{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "daXuA3h2b53x"
   },
   "source": [
    "# Exercise 9.2 - Solution\n",
    "Large arrays of radio antennas can be used to measure cosmic rays by recording the electromagnetic radiation generated in the atmosphere.\n",
    "These radio signals are strongly contaminated by galactic noise as well as signals from human origin. Since these signals appear to be similar to the background, the discovery of cosmic-ray events can be challenging.\n",
    "## Identification of signals\n",
    "In this exercise, we design an RNN to classify if the recorded radio signals contain a cosmic-ray event or only noise.\n",
    "\n",
    "The signal-to-noise ratio (SNR) of a measured trace $S(t)$ is defined as follows:\n",
    "\n",
    "$$\\mathrm{SNR}=\\frac{S^{\\mathrm{signal}}(t)_\\mathrm{max}}{\\mathrm{RMS}[S(t)]},$$  \n",
    "where $S^{\\mathrm{signal}}(t)_\\mathrm{max}$ denotes the maximum amplitude of the (true) signal.\n",
    "\n",
    "Typical cosmic-ray observatories enable a precise reconstruction at an SNR of roughly 3.\n",
    "\n",
    "We choose a challenging setup in this task and try to identify cosmic-ray events in signal traces with an SNR of 2.  \n",
    "Training RNNs can be computationally demanding, thus, we recommend to use a GPU for this task."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "yMkxS2Zrb533",
    "outputId": "0ca1dc6b-3095-46e5-d809-b95bbfaf7525"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "keras version 2.4.0\n"
     ]
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "layers = keras.layers\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "b7A1BSCHb539"
   },
   "source": [
    "### Load and prepare dataset\n",
    "In this task, we use a simulation of cosmic-ray-induced air showers that are measured by radion antennas.  \n",
    "For more information, see https://arxiv.org/abs/1901.04079.  \n",
    "The task is to design an RNN which is able to identify if the measured signal traces (shortened to 500 time steps) contains a signal or not."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "ss_iQSeWb54A"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import gdown\n",
    "url = \"https://drive.google.com/u/0/uc?export=download&confirm=HgGH&id=1R-qfxO1jVh88TC9Gnm9JGMomSRg0Zpkx\"\n",
    "output = 'radio_data.npz'\n",
    "\n",
    "if os.path.exists(output) == False:\n",
    "    gdown.download(url, output, quiet=True)\n",
    "\n",
    "f = np.load(output)\n",
    "n_train = 40000\n",
    "\n",
    "x_train, x_test = f[\"traces\"][:n_train], f[\"traces\"][n_train:]  # measured traces (signal + colored noise)\n",
    "signals = f[\"signals\"]  # signal part (only available for cosmic-ray events)\n",
    "\n",
    "labels = (signals.std(axis=-1) != 0).astype(float)  # define training label (1=cosmic event, 0=noise)\n",
    "y_train, y_test = labels[:n_train], labels[n_train:]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hpgUu--lb54B"
   },
   "source": [
    "## Plot example signal traces\n",
    "Left: signal trace containing a cosmic-ray event. The underlying cosmic-ray signal is shown in red, the backgrounds + signal is shown in blue.\n",
    "Right: background noise."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297
    },
    "id": "BiswHp3_b54D",
    "outputId": "9a0bc7a6-3c22-4b9c-b5d3-0728ee629593"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "fs = 180e6  # Sampling frequency of antenna setup 180 MHz\n",
    "t = np.arange(500) / fs * 1e6\n",
    "idx = np.random.randint(0, labels.sum()-1)\n",
    "idx2 =  np.random.randint(0, n_train - labels.sum())\n",
    "\n",
    "plt.figure(1, (12, 4))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(t, np.real(f[\"traces\"][labels.astype(bool)][idx]), linewidth = 1, color=\"b\", label=\"Measured trace\")\n",
    "plt.plot(t, np.real(signals[labels.astype(bool)][idx]), linewidth = 1, color=\"r\", label=\"CR signal\")\n",
    "plt.ylabel('Amplitude / mV')\n",
    "plt.xlabel('Time / $\\mu \\mathrm{s}$')\n",
    "plt.legend()\n",
    "plt.title(\"Cosmic-ray event\")\n",
    "plt.subplot(1, 2, 2)\n",
    "\n",
    "plt.plot(t, np.real(x_train[~y_train.astype(bool)][idx2]), linewidth = 1, color=\"b\", label=\"Measured trace\")\n",
    "plt.ylabel('Amplitude / mV')\n",
    "plt.xlabel('Time / $\\mu \\mathrm{s}$')\n",
    "plt.legend()\n",
    "plt.title(\"Noise event\")\n",
    "\n",
    "plt.grid(True)\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "id": "oRPX-EYOb54L"
   },
   "outputs": [],
   "source": [
    "sigma = x_train.std()\n",
    "x_train /= sigma\n",
    "x_test /= sigma"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "lOanYtv_b54G"
   },
   "source": [
    "### Define RNN model\n",
    "In the following, design a cosmic-ray model to identify cosmic-ray events using an RNN-based classifier."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "kBzMjinOb54G",
    "outputId": "1e892def-6270-4178-b372-65210f5d8f28"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "bidirectional (Bidirectional (None, 500, 64)           8704      \n",
      "_________________________________________________________________\n",
      "bidirectional_1 (Bidirection (None, 500, 128)          66048     \n",
      "_________________________________________________________________\n",
      "bidirectional_2 (Bidirection (None, 500, 20)           11120     \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 10000)             0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 10000)             0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1)                 10001     \n",
      "=================================================================\n",
      "Total params: 95,873\n",
      "Trainable params: 95,873\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = keras.models.Sequential()\n",
    "model.add(layers.Bidirectional(layers.LSTM(32, return_sequences=True), input_shape=(500,1)))\n",
    "model.add(layers.Bidirectional(layers.LSTM(64, return_sequences=True)))\n",
    "model.add(layers.Bidirectional(layers.LSTM(10, return_sequences=True)))\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dropout(0.3))\n",
    "model.add(layers.Dense(1, activation=\"sigmoid\"))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Bw8tkc1Xb54I"
   },
   "source": [
    "#### Pre-processing of data and RNN training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "fjncDgOcb54N",
    "outputId": "1fd1ec77-5711-401e-9c40-336c3a588f30"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "282/282 [==============================] - 42s 119ms/step - loss: 0.6340 - accuracy: 0.6387 - val_loss: 0.5846 - val_accuracy: 0.6823\n",
      "Epoch 2/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.5897 - accuracy: 0.6818 - val_loss: 0.5826 - val_accuracy: 0.6845\n",
      "Epoch 3/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5875 - accuracy: 0.6848 - val_loss: 0.5851 - val_accuracy: 0.6892\n",
      "Epoch 4/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5857 - accuracy: 0.6859 - val_loss: 0.5800 - val_accuracy: 0.6848\n",
      "Epoch 5/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5858 - accuracy: 0.6824 - val_loss: 0.6100 - val_accuracy: 0.6702\n",
      "Epoch 6/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5835 - accuracy: 0.6849 - val_loss: 0.5897 - val_accuracy: 0.6783\n",
      "Epoch 7/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5854 - accuracy: 0.6858 - val_loss: 0.5785 - val_accuracy: 0.6892\n",
      "Epoch 8/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5829 - accuracy: 0.6884 - val_loss: 0.5816 - val_accuracy: 0.6862\n",
      "Epoch 9/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5813 - accuracy: 0.6898 - val_loss: 0.5961 - val_accuracy: 0.6787\n",
      "Epoch 10/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5830 - accuracy: 0.6883 - val_loss: 0.5809 - val_accuracy: 0.6865\n",
      "Epoch 11/100\n",
      "282/282 [==============================] - 32s 115ms/step - loss: 0.5805 - accuracy: 0.6916 - val_loss: 0.5817 - val_accuracy: 0.6880\n",
      "Epoch 12/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5797 - accuracy: 0.6904 - val_loss: 0.5816 - val_accuracy: 0.6835\n",
      "\n",
      "Epoch 00012: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n",
      "Epoch 13/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5750 - accuracy: 0.6952 - val_loss: 0.5822 - val_accuracy: 0.6842\n",
      "Epoch 14/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5751 - accuracy: 0.6959 - val_loss: 0.5899 - val_accuracy: 0.6810\n",
      "Epoch 15/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5748 - accuracy: 0.6939 - val_loss: 0.5907 - val_accuracy: 0.6798\n",
      "Epoch 16/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5753 - accuracy: 0.6951 - val_loss: 0.5863 - val_accuracy: 0.6840\n",
      "Epoch 17/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5751 - accuracy: 0.6964 - val_loss: 0.5805 - val_accuracy: 0.6888\n",
      "\n",
      "Epoch 00017: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n",
      "Epoch 18/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5711 - accuracy: 0.6990 - val_loss: 0.5811 - val_accuracy: 0.6852\n",
      "Epoch 19/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5713 - accuracy: 0.6986 - val_loss: 0.5826 - val_accuracy: 0.6875\n",
      "Epoch 20/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5698 - accuracy: 0.6998 - val_loss: 0.5816 - val_accuracy: 0.6837\n",
      "Epoch 21/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5692 - accuracy: 0.6988 - val_loss: 0.5863 - val_accuracy: 0.6852\n",
      "Epoch 22/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5689 - accuracy: 0.7021 - val_loss: 0.5779 - val_accuracy: 0.6877\n",
      "Epoch 23/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5624 - accuracy: 0.7029 - val_loss: 0.5693 - val_accuracy: 0.6982\n",
      "Epoch 24/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5567 - accuracy: 0.7089 - val_loss: 0.5639 - val_accuracy: 0.7030\n",
      "Epoch 25/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5523 - accuracy: 0.7124 - val_loss: 0.5647 - val_accuracy: 0.7020\n",
      "Epoch 26/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5469 - accuracy: 0.7161 - val_loss: 0.5509 - val_accuracy: 0.7138\n",
      "Epoch 27/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5368 - accuracy: 0.7208 - val_loss: 0.5293 - val_accuracy: 0.7272\n",
      "Epoch 28/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5397 - accuracy: 0.7194 - val_loss: 0.5640 - val_accuracy: 0.7000\n",
      "Epoch 29/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.5312 - accuracy: 0.7279 - val_loss: 0.5275 - val_accuracy: 0.7247\n",
      "Epoch 30/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.5030 - accuracy: 0.7456 - val_loss: 0.4799 - val_accuracy: 0.7620\n",
      "Epoch 31/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.4591 - accuracy: 0.7753 - val_loss: 0.3993 - val_accuracy: 0.8167\n",
      "Epoch 32/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.4390 - accuracy: 0.7870 - val_loss: 0.3567 - val_accuracy: 0.8440\n",
      "Epoch 33/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.4531 - accuracy: 0.7831 - val_loss: 0.4448 - val_accuracy: 0.7890\n",
      "Epoch 34/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.3567 - accuracy: 0.8382 - val_loss: 0.4410 - val_accuracy: 0.7857\n",
      "Epoch 35/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.3203 - accuracy: 0.8605 - val_loss: 0.2384 - val_accuracy: 0.9038\n",
      "Epoch 36/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.2259 - accuracy: 0.9109 - val_loss: 0.2173 - val_accuracy: 0.9137\n",
      "Epoch 37/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.5097 - accuracy: 0.7615 - val_loss: 0.5205 - val_accuracy: 0.7337\n",
      "Epoch 38/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.4893 - accuracy: 0.7575 - val_loss: 0.4406 - val_accuracy: 0.7860\n",
      "Epoch 39/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.4003 - accuracy: 0.8154 - val_loss: 0.4422 - val_accuracy: 0.7965\n",
      "Epoch 40/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.3712 - accuracy: 0.8314 - val_loss: 0.3695 - val_accuracy: 0.8388\n",
      "Epoch 41/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.3175 - accuracy: 0.8637 - val_loss: 0.3010 - val_accuracy: 0.8750\n",
      "\n",
      "Epoch 00041: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n",
      "Epoch 42/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.2289 - accuracy: 0.9109 - val_loss: 0.2339 - val_accuracy: 0.9020\n",
      "Epoch 43/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1981 - accuracy: 0.9239 - val_loss: 0.1823 - val_accuracy: 0.9312\n",
      "Epoch 44/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.1666 - accuracy: 0.9381 - val_loss: 0.1497 - val_accuracy: 0.9433\n",
      "Epoch 45/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1438 - accuracy: 0.9469 - val_loss: 0.1193 - val_accuracy: 0.9592\n",
      "Epoch 46/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1368 - accuracy: 0.9502 - val_loss: 0.1190 - val_accuracy: 0.9572\n",
      "Epoch 47/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1219 - accuracy: 0.9573 - val_loss: 0.1381 - val_accuracy: 0.9460\n",
      "Epoch 48/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1132 - accuracy: 0.9605 - val_loss: 0.1858 - val_accuracy: 0.9273\n",
      "Epoch 49/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.1147 - accuracy: 0.9595 - val_loss: 0.0818 - val_accuracy: 0.9720\n",
      "Epoch 50/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1035 - accuracy: 0.9638 - val_loss: 0.0860 - val_accuracy: 0.9707\n",
      "Epoch 51/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.1018 - accuracy: 0.9651 - val_loss: 0.0693 - val_accuracy: 0.9785\n",
      "Epoch 52/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0805 - accuracy: 0.9726 - val_loss: 0.0606 - val_accuracy: 0.9803\n",
      "Epoch 53/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.1922 - accuracy: 0.9281 - val_loss: 0.0823 - val_accuracy: 0.9768\n",
      "Epoch 54/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0765 - accuracy: 0.9749 - val_loss: 0.0664 - val_accuracy: 0.9787\n",
      "Epoch 55/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0691 - accuracy: 0.9771 - val_loss: 0.0719 - val_accuracy: 0.9778\n",
      "Epoch 56/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0699 - accuracy: 0.9770 - val_loss: 0.0511 - val_accuracy: 0.9845\n",
      "Epoch 57/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0797 - accuracy: 0.9731 - val_loss: 0.2559 - val_accuracy: 0.8903\n",
      "Epoch 58/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0792 - accuracy: 0.9739 - val_loss: 0.0531 - val_accuracy: 0.9833\n",
      "Epoch 59/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0639 - accuracy: 0.9793 - val_loss: 0.0669 - val_accuracy: 0.9820\n",
      "Epoch 60/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0652 - accuracy: 0.9791 - val_loss: 0.0574 - val_accuracy: 0.9790\n",
      "Epoch 61/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0591 - accuracy: 0.9812 - val_loss: 0.0805 - val_accuracy: 0.9820\n",
      "\n",
      "Epoch 00061: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n",
      "Epoch 62/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0427 - accuracy: 0.9866 - val_loss: 0.0608 - val_accuracy: 0.9803\n",
      "Epoch 63/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0414 - accuracy: 0.9868 - val_loss: 0.0434 - val_accuracy: 0.9893\n",
      "Epoch 64/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0424 - accuracy: 0.9862 - val_loss: 0.0377 - val_accuracy: 0.9918\n",
      "Epoch 65/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0424 - accuracy: 0.9859 - val_loss: 0.0402 - val_accuracy: 0.9872\n",
      "Epoch 66/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0430 - accuracy: 0.9868 - val_loss: 0.0379 - val_accuracy: 0.9883\n",
      "Epoch 67/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0412 - accuracy: 0.9865 - val_loss: 0.0324 - val_accuracy: 0.9900\n",
      "Epoch 68/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0383 - accuracy: 0.9883 - val_loss: 0.0414 - val_accuracy: 0.9890\n",
      "Epoch 69/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0394 - accuracy: 0.9878 - val_loss: 0.0321 - val_accuracy: 0.9908\n",
      "Epoch 70/100\n",
      "282/282 [==============================] - 32s 112ms/step - loss: 0.0350 - accuracy: 0.9892 - val_loss: 0.0372 - val_accuracy: 0.9887\n",
      "Epoch 71/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0379 - accuracy: 0.9879 - val_loss: 0.0342 - val_accuracy: 0.9893\n",
      "Epoch 72/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0532 - accuracy: 0.9840 - val_loss: 0.0338 - val_accuracy: 0.9902\n",
      "Epoch 73/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0390 - accuracy: 0.9879 - val_loss: 0.0338 - val_accuracy: 0.9910\n",
      "Epoch 74/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0392 - accuracy: 0.9875 - val_loss: 0.0455 - val_accuracy: 0.9885\n",
      "\n",
      "Epoch 00074: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n",
      "Epoch 75/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0332 - accuracy: 0.9896 - val_loss: 0.0285 - val_accuracy: 0.9933\n",
      "Epoch 76/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0301 - accuracy: 0.9906 - val_loss: 0.0382 - val_accuracy: 0.9915\n",
      "Epoch 77/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0296 - accuracy: 0.9910 - val_loss: 0.0328 - val_accuracy: 0.9900\n",
      "Epoch 78/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0297 - accuracy: 0.9910 - val_loss: 0.0326 - val_accuracy: 0.9898\n",
      "Epoch 79/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0312 - accuracy: 0.9907 - val_loss: 0.0296 - val_accuracy: 0.9935\n",
      "Epoch 80/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0285 - accuracy: 0.9913 - val_loss: 0.0797 - val_accuracy: 0.9775\n",
      "\n",
      "Epoch 00080: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n",
      "Epoch 81/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0264 - accuracy: 0.9922 - val_loss: 0.0282 - val_accuracy: 0.9937\n",
      "Epoch 82/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0254 - accuracy: 0.9923 - val_loss: 0.0277 - val_accuracy: 0.9935\n",
      "Epoch 83/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0264 - accuracy: 0.9918 - val_loss: 0.0257 - val_accuracy: 0.9933\n",
      "Epoch 84/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0255 - accuracy: 0.9921 - val_loss: 0.0284 - val_accuracy: 0.9937\n",
      "Epoch 85/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0248 - accuracy: 0.9923 - val_loss: 0.0283 - val_accuracy: 0.9933\n",
      "Epoch 86/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0242 - accuracy: 0.9927 - val_loss: 0.0282 - val_accuracy: 0.9920\n",
      "Epoch 87/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0251 - accuracy: 0.9921 - val_loss: 0.0273 - val_accuracy: 0.9925\n",
      "Epoch 88/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0246 - accuracy: 0.9927 - val_loss: 0.0349 - val_accuracy: 0.9923\n",
      "\n",
      "Epoch 00088: ReduceLROnPlateau reducing learning rate to 1e-05.\n",
      "Epoch 89/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0236 - accuracy: 0.9931 - val_loss: 0.0277 - val_accuracy: 0.9925\n",
      "Epoch 90/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0237 - accuracy: 0.9928 - val_loss: 0.0260 - val_accuracy: 0.9935\n",
      "Epoch 91/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0242 - accuracy: 0.9928 - val_loss: 0.0257 - val_accuracy: 0.9940\n",
      "Epoch 92/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0242 - accuracy: 0.9926 - val_loss: 0.0258 - val_accuracy: 0.9942\n",
      "Epoch 93/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0239 - accuracy: 0.9928 - val_loss: 0.0253 - val_accuracy: 0.9942\n",
      "Epoch 94/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0237 - accuracy: 0.9930 - val_loss: 0.0245 - val_accuracy: 0.9935\n",
      "Epoch 95/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0234 - accuracy: 0.9931 - val_loss: 0.0249 - val_accuracy: 0.9940\n",
      "Epoch 96/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0235 - accuracy: 0.9931 - val_loss: 0.0247 - val_accuracy: 0.9935\n",
      "Epoch 97/100\n",
      "282/282 [==============================] - 32s 114ms/step - loss: 0.0234 - accuracy: 0.9927 - val_loss: 0.0251 - val_accuracy: 0.9930\n",
      "Epoch 98/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0234 - accuracy: 0.9931 - val_loss: 0.0254 - val_accuracy: 0.9942\n",
      "Epoch 99/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0225 - accuracy: 0.9935 - val_loss: 0.0270 - val_accuracy: 0.9942\n",
      "Epoch 100/100\n",
      "282/282 [==============================] - 32s 113ms/step - loss: 0.0228 - accuracy: 0.9931 - val_loss: 0.0274 - val_accuracy: 0.9940\n"
     ]
    }
   ],
   "source": [
    "model.compile(\n",
    "    loss='binary_crossentropy',\n",
    "    optimizer=keras.optimizers.Adam(1e-3, decay=0.00008),\n",
    "    metrics=['accuracy'])\n",
    "\n",
    "\n",
    "results = model.fit(x_train[...,np.newaxis], y_train,\n",
    "                    batch_size=128,\n",
    "                    epochs=100,\n",
    "                    verbose=1,\n",
    "                    validation_split=0.1,\n",
    "                    callbacks = [keras.callbacks.ReduceLROnPlateau(factor=0.5, patience=5, verbose=1, min_lr=1e-5),\n",
    "                                 keras.callbacks.EarlyStopping(patience=15, verbose=1)]\n",
    "                    )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "xzutH-Qkb54P",
    "outputId": "dbf0745e-dab6-4073-9466-c42bf6e10c0e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 13s 40ms/step - loss: 0.0250 - accuracy: 0.9920\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.024966726079583168, 0.9919999837875366]"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(x_test[...,np.newaxis], y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "A16SCLCBb54R"
   },
   "source": [
    "### Plot loss and accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 297
    },
    "id": "8gDw4qLVb54R",
    "outputId": "5f6afb0c-e345-4736-cf3c-e7776f9b5170"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(1, (12, 4))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(results.history['loss'])\n",
    "plt.plot(results.history['val_loss'])\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper right')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(results.history['accuracy'])\n",
    "plt.plot(results.history['val_accuracy'])\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'val'], loc='upper left')\n",
    "plt.tight_layout()"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "Exercise_9_2_solution.ipynb",
   "provenance": []
  },
  "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}