ixxi-dante/nw2vec

View on GitHub
projects/correctness/gae-reproduction-citeseer-nw2vec-multitask_arch-shallow_adj-ov=0.ipynb

Summary

Maintainability
Test Coverage
{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train nw2vec with the original VGAE architecture for Cora embeddings + a kernel for the scalar product decoding + an intermediate layer in decoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "# Train on CPU (hide GPU) due to memory constraints\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = \"\"\n",
    "\n",
    "import contextlib\n",
    "\n",
    "import numpy as np\n",
    "import keras\n",
    "from keras_tqdm import TQDMNotebookCallback as TQDMCallback\n",
    "\n",
    "from nw2vec import layers\n",
    "from nw2vec import ae\n",
    "from nw2vec import utils\n",
    "from nw2vec import batching\n",
    "import settings\n",
    "\n",
    "from gae.input_data import load_data\n",
    "from gae.preprocessing import sparse_to_tuple, mask_test_edges"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "@contextlib.contextmanager\n",
    "def gae_directory():\n",
    "    working_directory = os.path.abspath(os.curdir)\n",
    "    try:\n",
    "        # Move to the GAE directory\n",
    "        os.chdir('../../gae')\n",
    "        yield\n",
    "    finally:\n",
    "        # Move back\n",
    "        os.chdir(working_directory)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load data\n",
    "with gae_directory():\n",
    "    adj, features = load_data('citeseer')\n",
    "    features = features.toarray()\n",
    "\n",
    "#adj_train = mask_test_edges(adj)[0]\n",
    "adj_train, train_edges, val_edges, val_edges_false, test_edges, test_edges_false = mask_test_edges(adj)\n",
    "assert adj_train.diagonal().sum() == 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_p_builder(dims, feature_codec='SigmoidBernoulli', adj_kernel=None, use_bias=False,\n",
    "                    embedding_slices=None, with_l1=True):\n",
    "\n",
    "    # Validate arguments and set default values\n",
    "    assert feature_codec in ['SigmoidBernoulli', 'OrthogonalGaussian', 'SoftmaxMultinomial']\n",
    "    if embedding_slices is None:\n",
    "        adj_embedding_slice, v_embedding_slice = [slice(None), slice(None)]\n",
    "    else:\n",
    "        adj_embedding_slice, v_embedding_slice = embedding_slices\n",
    "\n",
    "    # Extract the dimensions we use.\n",
    "    dim_data, dim_l1, _, _ = dims\n",
    "\n",
    "    def p_builder(p_input):\n",
    "        # Get slices of the embeddings for each prediction\n",
    "        p_input_adj = layers.InnerSlice(adj_embedding_slice)(p_input)\n",
    "        p_input_v = layers.InnerSlice(v_embedding_slice)(p_input)\n",
    "        \n",
    "        p_penultimate_adj = p_input_adj\n",
    "        \n",
    "        if with_l1:\n",
    "            p_penultimate_v = keras.layers.Dense(\n",
    "                dim_l1, use_bias=use_bias, activation='relu',\n",
    "                kernel_regularizer='l2', bias_regularizer='l2',\n",
    "                name='p_layer1_v'\n",
    "            )(p_input_v)\n",
    "        else:\n",
    "            p_penultimate_v = p_input_v\n",
    "\n",
    "        # Prepare kwargs for the Bilinear adj decoder, then build it.\n",
    "        adj_kwargs = {}\n",
    "        if adj_kernel is not None:\n",
    "            adj_kwargs['fixed_kernel'] = adj_kernel\n",
    "        else:\n",
    "            adj_kwargs['kernel_regularizer'] = 'l2'\n",
    "        p_adj = layers.Bilinear(0, use_bias=use_bias, name='p_adj',\n",
    "                                bias_regularizer='l2',\n",
    "                                **adj_kwargs)([p_penultimate_adj, p_penultimate_adj])\n",
    "\n",
    "        # Finally build the feature decoder according to the requested codec.\n",
    "        if feature_codec in ['SigmoidBernoulli', 'SoftmaxMultinomial']:\n",
    "            p_v = keras.layers.Dense(dim_data, use_bias=use_bias,\n",
    "                                     kernel_regularizer='l2', bias_regularizer='l2',\n",
    "                                     name='p_v')(p_penultimate_v)\n",
    "        else:\n",
    "            assert feature_codec == 'OrthogonalGaussian'\n",
    "            p_v_μ_flat = keras.layers.Dense(dim_data, use_bias=use_bias,\n",
    "                                            kernel_regularizer='l2', bias_regularizer='l2',\n",
    "                                            name='p_v_mu_flat')(p_penultimate_v)\n",
    "            p_v_logS_flat = keras.layers.Dense(dim_data, use_bias=use_bias,\n",
    "                                               kernel_regularizer='l2', bias_regularizer='l2',\n",
    "                                               name='p_v_logS_flat')(p_penultimate_v)\n",
    "            p_v = keras.layers.Concatenate(name='p_v_mulogS_flat')([p_v_μ_flat, p_v_logS_flat])\n",
    "\n",
    "        return ([p_adj, p_v], ('SigmoidBernoulliScaledAdjacency', feature_codec))\n",
    "\n",
    "    return p_builder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_ξ_samples = 1\n",
    "n_nodes = adj_train.shape[0]\n",
    "dim_data, dim_l1, dim_ξ_adj, dim_ξ_v = features.shape[1], 32, 16, 16\n",
    "overlap = 0\n",
    "dims = (dim_data, dim_l1, dim_ξ_adj, dim_ξ_v)\n",
    "loss_weights = {\n",
    "    'q_mulogS_flat': 1e-2 * 1.0 / (dim_ξ_adj - overlap + dim_ξ_v),\n",
    "    'p_adj': 1.0 / (n_nodes * np.log(2)),\n",
    "    'p_v': 1.0 / (features.shape[1] * np.log(2)),\n",
    "}\n",
    "\n",
    "# Actual VAE\n",
    "q_model, q_codecs = ae.build_q(dims, overlap=overlap,\n",
    "                               fullbatcher=batching.fullbatches, minibatcher=batching.pq_batches)\n",
    "p_builder = build_p_builder(dims,\n",
    "                               feature_codec='SigmoidBernoulli',\n",
    "                               adj_kernel=np.eye(16),\n",
    "                               embedding_slices=[slice(dim_ξ_adj),\n",
    "                                                 slice(dim_ξ_adj - overlap, dim_ξ_adj - overlap + dim_ξ_v)],\n",
    "                               with_l1=True)\n",
    "vae, vae_codecs = ae.build_vae(\n",
    "    (q_model, q_codecs), p_builder,\n",
    "    n_ξ_samples,\n",
    "    loss_weights=loss_weights\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def target_func(batch_adj, required_nodes, final_nodes):\n",
    "    return [\n",
    "        np.zeros(1), # ignored\n",
    "        utils.expand_dims_tile(utils.expand_dims_tile(batch_adj + np.eye(batch_adj.shape[0]), 0, n_ξ_samples), 0, 1),\n",
    "        utils.expand_dims_tile(features[final_nodes], 1, n_ξ_samples),\n",
    "    ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/sl/.conda/envs/base36/lib/python3.6/site-packages/tensorflow/python/ops/gradients_impl.py:100: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory.\n",
      "  \"Converting sparse IndexedSlices to a dense Tensor of unknown shape. \"\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "195cefbddee44765a3f5165e48f56119",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(IntProgress(value=0, description='Training', max=200), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "n_epochs = 200\n",
    "\n",
    "history = vae.fit_fullbatches(batcher_kws={'adj': adj_train, 'features': features, 'target_func': target_func},\n",
    "                              epochs=n_epochs,\n",
    "                              verbose=0, callbacks=[TQDMCallback(show_inner=False)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "history = {'history': history.history}"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axes = plt.subplots(1, len(history['history']), figsize=(len(history['history']) * 5, 4))\n",
    "for i, (title, values) in enumerate(history['history'].items()):\n",
    "    axes[i].plot(np.array(values))\n",
    "    axes[i].set_title(title)\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Precision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import scipy\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.metrics import average_precision_score\n",
    "from keras import backend as K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_roc_score(edges_pos, edges_neg, adj_pred):\n",
    "    # Predict on test set of edges\n",
    "    preds = []\n",
    "    pos = []\n",
    "    for e in edges_pos:\n",
    "        preds.append(adj_pred[e[0], e[1]])\n",
    "        pos.append(adj[e[0], e[1]])\n",
    "\n",
    "    preds_neg = []\n",
    "    neg = []\n",
    "    for e in edges_neg:\n",
    "        preds_neg.append(adj_pred[e[0], e[1]])\n",
    "        neg.append(adj[e[0], e[1]])\n",
    "\n",
    "    preds_all = np.hstack([preds, preds_neg])\n",
    "    labels_all = np.hstack([np.ones(len(preds)), np.zeros(len(preds))])\n",
    "    roc_score = roc_auc_score(labels_all, preds_all)\n",
    "    ap_score = average_precision_score(labels_all, preds_all)\n",
    "\n",
    "    return roc_score, ap_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "q_preds, adj_preds, v_preds = zip(*[vae.predict_fullbatch(adj=adj_train, features=features) for _ in range(10)])\n",
    "\n",
    "q_preds = np.array(q_preds)\n",
    "adj_preds = np.array(adj_preds)\n",
    "\n",
    "q_pred = q_preds.mean(0)\n",
    "adj_pred = scipy.special.expit(adj_preds[:, 0, :, :, :]).mean(1).mean(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.9262214708368555, 0.9375222651716126)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_roc_score(test_edges, test_edges_false, adj_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for layer in vae.layers:\n",
    "    if hasattr(layer, 'kernel'):\n",
    "        fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(6, 4))#, sharey=True)\n",
    "        im1 = ax1.imshow(K.eval(layer.kernel).T)\n",
    "        ax1.set_title('{} kernel'.format(layer.name))\n",
    "        plt.colorbar(im1, ax=ax1)\n",
    "        if hasattr(layer, 'bias') and layer.bias is not None:\n",
    "            im2 = ax2.imshow(K.eval(K.expand_dims(layer.bias, -1)))\n",
    "            ax2.set_title('{} bias'.format(layer.name))\n",
    "            plt.colorbar(im2, ax=ax2)\n",
    "        else:\n",
    "            ax2.set_visible(False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}