example.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "np.random.seed(12)\n",
    "\n",
    "from keras.layers import Input\n",
    "from keras.layers import Dense\n",
    "from keras.layers import LSTM\n",
    "from keras.layers import Masking\n",
    "\n",
    "from keras import optimizers\n",
    "from keras import models\n",
    "from attention.layers import AttentionLayer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generating the dataset\n",
    "\n",
    "Each sample is a \"tape\" formed by two rows: the top row $y[0]$ contains random numbers sampled from the interval $[0, 1)$, while the second, $y[1]$, is formed by a string of zeros, except for one position, which has a one. The model is trained as a regressor to produce on the output the value from the first row of the column marked as 1. For example, given:\n",
    "\n",
    "$$ X = \n",
    "\\begin{bmatrix}\n",
    "    0.13 & 0.01 & 0.11 & 0.32 & 0.24 & 0.01 \\\\\n",
    "    0    & 0    & 0    & 0    &   1  & 0    \\\\\n",
    "\\end{bmatrix},\n",
    "$$\n",
    "\n",
    "the model should produce $\\mathcal{M}_\\theta(X) = 0.24$. We train two models, one without attention and other with, as shown below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "N_SAMPLES = 1000\n",
    "TIMESTEPS = 64\n",
    "\n",
    "# (b, t, d)\n",
    "X = np.random.rand(N_SAMPLES, TIMESTEPS, 1)\n",
    "F = np.zeros(shape=(N_SAMPLES, TIMESTEPS, 1))\n",
    "X = np.concatenate((X, F), axis=2)\n",
    "\n",
    "Y = list()\n",
    "\n",
    "correct_timesteps = np.random.randint(low=0, high=TIMESTEPS, size=(N_SAMPLES))\n",
    "for sample, timestep in enumerate(correct_timesteps):\n",
    "    X[sample][timestep][1] = 1\n",
    "    \n",
    "    Y.append(X[sample][timestep][0])\n",
    "    \n",
    "Y = np.asarray(Y).reshape(-1, 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Defining a simple and an attention model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_vanilla_model():\n",
    "    model_in = Input(shape=(TIMESTEPS, 2), name='sequence-in')\n",
    "    vectors  = LSTM(units=4, name='lstm')(model_in)\n",
    "    output   = Dense(units=1, activation='linear')(vectors)\n",
    "\n",
    "    # (b, t, d)\n",
    "    model = models.Model(inputs=[model_in], outputs=[output])\n",
    "    model.summary()\n",
    "\n",
    "    model.compile(optimizers.Adam(1e-2), 'mse', metrics=['mse'])\n",
    "    return model\n",
    "\n",
    "\n",
    "def build_attention_model():\n",
    "    model_in = Input(shape=(None, 2), name='sequence-in')\n",
    "    masked   = Masking(name='mask')(model_in)\n",
    "    vectors  = LSTM(units=4, name='lstm', return_sequences=True)(masked)\n",
    "\n",
    "    descr    = AttentionLayer(name='attention')(vectors)\n",
    "    output   = Dense(units=1, activation='linear')(descr)\n",
    "\n",
    "    # (b, t, d)\n",
    "    model = models.Model(inputs=[model_in], outputs=[output])\n",
    "    model.summary()\n",
    "\n",
    "    model.compile(optimizers.Adam(5e-2), 'mse', metrics=['mse'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training them"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequence-in (InputLayer)     (None, 64, 2)             0         \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, 4)                 112       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 117\n",
      "Trainable params: 117\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.1477 - mean_squared_error: 0.1477\n",
      "Epoch 2/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0868 - mean_squared_error: 0.0868\n",
      "Epoch 3/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0860 - mean_squared_error: 0.0860\n",
      "Epoch 4/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0859 - mean_squared_error: 0.0859\n",
      "Epoch 5/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0856 - mean_squared_error: 0.0856\n",
      "Epoch 6/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0859 - mean_squared_error: 0.0859\n",
      "Epoch 7/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0861 - mean_squared_error: 0.0861\n",
      "Epoch 8/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0866 - mean_squared_error: 0.0866\n",
      "Epoch 9/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0859 - mean_squared_error: 0.0859\n",
      "Epoch 10/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0860 - mean_squared_error: 0.0860\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f4c98d57550>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vanilla_model = build_vanilla_model()\n",
    "vanilla_model.fit(X, Y, batch_size=32, epochs=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequence-in (InputLayer)     (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "mask (Masking)               (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, None, 4)           112       \n",
      "_________________________________________________________________\n",
      "attention (AttentionLayer)   (None, 4)                 24        \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 141\n",
      "Trainable params: 141\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/10\n",
      "1000/1000 [==============================] - 4s 4ms/step - loss: 0.1063 - mean_squared_error: 0.1063\n",
      "Epoch 2/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0822 - mean_squared_error: 0.0822\n",
      "Epoch 3/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0094 - mean_squared_error: 0.0094\n",
      "Epoch 4/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 0.0020 - mean_squared_error: 0.0020\n",
      "Epoch 5/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 8.2353e-04 - mean_squared_error: 8.2353e-04\n",
      "Epoch 6/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 2.9403e-04 - mean_squared_error: 2.9403e-04\n",
      "Epoch 7/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 1.6836e-04 - mean_squared_error: 1.6836e-04\n",
      "Epoch 8/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 1.3372e-04 - mean_squared_error: 1.3372e-04\n",
      "Epoch 9/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 1.4867e-04 - mean_squared_error: 1.4867e-04\n",
      "Epoch 10/10\n",
      "1000/1000 [==============================] - 3s 3ms/step - loss: 1.6271e-04 - mean_squared_error: 1.6271e-04\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f4c8c762be0>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "attention_model = build_attention_model()\n",
    "attention_model.predict(X).shape\n",
    "attention_model.fit(X, Y, batch_size=32, epochs=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Comparing loss functinos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f4c21dc92e8>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(vanilla_model.history.history['loss'], 'o-', alpha=0.8, label='Without attention')\n",
    "plt.plot(attention_model.history.history['loss'], 'o-', alpha=0.8, label='With attention')\n",
    "plt.grid(alpha=0.4)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualizing attention weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def copy_attention_model(model):    \n",
    "    model_in = Input(shape=(None, 2), name='sequence-in')\n",
    "    masked   = Masking(name='mask')(model_in)\n",
    "    vectors  = LSTM(units=4, name='lstm', return_sequences=True, weights=model.layers[2].get_weights())(masked)\n",
    "\n",
    "    attention = AttentionLayer(name='attention', return_attention=True, weights=model.layers[3].get_weights())(vectors)\n",
    "    \n",
    "    new_model = models.Model(inputs=[model_in], outputs=[attention])\n",
    "    new_model.summary()\n",
    "\n",
    "    return new_model\n",
    "\n",
    "def plot_attention_graph(selected_samples, attention_coefs, selected_values):\n",
    "    for sample_id, attention, y in zip(selected_samples, attention_coefs, selected_values):\n",
    "        plt.figure(figsize=(20, 3))\n",
    "        plt.title(f'Sample #{sample_id}. Correct value @ index {y}')\n",
    "        plt.xlabel('Timesteps')\n",
    "        plt.ylabel('Attention coefficient')\n",
    "\n",
    "        plt.grid(alpha=0.4)\n",
    "        plt.plot(attention)\n",
    "        \n",
    "        max_attention = np.max(attention)\n",
    "\n",
    "        U = np.linspace(0, max_attention, 20)\n",
    "        y = np.ones_like(U) * y\n",
    "        plt.plot(y, U, 'r', alpha=0.8, label='Correct timestep')\n",
    "\n",
    "        plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequence-in (InputLayer)     (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "mask (Masking)               (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, None, 4)           112       \n",
      "_________________________________________________________________\n",
      "attention (AttentionLayer)   (None, None)              24        \n",
      "=================================================================\n",
      "Total params: 136\n",
      "Trainable params: 136\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "coef_model = copy_attention_model(attention_model)\n",
    "\n",
    "\n",
    "N_PICKS = 5\n",
    "\n",
    "selected_samples = np.random.randint(0, high=N_SAMPLES, size=N_PICKS)\n",
    "selected_values  = correct_timesteps[selected_samples]\n",
    "attention_coefs  = coef_model.predict(X[selected_samples])\n",
    "\n",
    "plot_attention_graph(selected_samples, attention_coefs, selected_values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training a model with attention layer and masking\n",
    "\n",
    "Notice that the model is trained for fewer epochs to make the \"bump\" on the attention coefficients between the masked and non-masked regions visible. It's actually a bit harder to stop the training in a good spot by just changing the amount of epochs. If the loss is too low, the bump is not visible, if it is too high, the attention simply makes no sense, since the model is still learning to what attend to."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequence-in (InputLayer)     (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "mask (Masking)               (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, None, 4)           112       \n",
      "_________________________________________________________________\n",
      "attention (AttentionLayer)   (None, 4)                 24        \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 1)                 5         \n",
      "=================================================================\n",
      "Total params: 141\n",
      "Trainable params: 141\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Epoch 1/2\n",
      "1000/1000 [==============================] - 6s 6ms/step - loss: 0.0933 - mean_squared_error: 0.0933\n",
      "Epoch 2/2\n",
      "1000/1000 [==============================] - 5s 5ms/step - loss: 0.0263 - mean_squared_error: 0.0263\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "sequence-in (InputLayer)     (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "mask (Masking)               (None, None, 2)           0         \n",
      "_________________________________________________________________\n",
      "lstm (LSTM)                  (None, None, 4)           112       \n",
      "_________________________________________________________________\n",
      "attention (AttentionLayer)   (None, None)              24        \n",
      "=================================================================\n",
      "Total params: 136\n",
      "Trainable params: 136\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "AMOUNT_BLANKS = 16\n",
    "\n",
    "# prepending each sample is AMOUNT_BLANKS blank timesteps\n",
    "O = np.zeros(shape=(1000, AMOUNT_BLANKS, 2))\n",
    "O = np.concatenate((O, X), axis=1)\n",
    "\n",
    "# training the model\n",
    "masked_model = build_attention_model()\n",
    "masked_model.fit(O, Y, batch_size=16, epochs=2)\n",
    "\n",
    "# building the model to get the attention weights\n",
    "masked_model_coefs = copy_attention_model(masked_model)\n",
    "\n",
    "# picking some samples for visualization\n",
    "selected_samples = np.random.randint(0, high=N_SAMPLES, size=N_PICKS)\n",
    "selected_values  = correct_timesteps[selected_samples] + AMOUNT_BLANKS\n",
    "attention_coefs  = masked_model_coefs.predict(O[selected_samples])\n",
    "\n",
    "# plotting their attention weights\n",
    "plot_attention_graph(selected_samples, attention_coefs, selected_values)"
   ]
  }
 ],
 "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}