 'actress'),\n",
       " ('planner', 'homemaker'),\n",
       " ('playwright', 'actress'),\n",
       " ('plumber', 'housewife'),\n",
       " ('policeman', 'businesswoman'),\n",
       " ('politician', 'businesswoman'),\n",
       " ('pollster', 'businesswoman'),\n",
       " ('preacher', 'businesswoman'),\n",
       " ('president', 'businesswoman'),\n",
       " ('priest', 'housekeeper'),\n",
       " ('principal', 'actress'),\n",
       " ('prisoner', 'businesswoman'),\n",
       " ('professor', 'actress'),\n",
       " ('programmer', 'actress'),\n",
       " ('promoter', 'businesswoman'),\n",
       " ('proprietor', 'businesswoman'),\n",
       " ('prosecutor', 'actress'),\n",
       " ('protagonist', 'actress'),\n",
       " ('protege', 'businesswoman'),\n",
       " ('protester', 'businesswoman'),\n",
       " ('provost', 'librarian'),\n",
       " ('psychiatrist', 'actress'),\n",
       " ('psychologist', 'actress'),\n",
       " ('publicist', 'actress'),\n",
       " ('pundit', 'columnist'),\n",
       " ('radiologist', 'actress'),\n",
       " ('ranger', 'naturalist'),\n",
       " ('realtor', 'businesswoman'),\n",
       " ('receptionist', 'actress'),\n",
       " ('researcher', 'actress'),\n",
       " ('restaurateur', 'businesswoman'),\n",
       " ('sailor', 'adventurer'),\n",
       " ('salesman', 'businesswoman'),\n",
       " ('saxophonist', 'actress'),\n",
       " ('scholar', 'businesswoman'),\n",
       " ('scientist', 'actress'),\n",
       " ('screenwriter', 'actress'),\n",
       " ('sculptor', 'actress'),\n",
       " ('secretary', 'actress'),\n",
       " ('senator', 'businesswoman'),\n",
       " ('sergeant', 'businesswoman'),\n",
       " ('servant', 'businesswoman'),\n",
       " ('serviceman', 'actress'),\n",
       " ('shopkeeper', 'businesswoman'),\n",
       " ('singer', 'actress'),\n",
       " ('skipper', 'businesswoman'),\n",
       " ('socialite', 'actress'),\n",
       " ('sociologist', 'actress'),\n",
       " ('soldier', 'businesswoman'),\n",
       " ('solicitor', 'paralegal'),\n",
       " ('soloist', 'ballerina'),\n",
       " ('sportsman', 'businesswoman'),\n",
       " ('sportswriter', 'actress'),\n",
       " ('statesman', 'businesswoman'),\n",
       " ('steward', 'businesswoman'),\n",
       " ('stockbroker', 'businesswoman'),\n",
       " ('strategist', 'businesswoman'),\n",
       " ('student', 'businesswoman'),\n",
       " ('stylist', 'businesswoman'),\n",
       " ('substitute', 'businesswoman'),\n",
       " ('superintendent', 'businesswoman'),\n",
       " ('surgeon', 'businesswoman'),\n",
       " ('surveyor', 'businesswoman'),\n",
       " ('teacher', 'businesswoman'),\n",
       " ('technician', 'actress'),\n",
       " ('teenager', 'actress'),\n",
       " ('therapist', 'actress'),\n",
       " ('trader', 'entrepreneur'),\n",
       " ('treasurer', 'businesswoman'),\n",
       " ('trooper', 'businesswoman'),\n",
       " ('trucker', 'homemaker'),\n",
       " ('trumpeter', 'businesswoman'),\n",
       " ('tycoon', 'industrialist'),\n",
       " ('understudy', 'actress'),\n",
       " ('valedictorian', 'actress'),\n",
       " ('violinist', 'actress'),\n",
       " ('vocalist', 'actress'),\n",
       " ('waiter', 'waitress'),\n",
       " ('waitress', 'actress'),\n",
       " ('warden', 'superintendent'),\n",
       " ('warrior', 'businesswoman'),\n",
       " ('welder', 'electrician'),\n",
       " ('worker', 'homemaker'),\n",
       " ('wrestler', 'businesswoman'),\n",
       " ('writer', 'actress')]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding.analogies(analogy_direction=direction2, tokens=titles)"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyOsxh3VYsE0zEZginQzHrok",
   "collapsed_sections": [],
   "name": "Assignment 3 - Gendered Directions in Embeddings.ipynb",
   "provenance": []
  },
  "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": "3.10.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}