BERT_word_list_CN_paper.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "BERT_word_list_CN.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "include_colab_link": true
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.4"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jsedoc/ConceptorDebias/blob/master/BERT_word_list_CN_paper.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "C5UBFar7so29"
      },
      "source": [
        "# Setup BERT"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "GbLgfNRusWBt",
        "outputId": "1dea4e70-f592-4b6e-de1e-bfa619b533ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "!pip install flair"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting flair\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/4e/3a/2e777f65a71c1eaa259df44c44e39d7071ba8c7780a1564316a38bf86449/flair-0.4.2-py3-none-any.whl (136kB)\n",
            "\u001b[K     |████████████████████████████████| 143kB 4.9MB/s \n",
            "\u001b[?25hCollecting sqlitedict>=1.6.0 (from flair)\n",
            "  Downloading https://files.pythonhosted.org/packages/0f/1c/c757b93147a219cf1e25cef7e1ad9b595b7f802159493c45ce116521caff/sqlitedict-1.6.0.tar.gz\n",
            "Requirement already satisfied: pytest>=3.6.4 in /usr/local/lib/python3.6/dist-packages (from flair) (3.6.4)\n",
            "Collecting pytorch-pretrained-bert>=0.6.1 (from flair)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/d7/e0/c08d5553b89973d9a240605b9c12404bcf8227590de62bae27acbcfe076b/pytorch_pretrained_bert-0.6.2-py3-none-any.whl (123kB)\n",
            "\u001b[K     |████████████████████████████████| 133kB 38.5MB/s \n",
            "\u001b[?25hRequirement already satisfied: torch>=1.0.0 in /usr/local/lib/python3.6/dist-packages (from flair) (1.1.0)\n",
            "Collecting mpld3==0.3 (from flair)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/91/95/a52d3a83d0a29ba0d6898f6727e9858fe7a43f6c2ce81a5fe7e05f0f4912/mpld3-0.3.tar.gz (788kB)\n",
            "\u001b[K     |████████████████████████████████| 798kB 52.3MB/s \n",
            "\u001b[?25hCollecting segtok>=1.5.7 (from flair)\n",
            "  Downloading https://files.pythonhosted.org/packages/1d/59/6ed78856ab99d2da04084b59e7da797972baa0efecb71546b16d48e49d9b/segtok-1.5.7.tar.gz\n",
            "Requirement already satisfied: hyperopt>=0.1.1 in /usr/local/lib/python3.6/dist-packages (from flair) (0.1.2)\n",
            "Requirement already satisfied: tqdm>=4.26.0 in /usr/local/lib/python3.6/dist-packages (from flair) (4.28.1)\n",
            "Collecting deprecated>=1.2.4 (from flair)\n",
            "  Downloading https://files.pythonhosted.org/packages/88/0e/9d5a1a8cd7130c49334cce7b8167ceda63d6a329c8ea65b626116bc9e9e6/Deprecated-1.2.6-py2.py3-none-any.whl\n",
            "Collecting bpemb>=0.2.9 (from flair)\n",
            "  Downloading https://files.pythonhosted.org/packages/bc/70/468a9652095b370f797ed37ff77e742b11565c6fd79eaeca5f2e50b164a7/bpemb-0.3.0-py3-none-any.whl\n",
            "Collecting regex (from flair)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/6f/4e/1b178c38c9a1a184288f72065a65ca01f3154df43c6ad898624149b8b4e0/regex-2019.06.08.tar.gz (651kB)\n",
            "\u001b[K     |████████████████████████████████| 655kB 45.9MB/s \n",
            "\u001b[?25hRequirement already satisfied: sklearn in /usr/local/lib/python3.6/dist-packages (from flair) (0.0)\n",
            "Requirement already satisfied: urllib3<1.25,>=1.20 in /usr/local/lib/python3.6/dist-packages (from flair) (1.24.3)\n",
            "Requirement already satisfied: gensim>=3.4.0 in /usr/local/lib/python3.6/dist-packages (from flair) (3.6.0)\n",
            "Requirement already satisfied: matplotlib>=2.2.3 in /usr/local/lib/python3.6/dist-packages (from flair) (3.0.3)\n",
            "Requirement already satisfied: tabulate in /usr/local/lib/python3.6/dist-packages (from flair) (0.8.3)\n",
            "Requirement already satisfied: attrs>=17.4.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (19.1.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (41.0.1)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (1.12.0)\n",
            "Requirement already satisfied: py>=1.5.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (1.8.0)\n",
            "Requirement already satisfied: atomicwrites>=1.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (1.3.0)\n",
            "Requirement already satisfied: more-itertools>=4.0.0 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (7.1.0)\n",
            "Requirement already satisfied: pluggy<0.8,>=0.5 in /usr/local/lib/python3.6/dist-packages (from pytest>=3.6.4->flair) (0.7.1)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from pytorch-pretrained-bert>=0.6.1->flair) (2.21.0)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from pytorch-pretrained-bert>=0.6.1->flair) (1.16.4)\n",
            "Requirement already satisfied: boto3 in /usr/local/lib/python3.6/dist-packages (from pytorch-pretrained-bert>=0.6.1->flair) (1.9.189)\n",
            "Requirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair) (0.16.0)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair) (2.3)\n",
            "Requirement already satisfied: pymongo in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair) (3.8.0)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.6/dist-packages (from hyperopt>=0.1.1->flair) (1.3.0)\n",
            "Requirement already satisfied: wrapt<2,>=1.10 in /usr/local/lib/python3.6/dist-packages (from deprecated>=1.2.4->flair) (1.11.2)\n",
            "Collecting sentencepiece (from bpemb>=0.2.9->flair)\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/00/95/7f357995d5eb1131aa2092096dca14a6fc1b1d2860bd99c22a612e1d1019/sentencepiece-0.1.82-cp36-cp36m-manylinux1_x86_64.whl (1.0MB)\n",
            "\u001b[K     |████████████████████████████████| 1.0MB 36.2MB/s \n",
            "\u001b[?25hRequirement already satisfied: scikit-learn in /usr/local/lib/python3.6/dist-packages (from sklearn->flair) (0.21.2)\n",
            "Requirement already satisfied: smart-open>=1.2.1 in /usr/local/lib/python3.6/dist-packages (from gensim>=3.4.0->flair) (1.8.4)\n",
            "Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair) (2.5.3)\n",
            "Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair) (2.4.0)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair) (1.1.0)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=2.2.3->flair) (0.10.0)\n",
            "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair) (3.0.4)\n",
            "Requirement already satisfied: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair) (2.8)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->pytorch-pretrained-bert>=0.6.1->flair) (2019.6.16)\n",
            "Requirement already satisfied: s3transfer<0.3.0,>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair) (0.2.1)\n",
            "Requirement already satisfied: jmespath<1.0.0,>=0.7.1 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair) (0.9.4)\n",
            "Requirement already satisfied: botocore<1.13.0,>=1.12.189 in /usr/local/lib/python3.6/dist-packages (from boto3->pytorch-pretrained-bert>=0.6.1->flair) (1.12.189)\n",
            "Requirement already satisfied: decorator>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from networkx->hyperopt>=0.1.1->flair) (4.4.0)\n",
            "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.6/dist-packages (from scikit-learn->sklearn->flair) (0.13.2)\n",
            "Requirement already satisfied: boto>=2.32 in /usr/local/lib/python3.6/dist-packages (from smart-open>=1.2.1->gensim>=3.4.0->flair) (2.49.0)\n",
            "Requirement already satisfied: docutils>=0.10 in /usr/local/lib/python3.6/dist-packages (from botocore<1.13.0,>=1.12.189->boto3->pytorch-pretrained-bert>=0.6.1->flair) (0.14)\n",
            "Building wheels for collected packages: sqlitedict, mpld3, segtok, regex\n",
            "  Building wheel for sqlitedict (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Stored in directory: /root/.cache/pip/wheels/bd/57/d3/907c3ee02d35e66f674ad0106e61f06eeeb98f6ee66a6cc3fe\n",
            "  Building wheel for mpld3 (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Stored in directory: /root/.cache/pip/wheels/c0/47/fb/8a64f89aecfe0059830479308ad42d62e898a3e3cefdf6ba28\n",
            "  Building wheel for segtok (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Stored in directory: /root/.cache/pip/wheels/15/ee/a8/6112173f1386d33eebedb3f73429cfa41a4c3084556bcee254\n",
            "  Building wheel for regex (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Stored in directory: /root/.cache/pip/wheels/35/e4/80/abf3b33ba89cf65cd262af8a22a5a999cc28fbfabea6b38473\n",
            "Successfully built sqlitedict mpld3 segtok regex\n",
            "Installing collected packages: sqlitedict, regex, pytorch-pretrained-bert, mpld3, segtok, deprecated, sentencepiece, bpemb, flair\n",
            "Successfully installed bpemb-0.3.0 deprecated-1.2.6 flair-0.4.2 mpld3-0.3 pytorch-pretrained-bert-0.6.2 regex-2019.6.8 segtok-1.5.7 sentencepiece-0.1.82 sqlitedict-1.6.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "sAxGdPUKsZB5",
        "outputId": "3fe861ac-0d68-4388-f686-48a4c4eb8531",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "from flair.embeddings import BertEmbeddings\n",
        "from flair.data import Sentence\n",
        "\n",
        "\n",
        "# init embedding\n",
        "embedding = BertEmbeddings('bert-large-cased', layers=\"20,21,22,23\")"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 231508/231508 [00:00<00:00, 1248277.29B/s]\n",
            "100%|██████████| 1248501532/1248501532 [00:32<00:00, 38509080.81B/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "-lJ5H2KrsL9f",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import torch\n",
        "t = np.transpose\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "j91LrkyBsu-R"
      },
      "source": [
        "# Get Brown Corpus"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "5GVVfJkDs0fh",
        "outputId": "ca7034f3-a549-441d-8138-2592b142c1e6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "from tqdm import tqdm\n",
        "\n",
        "import nltk\n",
        "nltk.download('brown')\n",
        "\n",
        "from nltk.corpus import brown\n",
        "brown_corpus = brown.sents()[:20000]"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package brown to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/brown.zip.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Uo4Qu3fntX5f",
        "outputId": "145b6484-1bbe-4dd9-f9aa-551885bfbf3a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "brown_e = []\n",
        "for s in tqdm(brown_corpus):\n",
        "    sentence = Sentence(' '.join(s))\n",
        "    embedding.embed(sentence)\n",
        "    sent_emb = torch.stack([token.embedding for token in sentence]).numpy()\n",
        "    brown_e.append(sent_emb)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  7%|▋         | 1499/20000 [01:18<14:03, 21.94it/s]"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Ml280Y7Qkah8",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "t = np.transpose\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "mghOzcVXklz3",
        "colab": {}
      },
      "source": [
        "import matplotlib as mpl\n",
        "from mpl_toolkits.mplot3d import Axes3D\n",
        "import numpy as np\n",
        "import matplotlib.cm as cm\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "from   sklearn.decomposition import PCA\n",
        "from   sklearn.manifold import TSNE, Isomap, LocallyLinearEmbedding, MDS, SpectralEmbedding\n",
        "from   sklearn.preprocessing import StandardScaler\n",
        "\n",
        "np.set_printoptions(suppress=True)\n",
        "np.set_printoptions(precision=4)\n",
        "plt_style = 'seaborn-talk'\n",
        "import seaborn as sns; sns.set()\n",
        "\n",
        "# Experimental: HDBScan is a state-of-the-art clustering algorithm\n",
        "hdbscan_available = True\n",
        "try:\n",
        "    import hdbscan\n",
        "except ImportError:\n",
        "    hdbscan_available = False"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "GXRIbtOmksRo",
        "colab": {}
      },
      "source": [
        "def do_plot(X_fit, title=None, labels=['']):\n",
        "    dimension = X_fit.shape[1]\n",
        "    label_types = sorted(list(set(labels)))\n",
        "    num_labels = len(label_types)\n",
        "    colors = cm.Accent(np.linspace(0, 1, num_labels))\n",
        "    print(X_fit.shape, label_types, num_labels, colors)\n",
        "    with plt.style.context(plt_style):\n",
        "        fig = plt.figure()\n",
        "        \n",
        "        if dimension == 2:\n",
        "            ax = fig.add_subplot(111)\n",
        "            for lab, col in zip(label_types, colors):\n",
        "                if num_labels>1:\n",
        "                    idxs = [i for i,v in enumerate(labels) if v == lab]\n",
        "                    import pdb; pdb.set_trace()\n",
        "                    df = pd.DataFrame([[X_fit[i, 0] for i in idxs], [X_fit[i, 1] for i in idxs], [lab for i in idxs]], columns=['X','Y','label'])\n",
        "                    if tdf != None:\n",
        "                        tdf.append(df)\n",
        "                    else:\n",
        "                        tdf = df\n",
        "                    #mt = [j for i,v in enumerate(labels) if v == lab]\n",
        "                    #markers = {\"Mary\": \"s\", \"John\": \"X\"}\n",
        "                    #ax = sns.scatterplot([X_fit[i, 0] for i in idxs],\n",
        "                    #           [X_fit[i, 1] for i in idxs],\n",
        "                    #           c=col, label=lab, style=pd.Series(col))\n",
        "                else:\n",
        "                    ax.scatter(X_fit[:, 0],\n",
        "                               X_fit[:, 1],\n",
        "                               c=col)\n",
        "        elif dimension == 3:\n",
        "            ax = fig.add_subplot(111, projection='3d')\n",
        "            for lab, col in zip(label_types, colors):\n",
        "                ax.scatter(X_fit[labels==lab, 0],\n",
        "                           X_fit[labels==lab, 1],\n",
        "                           X_fit[labels==lab, 2],\n",
        "                           c=col)\n",
        "        else:\n",
        "            raise Exception('Unknown dimension: %d' % dimension)\n",
        "        ax = sns.scatterplot(tdf.X,tdf.Y, hue=tdf.label, style=tdf.label)\n",
        "        plt.title(title)\n",
        "        if num_labels>1:\n",
        "            ax.legend()\n",
        "        plt.show()\n",
        "        return tdf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hBtX7HE8JGKp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_df(X_fit, labels):\n",
        "    tdf =''\n",
        "    label_types = sorted(list(set(labels)))\n",
        "    print(label_types)\n",
        "    for lab in label_types:\n",
        "        idxs = [i for i,v in enumerate(labels) if v == lab]\n",
        "        Z = [[X_fit[i, 0] for i in idxs], [X_fit[i, 1] for i in idxs], [lab for i in idxs]]\n",
        "        df = pd.DataFrame(np.array(Z).T, columns=['PC1','PC2','word'])\n",
        "        if type(tdf) != type(''):\n",
        "            tdf = tdf.append(df)\n",
        "        else:\n",
        "            tdf = df\n",
        "    return tdf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "90yn5fONJvEv",
        "colab_type": "text"
      },
      "source": [
        "# New Section"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jWID015cJwGa",
        "colab_type": "code",
        "outputId": "4524f970-737f-4e2f-beaf-aa3f01725152",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['he','she'])\n",
        "pca = PCA(n_components=2)\n",
        "pca.fit(X)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['executive','relatives','wedding','salary'])\n",
        "df = create_df(pca.transform(X), labels=labels)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['executive', 'relatives', 'salary', 'wedding']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "syfGRK5aLNm4",
        "colab_type": "code",
        "outputId": "e10e9096-b837-42ce-8fc7-dbe593658b89",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        }
      },
      "source": [
        "with plt.style.context(plt_style):\n",
        "    fig = plt.figure()\n",
        "    ax = sns.scatterplot(x=df.PC1.astype('float64'), y=df.PC2.astype('float64'), hue=df.word, style=df.word)\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAowAAAHHCAYAAAA1RAnkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nimJYrAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 748.8x514.8 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E_7B1L_keYoi",
        "colab_type": "code",
        "outputId": "8f709483-fb33-45bd-de6f-2937050d13d5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 596
        }
      },
      "source": [
        "AB, AB_labels, _ = pick_embeddings(brown_corpus, brown_embs, ['he', 'she'])\n",
        "XY, XY_labels, _ = pick_embeddings(brown_corpus, brown_embs, ['wedding', 'relatives','executive', 'salary'])#WEATLists.W_6_Career + WEATLists.W_6_Family)\n",
        "\n",
        "Z = np.matrix(AB)\n",
        "L = np.matrix(XY)\n",
        "GnegC, GnewX, GR = post_process_cn_matrix(Z.T)\n",
        "\n",
        "pca.fit(GnewX)\n",
        "df = create_df(pca.transform((GnegC*L.T).T), labels=labels)\n",
        "\n",
        "with plt.style.context(plt_style):\n",
        "    fig = plt.figure()\n",
        "    ax = sns.scatterplot(x=df.PC1.astype('float64'), y=df.PC2.astype('float64'), hue=df.word, style=df.word)\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "starting...\n",
            "(4096, 2088)\n",
            "R calculated\n",
            "C calculated\n",
            "negC calculated\n",
            "(2088, 4096)\n",
            "['executive', 'relatives', 'salary', 'wedding']\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqEAAAHHCAYAAACP54GEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
            "text/plain": [
              "<Figure size 748.8x514.8 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "2XB7QFvHkxWq",
        "colab": {}
      },
      "source": [
        "def pick_embeddings(corpus, sent_embs, word_list):\n",
        "    X = []\n",
        "    labels = []\n",
        "    sents = []\n",
        "    for i, s in enumerate(corpus):\n",
        "        for j, w in enumerate(s):\n",
        "            if w in word_list:\n",
        "                X.append(sent_embs[i][j])\n",
        "                labels.append(w)\n",
        "                sents.append(s)\n",
        "    return (X, labels, sents)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "4QGjzpJOlEni",
        "colab": {}
      },
      "source": [
        "#brown_corpus = brown_corpus[:20000]\n",
        "brown_embs = brown_e"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "797rJIVxk1wH",
        "outputId": "d600a5ce-f820-4280-9b22-7de460accec8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 8132
        }
      },
      "source": [
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['husband', 'wife', 'Mary', 'John'])\n",
        "pca = PCA(n_components=2)\n",
        "pca.fit(X)\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['man', 'woman'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['himself', 'herself'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['boy', 'girl'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['Mary', 'John'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['husband', 'wife'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)\n",
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['David', 'Lauren', 'Robert', 'Bonnie'])\n",
        "do_plot(pca.transform(X), 'PCA', labels=labels)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(386, 2) ['John', 'Mary', 'husband', 'wife'] 4 [[0.498  0.7882 0.498  1.    ]\n",
            " [0.9922 0.7529 0.5255 1.    ]\n",
            " [0.9412 0.0078 0.498  1.    ]\n",
            " [0.4    0.4    0.4    1.    ]]\n",
            "> <ipython-input-62-653900a4aa6b>(16)do_plot()\n",
            "-> df = pd.DataFrame([[X_fit[i, 0] for i in idxs], [X_fit[i, 1] for i in idxs], [lab for i in idxs]], columns=['X','Y','label'])\n",
            "(Pdb) [X_fit[i, 0] for i in idxs]\n",
            "*** NameError: name 'X_fit' is not defined\n",
            "(Pdb) p X_fit\n",
            "array([[ 18.0313,  -3.7581],\n",
            "       [ 19.0955,   2.4399],\n",
            "       [-12.7979,  -8.5769],\n",
            "       [-14.6823,  -8.9644],\n",
            "       [-14.6619,  -5.271 ],\n",
            "       [ -7.3371,  -9.4487],\n",
            "       [-14.8061,  -8.9884],\n",
            "       [-16.2526, -12.9182],\n",
            "       [-12.9015,  -1.8052],\n",
            "       [-10.2424,  -6.0034],\n",
            "       [ 20.4701,  -0.9063],\n",
            "       [-14.2221,   0.204 ],\n",
            "       [-10.9482,  -6.4139],\n",
            "       [ 16.5498,  -2.6594],\n",
            "       [ -9.5294,  -2.578 ],\n",
            "       [-11.9551,  -4.6669],\n",
            "       [-12.3487,  -7.1036],\n",
            "       [ -4.516 ,  -4.5759],\n",
            "       [ -8.1343,  -3.3622],\n",
            "       [ 18.1088,  -1.195 ],\n",
            "       [ -7.6131,  -3.3196],\n",
            "       [ 18.0675,  -1.6446],\n",
            "       [-12.8327,  -8.0355],\n",
            "       [-12.8716,  -8.1181],\n",
            "       [ 11.8843,  -2.8036],\n",
            "       [ 13.8737,  -3.0561],\n",
            "       [ 16.8647,  -1.9168],\n",
            "       [ 15.6406,  -3.4625],\n",
            "       [ 16.1739,  -1.6466],\n",
            "       [-12.7845,  -6.9315],\n",
            "       [-13.1548,  -7.5738],\n",
            "       [-14.834 , -10.6302],\n",
            "       [-11.2769,  -1.9256],\n",
            "       [ -8.9878,  -3.7292],\n",
            "       [-16.2687,  -7.656 ],\n",
            "       [-17.8634, -11.0676],\n",
            "       [ 15.5496,  -1.7593],\n",
            "       [ 12.2788,  -3.7692],\n",
            "       [-15.8078,  -6.3938],\n",
            "       [ -8.184 ,  -9.5494],\n",
            "       [-17.7809, -11.6282],\n",
            "       [-16.6978, -11.0285],\n",
            "       [-14.6042,  -3.3223],\n",
            "       [-16.7055,  -8.3594],\n",
            "       [-16.4565, -10.4068],\n",
            "       [ -9.1915,  -3.798 ],\n",
            "       [-18.2668,  -9.7305],\n",
            "       [-13.3344,  -9.5565],\n",
            "       [-13.6886,  -4.1997],\n",
            "       [ 18.7158,  -1.9051],\n",
            "       [ 17.6247,  -2.8287],\n",
            "       [ 13.5723,  -0.5717],\n",
            "       [ 19.4328,  -0.6114],\n",
            "       [-14.875 , -10.7645],\n",
            "       [ -8.2094,  -6.4377],\n",
            "       [-13.2526,  -9.8063],\n",
            "       [ 19.1146,  -1.2882],\n",
            "       [ 17.5439,  -1.4988],\n",
            "       [-15.7804,  -4.0984],\n",
            "       [-12.0553,  -4.7147],\n",
            "       [-13.9113,  -7.6418],\n",
            "       [ 17.5342,   0.8591],\n",
            "       [ 19.3589,  -1.9483],\n",
            "       [ 18.1252,   0.055 ],\n",
            "       [ -7.7449,  -5.5366],\n",
            "       [ 16.1779,  -1.6252],\n",
            "       [ 14.6368,  -3.9582],\n",
            "       [ 13.2238,  -3.7729],\n",
            "       [-12.7023, -11.1153],\n",
            "       [-14.8487,  -5.2454],\n",
            "       [-11.6043,  -5.7246],\n",
            "       [-12.0646,  -5.9769],\n",
            "       [-11.1572,  -8.3106],\n",
            "       [-14.9696, -10.8035],\n",
            "       [-16.5001,  -6.1421],\n",
            "       [-18.4644,  -3.7953],\n",
            "       [-15.0457,  -6.4682],\n",
            "       [ 18.9922,  -1.9673],\n",
            "       [-14.4105,  -8.1337],\n",
            "       [ -4.5695,  -5.7811],\n",
            "       [-15.0791,  -7.8566],\n",
            "       [-11.3036,  -6.7072],\n",
            "       [ 18.604 ,   1.7101],\n",
            "       [ 15.4896,  -0.3646],\n",
            "       [ 19.0106,   1.5385],\n",
            "       [ 16.3645,   0.6988],\n",
            "       [ 13.5209,  -4.3196],\n",
            "       [  9.2441,   1.3542],\n",
            "       [ 17.8978,  -3.2794],\n",
            "       [ 19.4285,  -0.1132],\n",
            "       [ 17.2656,  -1.9592],\n",
            "       [ 16.4454,   0.1873],\n",
            "       [ 17.1482,  -0.5502],\n",
            "       [ 16.0873,  -0.9442],\n",
            "       [ 14.8076,  -0.9677],\n",
            "       [ -7.8998,  -4.9054],\n",
            "       [-13.9991,  -8.4741],\n",
            "       [-15.2525,  -5.5349],\n",
            "       [-16.2825,  -4.5441],\n",
            "       [-13.0311,  -8.7678],\n",
            "       [ 16.1368,  -1.3582],\n",
            "       [ 15.2463,  -1.9045],\n",
            "       [ 19.9924,  -0.6987],\n",
            "       [-16.4256,  -6.0456],\n",
            "       [ 12.5669,  -2.0096],\n",
            "       [-11.5109,  -6.7996],\n",
            "       [-14.2295,  -0.0812],\n",
            "       [ -3.8164,  -1.9302],\n",
            "       [-12.5669,   2.4752],\n",
            "       [ 18.2724,  -0.9265],\n",
            "       [ -3.4939,  -0.1523],\n",
            "       [ -5.179 ,   2.56  ],\n",
            "       [-12.4913,   1.0361],\n",
            "       [ -4.572 ,   0.6464],\n",
            "       [ 15.5807,  -0.392 ],\n",
            "       [-12.2845, -10.4884],\n",
            "       [-12.1129,  -9.2084],\n",
            "       [-11.0234,  -2.8456],\n",
            "       [-13.3542,  -5.2464],\n",
            "       [ 15.753 ,  -2.2512],\n",
            "       [-14.0889,  -5.5951],\n",
            "       [ 13.3895,  -4.4021],\n",
            "       [-10.9797,  -4.5159],\n",
            "       [-12.1403,  -4.5494],\n",
            "       [-10.6075,  -5.912 ],\n",
            "       [ 17.2594,  -2.1393],\n",
            "       [ -6.0251,  -5.0598],\n",
            "       [ -9.0635,  -0.4305],\n",
            "       [ -9.8319,  -5.6064],\n",
            "       [-14.1572, -11.2256],\n",
            "       [-13.1253,  -6.1908],\n",
            "       [ -5.4138,   2.512 ],\n",
            "       [ -3.7425,   0.9837],\n",
            "       [ -6.0812,   3.942 ],\n",
            "       [ -4.7066,   2.472 ],\n",
            "       [-15.0809, -11.1623],\n",
            "       [-10.1866,  -4.4288],\n",
            "       [-11.1866,  -7.2679],\n",
            "       [ -9.7824,  -3.7339],\n",
            "       [ 15.4782,  -4.241 ],\n",
            "       [-12.1162,  16.6007],\n",
            "       [-12.1258,  19.2493],\n",
            "       [-12.0845,  18.0188],\n",
            "       [-11.7898,   7.2449],\n",
            "       [-11.9337,   8.6496],\n",
            "       [ -9.6777,  -3.8698],\n",
            "       [-12.9631,  -4.5381],\n",
            "       [-12.362 ,  -4.6517],\n",
            "       [ 18.8802,   0.4704],\n",
            "       [ 16.6572,  -0.5047],\n",
            "       [ -9.7435,  -3.6273],\n",
            "       [-11.5857,  -1.4175],\n",
            "       [ 12.965 ,  -0.9826],\n",
            "       [ 12.9519,  -0.837 ],\n",
            "       [ 11.9896,  -0.6012],\n",
            "       [  8.3136,  -1.8423],\n",
            "       [-11.722 ,   5.3982],\n",
            "       [-13.5947,  -5.8409],\n",
            "       [-14.9082,  -8.3243],\n",
            "       [-14.7843,  17.5565],\n",
            "       [-14.8148,  15.7933],\n",
            "       [-14.0067,  14.0959],\n",
            "       [-11.0585,  -2.6502],\n",
            "       [-12.4257,  -6.3769],\n",
            "       [ 18.211 ,  -0.8585],\n",
            "       [ 17.8805,  -0.3242],\n",
            "       [ -9.3991,  -5.4081],\n",
            "       [ 15.6962,  -0.9494],\n",
            "       [ 13.6506,  -3.3672],\n",
            "       [-13.0104,  -4.4741],\n",
            "       [-10.3937,  -4.7995],\n",
            "       [-12.8866,  -5.1086],\n",
            "       [-11.4749,  -1.9622],\n",
            "       [ 17.7622,   0.7267],\n",
            "       [ 12.548 ,  -0.5572],\n",
            "       [-14.6306,  -0.2712],\n",
            "       [ 12.1271,  -1.2321],\n",
            "       [ 15.4713,   1.5384],\n",
            "       [-10.9599,  -4.5591],\n",
            "       [ 18.0546,   1.0572],\n",
            "       [-14.507 ,  -2.9017],\n",
            "       [-11.9911,  -4.8306],\n",
            "       [ 14.8433,   0.9386],\n",
            "       [ 15.9948,   0.6773],\n",
            "       [ 16.4551,  -1.0069],\n",
            "       [ 14.5122,   0.4708],\n",
            "       [ 14.7987,   0.5169],\n",
            "       [ 17.2587,   1.4997],\n",
            "       [-12.2736,  -5.3373],\n",
            "       [-11.5044,  -4.8923],\n",
            "       [ 17.8445,   0.7846],\n",
            "       [-14.8645, -10.4237],\n",
            "       [-11.7934,  -5.1807],\n",
            "       [-11.1194,   1.1887],\n",
            "       [ -4.6697,  -4.7158],\n",
            "       [-10.5441,  -7.0362],\n",
            "       [-15.0328,  -4.4427],\n",
            "       [ -8.6377,  -3.4574],\n",
            "       [ 16.2762,   1.6224],\n",
            "       [ 14.1469,   0.4079],\n",
            "       [-12.1475,  -2.3288],\n",
            "       [ -4.2101,  10.4989],\n",
            "       [ 15.4075,   0.7356],\n",
            "       [ -4.8671,  11.5012],\n",
            "       [ -2.6789,   8.8373],\n",
            "       [-12.826 ,  20.8684],\n",
            "       [-12.6591,  24.1708],\n",
            "       [ -2.6221,  10.1825],\n",
            "       [-12.0853,  24.2363],\n",
            "       [ 18.6358,   1.7309],\n",
            "       [-17.45  ,  -4.4667],\n",
            "       [-10.5045,  -5.0952],\n",
            "       [-13.567 ,  28.1876],\n",
            "       [-13.6732,  27.7744],\n",
            "       [-13.6998,  27.8725],\n",
            "       [-13.2199,  25.3441],\n",
            "       [-12.9052,  26.5104],\n",
            "       [-13.5196,  28.3129],\n",
            "       [-12.9639,  27.2187],\n",
            "       [-13.2455,  26.7241],\n",
            "       [-13.6377,  28.5375],\n",
            "       [-12.7614,  26.7272],\n",
            "       [-13.6822,  28.5174],\n",
            "       [-13.9334,  28.4626],\n",
            "       [-13.1103,  24.5549],\n",
            "       [-13.661 ,  26.3415],\n",
            "       [-13.5325,  23.9553],\n",
            "       [-13.4975,  28.3892],\n",
            "       [-12.5854,  25.65  ],\n",
            "       [-13.2199,  25.3441],\n",
            "       [-13.2854,  25.7473],\n",
            "       [-12.6656,  26.3665],\n",
            "       [-11.0867,  -3.1766],\n",
            "       [-12.4881,   0.31  ],\n",
            "       [ -9.7183,  -2.0367],\n",
            "       [-13.4546,  19.2821],\n",
            "       [-16.2914,  -6.8325],\n",
            "       [ 10.0237,  -1.05  ],\n",
            "       [ -3.9227,  -3.1473],\n",
            "       [ -6.1122,   0.3211],\n",
            "       [-10.5293,  -2.8738],\n",
            "       [-16.3015,  -7.7317],\n",
            "       [ -3.0923,  10.0575],\n",
            "       [-13.8608,  -6.2689],\n",
            "       [ -9.6847,  -0.0113],\n",
            "       [-14.0893,  -5.3578],\n",
            "       [ 15.2779,   1.0303],\n",
            "       [ 17.3633,  -0.2504],\n",
            "       [ 18.6569,   1.1002],\n",
            "       [-12.5791,  -6.1852],\n",
            "       [ 18.9848,  -2.4084],\n",
            "       [ 15.8198,  -1.3986],\n",
            "       [-12.1356,  -5.8426],\n",
            "       [ 12.5756,   0.6695],\n",
            "       [ 16.0023,  -1.1754],\n",
            "       [-13.183 ,  -3.7998],\n",
            "       [-13.2625,  -2.5819],\n",
            "       [ -9.3871,  -1.9175],\n",
            "       [-10.5166,  -0.4632],\n",
            "       [-11.1729,  -4.3606],\n",
            "       [-10.5554,  -1.451 ],\n",
            "       [-12.2659,  -3.777 ],\n",
            "       [-10.4202,  -2.1682],\n",
            "       [-11.7405,  -3.6779],\n",
            "       [ -9.609 ,  -4.0338],\n",
            "       [-15.732 , -11.1046],\n",
            "       [-12.1582,   3.1451],\n",
            "       [-11.5616,   1.3061],\n",
            "       [-11.9575,   2.6831],\n",
            "       [-10.9665,   3.1952],\n",
            "       [ 15.4542,   1.3891],\n",
            "       [ 13.0793,   1.7831],\n",
            "       [ 16.9083,   2.5782],\n",
            "       [ 13.3819,   2.1745],\n",
            "       [ 13.3825,   2.5109],\n",
            "       [ 16.7445,   2.8696],\n",
            "       [ 14.5626,   2.1458],\n",
            "       [ 17.4443,   2.481 ],\n",
            "       [ 12.5532,   1.1829],\n",
            "       [ 17.336 ,   2.107 ],\n",
            "       [ 17.002 ,   2.561 ],\n",
            "       [ 15.6117,   1.5752],\n",
            "       [ 15.1063,  -0.1524],\n",
            "       [ 12.8033,   2.7242],\n",
            "       [ 15.7077,   3.4617],\n",
            "       [ 15.5269,   1.4982],\n",
            "       [ 18.1999,   2.7473],\n",
            "       [ 17.8984,   2.4177],\n",
            "       [ 14.1176,   1.0865],\n",
            "       [ 16.3608,   2.395 ],\n",
            "       [ 15.2986,   2.4561],\n",
            "       [ 16.3492,   1.7808],\n",
            "       [ -7.7241,  -4.723 ],\n",
            "       [ 16.024 ,   1.4401],\n",
            "       [ 13.3814,   1.8787],\n",
            "       [ 18.1438,   0.5113],\n",
            "       [ 12.0429,   2.943 ],\n",
            "       [ 16.5337,   3.8047],\n",
            "       [ 15.5399,   3.1652],\n",
            "       [ 10.6776,   1.4848],\n",
            "       [ 17.3849,   0.3567],\n",
            "       [ 17.3284,  -0.6762],\n",
            "       [ 13.994 ,  -1.445 ],\n",
            "       [ 18.2777,  -0.9935],\n",
            "       [ 14.0667,  -0.2114],\n",
            "       [ 16.8949,   0.2935],\n",
            "       [ 16.2895,  -0.8042],\n",
            "       [ 14.2273,  -0.9125],\n",
            "       [ 19.3331,   1.7312],\n",
            "       [ 12.1701,  -0.3059],\n",
            "       [ 17.3514,   0.8566],\n",
            "       [ 14.7128,   0.3455],\n",
            "       [ 15.3933,   1.0195],\n",
            "       [ 15.1109,   1.7939],\n",
            "       [ 15.4043,   1.8137],\n",
            "       [ 14.0404,   2.0285],\n",
            "       [ 14.1973,   1.2322],\n",
            "       [ 17.1569,   2.042 ],\n",
            "       [ 17.1582,   0.9188],\n",
            "       [ 15.8395,   1.8564],\n",
            "       [ 16.5629,   1.7976],\n",
            "       [ 16.5972,   0.4655],\n",
            "       [ 16.9064,   3.0372],\n",
            "       [ 15.1472,   2.2743],\n",
            "       [ 16.4455,   2.059 ],\n",
            "       [ 17.3285,   0.8834],\n",
            "       [ 15.1011,  -0.2471],\n",
            "       [ 14.8886,   1.4239],\n",
            "       [ 19.5129,  -0.0128],\n",
            "       [ 16.7089,   1.0475],\n",
            "       [ 17.8769,   0.8613],\n",
            "       [ 17.2894,   1.1615],\n",
            "       [ 15.7136,   2.2997],\n",
            "       [ 11.9474,   1.7589],\n",
            "       [ 13.6939,   1.8341],\n",
            "       [ -6.6263,  -1.3478],\n",
            "       [ 19.7547,  -1.8707],\n",
            "       [ 16.3602,  -1.1086],\n",
            "       [ 15.2759,  -2.418 ],\n",
            "       [ 14.094 ,   0.385 ],\n",
            "       [ 13.999 ,  -2.0116],\n",
            "       [-11.923 ,  -3.3692],\n",
            "       [-13.1681,   0.3806],\n",
            "       [-12.0828,  -6.2921],\n",
            "       [ -9.4556,  -4.5525],\n",
            "       [ 14.8361,  -1.1361],\n",
            "       [ 16.1067,  -0.8351],\n",
            "       [ 15.2699,  -2.4787],\n",
            "       [ 16.3822,  -2.3255],\n",
            "       [-13.2375,   0.9068],\n",
            "       [ -9.5981,  -2.2685],\n",
            "       [-15.0944,  -8.751 ],\n",
            "       [ 19.0318,  -1.0299],\n",
            "       [ 18.4414,  -0.2182],\n",
            "       [-12.1257,  -4.8169],\n",
            "       [-11.7848,  -1.2999],\n",
            "       [-12.5143,   3.3512],\n",
            "       [-11.3663,  -0.4479],\n",
            "       [-11.8666,   3.99  ],\n",
            "       [-11.4226,  -2.0027],\n",
            "       [-11.8736,  -5.0301],\n",
            "       [ 16.2637,  -0.501 ],\n",
            "       [ 16.5174,  -1.3498],\n",
            "       [ 14.1335,  -4.9287],\n",
            "       [ 17.7159,  -1.5797],\n",
            "       [-13.7242,  -6.4173],\n",
            "       [-15.3301,  -3.0685],\n",
            "       [-14.8046,  -3.3854],\n",
            "       [ -9.6073,  -7.7262],\n",
            "       [ -5.9843,  -3.1625],\n",
            "       [ -5.0924,  -7.5895],\n",
            "       [-11.179 ,   3.3632],\n",
            "       [-12.7568,   5.7194],\n",
            "       [-12.5212,   1.3995],\n",
            "       [-13.1895,   5.7297],\n",
            "       [-10.8524,   2.6684],\n",
            "       [-13.3113,   4.1692],\n",
            "       [ -8.8547,  -0.0259],\n",
            "       [-11.5776,   4.5308],\n",
            "       [-12.3004,   3.1295],\n",
            "       [ -4.8829,   3.8866],\n",
            "       [ 18.9682,  -0.9781],\n",
            "       [-15.0649, -12.251 ],\n",
            "       [-10.0471,  -5.4876],\n",
            "       [ -8.4016,  -5.1606],\n",
            "       [ 15.7593,  -0.9218]])\n",
            "(Pdb) [X_fit[i, 0] for i in idxs]\n",
            "*** NameError: name 'X_fit' is not defined\n",
            "(Pdb) n\n",
            "AssertionError: 3 columns passed, passed data had 190 columns\n",
            "> <ipython-input-62-653900a4aa6b>(16)do_plot()\n",
            "-> df = pd.DataFrame([[X_fit[i, 0] for i in idxs], [X_fit[i, 1] for i in idxs], [lab for i in idxs]], columns=['X','Y','label'])\n",
            "(Pdb) n\n",
            "--Return--\n",
            "> <ipython-input-62-653900a4aa6b>(16)do_plot()->None\n",
            "-> df = pd.DataFrame([[X_fit[i, 0] for i in idxs], [X_fit[i, 1] for i in idxs], [lab for i in idxs]], columns=['X','Y','label'])\n",
            "(Pdb) n\n",
            "AssertionError: 3 columns passed, passed data had 190 columns\n",
            "> <ipython-input-63-1ec8ff75b2a4>(4)<module>()\n",
            "-> do_plot(pca.transform(X), 'PCA', labels=labels)\n",
            "--KeyboardInterrupt--\n",
            "(Pdb) quit\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "BdbQuit",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mBdbQuit\u001b[0m                                   Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-63-1ec8ff75b2a4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mpca\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPCA\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn_components\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mpca\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m \u001b[0mdo_plot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpca\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'PCA'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m \u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpick_embeddings\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbrown_corpus\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbrown_embs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m'man'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'woman'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0mdo_plot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpca\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'PCA'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mlabels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/lib/python3.6/bdb.py\u001b[0m in \u001b[0;36mtrace_dispatch\u001b[0;34m(self, frame, event, arg)\u001b[0m\n\u001b[1;32m     55\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdispatch_return\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mframe\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     56\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mevent\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'exception'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 57\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdispatch_exception\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mframe\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     58\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mevent\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'c_call'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrace_dispatch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/lib/python3.6/bdb.py\u001b[0m in \u001b[0;36mdispatch_exception\u001b[0;34m(self, frame, arg)\u001b[0m\n\u001b[1;32m    111\u001b[0m                     and arg[0] is StopIteration and arg[2] is None):\n\u001b[1;32m    112\u001b[0m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muser_exception\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mframe\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 113\u001b[0;31m                 \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquitting\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mBdbQuit\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    114\u001b[0m         \u001b[0;31m# Stop at the StopIteration or GeneratorExit exception when the user\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    115\u001b[0m         \u001b[0;31m# has set stopframe in a generator by issuing a return command, or a\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mBdbQuit\u001b[0m: "
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAoAAAAG1CAYAAABpkkWgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 748.8x514.8 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "zkcoBZSlAt0M",
        "outputId": "62a562dd-6c68-4d84-a14e-a3ea1f3459a1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        }
      },
      "source": [
        "# General word list\n",
        "!wget https://raw.githubusercontent.com/IlyaSemenov/wikipedia-word-frequency/master/results/enwiki-20150602-words-frequency.txt\n",
        "!git clone https://github.com/PrincetonML/SIF\n",
        "    \n",
        "# Gender word lists\n",
        "!git clone https://github.com/uclanlp/gn_glove\n",
        "!git clone https://github.com/uclanlp/corefBias\n",
        "!wget https://www.cs.cmu.edu/Groups/AI/areas/nlp/corpora/names/female.txt\n",
        "!wget https://www.cs.cmu.edu/Groups/AI/areas/nlp/corpora/names/male.txt\n",
        "    \n",
        "# our code for debiasing -- also includes word lists    \n",
        "!git clone https://github.com/jsedoc/ConceptorDebias"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2019-06-10 22:11:36--  https://raw.githubusercontent.com/IlyaSemenov/wikipedia-word-frequency/master/results/enwiki-20150602-words-frequency.txt\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 151.101.0.133, 151.101.64.133, 151.101.128.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|151.101.0.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 404 Not Found\n",
            "2019-06-10 22:11:36 ERROR 404: Not Found.\n",
            "\n",
            "fatal: destination path 'SIF' already exists and is not an empty directory.\n",
            "fatal: destination path 'gn_glove' already exists and is not an empty directory.\n",
            "fatal: destination path 'corefBias' already exists and is not an empty directory.\n",
            "--2019-06-10 22:11:42--  https://www.cs.cmu.edu/Groups/AI/areas/nlp/corpora/names/female.txt\n",
            "Resolving www.cs.cmu.edu (www.cs.cmu.edu)... 128.2.42.95\n",
            "Connecting to www.cs.cmu.edu (www.cs.cmu.edu)|128.2.42.95|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 35751 (35K) [text/plain]\n",
            "Saving to: ‘female.txt.1’\n",
            "\n",
            "female.txt.1        100%[===================>]  34.91K   172KB/s    in 0.2s    \n",
            "\n",
            "2019-06-10 22:11:43 (172 KB/s) - ‘female.txt.1’ saved [35751/35751]\n",
            "\n",
            "--2019-06-10 22:11:44--  https://www.cs.cmu.edu/Groups/AI/areas/nlp/corpora/names/male.txt\n",
            "Resolving www.cs.cmu.edu (www.cs.cmu.edu)... 128.2.42.95\n",
            "Connecting to www.cs.cmu.edu (www.cs.cmu.edu)|128.2.42.95|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 20466 (20K) [text/plain]\n",
            "Saving to: ‘male.txt.1’\n",
            "\n",
            "male.txt.1          100%[===================>]  19.99K  --.-KB/s    in 0s      \n",
            "\n",
            "2019-06-10 22:11:45 (255 MB/s) - ‘male.txt.1’ saved [20466/20466]\n",
            "\n",
            "fatal: destination path 'ConceptorDebias' already exists and is not an empty directory.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "0CiP73winLx4",
        "colab": {}
      },
      "source": [
        "#from ConceptorDebias.lists.load_word_lists import *\n",
        "#from ConceptorDebias.plot_pc import *\n",
        "from ConceptorDebias.Conceptors.conceptor_fxns import *"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "3tqDFzb3B1zK",
        "colab": {}
      },
      "source": [
        "X, labels, _ = pick_embeddings(brown_corpus, brown_embs, ['husband', 'wife', 'Mary', 'John'])\n",
        "#negC, newX, R = post_process_cn_matrix(embeddings.T)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "bGISyKZVCHD7",
        "colab": {}
      },
      "source": [
        "from ConceptorDebias.load_word_lists import *\n",
        "\n",
        "gender_list_pronouns = WEATLists.W_7_Male_terms + WEATLists.W_7_Female_terms + WEATLists.W_8_Male_terms + WEATLists.W_8_Female_terms\n",
        "gender_list_pronouns = list(set(gender_list_pronouns))\n",
        "\n",
        "gender_list_extended = male_vino_extra + female_vino_extra + male_gnGlove + female_gnGlove\n",
        "gender_list_extended = list(set(gender_list_extended))\n",
        "\n",
        "gender_list_propernouns = male_cmu + female_cmu\n",
        "gender_list_propernouns = list(set(gender_list_propernouns))\n",
        "\n",
        "gender_list_all = gender_list_pronouns + gender_list_extended + gender_list_propernouns\n",
        "gender_list_all = list(set(gender_list_all))\n",
        "\n",
        "race_list = WEATLists.W_3_Unused_full_list_European_American_names + WEATLists.W_3_European_American_names + WEATLists.W_3_Unused_full_list_African_American_names + WEATLists.W_3_African_American_names + WEATLists.W_4_Unused_full_list_European_American_names + WEATLists.W_4_European_American_names + WEATLists.W_4_Unused_full_list_African_American_names + WEATLists.W_4_African_American_names + WEATLists.W_5_Unused_full_list_European_American_names + WEATLists.W_5_European_American_names + WEATLists.W_5_Unused_full_list_African_American_names + WEATLists.W_5_African_American_names\n",
        "race_list = list(set(race_list))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "OXkbatJyDP0w",
        "colab": {}
      },
      "source": [
        "AB, AB_labels, _ = pick_embeddings(brown_corpus, brown_embs, ['he', 'she'])\n",
        "XY, XY_labels, _ = pick_embeddings(brown_corpus, brown_embs, ['wedding', 'relatives','executive', 'salary', 'he', 'she'])#WEATLists.W_6_Career + WEATLists.W_6_Family)\n",
        "\n",
        "pca = PCA(n_components=2)\n",
        "pca.fit(AB)\n",
        "do_plot(pca.transform(XY), 'PCA', labels=XY_labels)\n",
        "\n",
        "\n",
        "Z = np.matrix(AB)\n",
        "L = np.matrix(XY)\n",
        "GnegC, GnewX, GR = post_process_cn_matrix(Z.T)\n",
        "\n",
        "pca.fit(GnewX)\n",
        "do_plot(pca.transform((GnegC*L.T).T), 'PCA', labels=XY_labels)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X48zGrTcPnYs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "print(WEATLists.W_6_Career)\n",
        "print(WEATLists.W_6_Family)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gyiNA3z8PnYv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pca = PCA(n_components=2)\n",
        "pca.fit(AB)\n",
        "do_plot(pca.transform(AB), 'PCA', labels=labels)\n",
        "\n",
        "#Z = np.matrix(AB)\n",
        "#GnegC, GnewX, GR = post_process_cn_matrix(Z.T)\n",
        "#pca.fit(GnewX)\n",
        "#do_plot(pca.transform(GnegC*XY), 'PCA', labels=labels)"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}