make_plots.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from data_gen import pendulum\n",
    "from scipy.stats import norm\n",
    "from sklearn.metrics import mean_squared_error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import rcParams,rc\n",
    "rc('text', usetex=True)\n",
    "rcParams['font.family'] = 'serif'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def medians_from_scatter(x, y, n_bins=30, skip_end=5):\n",
    "    left_lim = np.amin(x)\n",
    "    right_lim = np.amax(x)\n",
    "    right_lim += (right_lim - left_lim)/100000.\n",
    "    bin_edges = np.linspace(left_lim, right_lim, n_bins+1)\n",
    "    bin_centers = (bin_edges[1:] + bin_edges[:-1])/2.\n",
    "    idxs = [np.logical_and(x >= bin_edges[i], x < bin_edges[i+1]) for i in np.arange(n_bins)]\n",
    "    \n",
    "    median = np.array([np.percentile(y[idx], 50) for idx in idxs[:-skip_end]])\n",
    "    lower = np.array([np.percentile(y[idx], 16) for idx in idxs[:-skip_end]])\n",
    "    upper = np.array([np.percentile(y[idx], 84) for idx in idxs[:-skip_end]])\n",
    "    \n",
    "    return bin_centers[:-skip_end], median, lower, upper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def find_coverage(res, unc):\n",
    "    sigma = np.abs(res/unc)\n",
    "    cdf_sigma = 2*norm.cdf(sigma)-1\n",
    "    x = np.linspace(0.01, 1, 100)\n",
    "    coverage = [sum(cdf_sigma <= k)/nobject for k in x]\n",
    "    return coverage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# make a test set\n",
    "nobject=10000\n",
    "\n",
    "feat_test_t1to10, y_test_t1to10, calc_y_test_t1to10, delta_y_test_t1to10 = pendulum(n=nobject, t_spread=[0.01, 0.1],\n",
    "                                                  ell_spread=[0.02, 0.02], seed=666)\n",
    "feat_test_t1to5, y_test_t1to5, calc_y_test_t1to5, delta_y_test_t1to5 = pendulum(n=nobject, t_spread=[0.01, 0.05],\n",
    "                                                  ell_spread=[0.02, 0.02], seed=666)\n",
    "feat_test_t1to20, y_test_t1to20, calc_y_test_t1to20, delta_y_test_t1to20 = pendulum(n=nobject, t_spread=[0.01, 0.2],\n",
    "                                                  ell_spread=[0.02, 0.02], seed=666)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_de_t1to10 = np.load('data/y_pred_test_de_trange1to10_40ep_001.npy')\n",
    "y_pred_alunc_de_t1to10 = np.load('data/y_pred_test_alunc_de_trange1to10_40ep_001.npy')\n",
    "y_pred_epunc_de_t1to10 = np.load('data/y_pred_test_epunc_de_trange1to10_40ep_001.npy')\n",
    "y_pred_prunc_de_t1to10 = np.load('data/y_pred_test_prunc_de_trange1to10_40ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_de_t1to20 = np.load('data/y_pred_test_de_trange1to20_40ep_001.npy')\n",
    "y_pred_alunc_de_t1to20 = np.load('data/y_pred_test_alunc_de_trange1to20_40ep_001.npy')\n",
    "y_pred_epunc_de_t1to20 = np.load('data/y_pred_test_epunc_de_trange1to20_40ep_001.npy')\n",
    "y_pred_prunc_de_t1to20 = np.load('data/y_pred_test_prunc_de_trange1to20_40ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_de_t1to5 = np.load('data/y_pred_test_de_trange1to5_40ep_001.npy')\n",
    "y_pred_alunc_de_t1to5 = np.load('data/y_pred_test_alunc_de_trange1to5_40ep_001.npy')\n",
    "y_pred_epunc_de_t1to5 = np.load('data/y_pred_test_epunc_de_trange1to5_40ep_001.npy')\n",
    "y_pred_prunc_de_t1to5 = np.load('data/y_pred_test_prunc_de_trange1to5_40ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_bnn_t1to10 = np.load('data/y_pred_test_bnn_trange1to10_200ep_001.npy')\n",
    "y_pred_alunc_bnn_t1to10 = np.load('data/y_pred_test_alunc_bnn_trange1to10_200ep_001.npy')\n",
    "y_pred_epunc_bnn_t1to10 = np.load('data/y_pred_test_epunc_bnn_trange1to10_200ep_001.npy')\n",
    "y_pred_prunc_bnn_t1to10 = np.load('data/y_pred_test_prunc_bnn_trange1to10_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_bnn_t1to20 = np.load('data/y_pred_test_bnn_trange1to20_200ep_001.npy')\n",
    "y_pred_alunc_bnn_t1to20 = np.load('data/y_pred_test_alunc_bnn_trange1to20_200ep_001.npy')\n",
    "y_pred_epunc_bnn_t1to20 = np.load('data/y_pred_test_epunc_bnn_trange1to20_200ep_001.npy')\n",
    "y_pred_prunc_bnn_t1to20 = np.load('data/y_pred_test_prunc_bnn_trange1to20_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_bnn_t1to5 = np.load('data/y_pred_test_bnn_trange1to5_200ep_001.npy')\n",
    "y_pred_alunc_bnn_t1to5 = np.load('data/y_pred_test_alunc_bnn_trange1to5_200ep_001.npy')\n",
    "y_pred_epunc_bnn_t1to5 = np.load('data/y_pred_test_epunc_bnn_trange1to5_200ep_001.npy')\n",
    "y_pred_prunc_bnn_t1to5 = np.load('data/y_pred_test_prunc_bnn_trange1to5_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_cd_t1to10 = np.load('data/y_pred_test_cd_trange1to10_200ep_001.npy')\n",
    "y_pred_alunc_cd_t1to10 = np.load('data/y_pred_test_alunc_cd_trange1to10_200ep_001.npy')\n",
    "y_pred_epunc_cd_t1to10 = np.load('data/y_pred_test_epunc_cd_trange1to10_200ep_001.npy')\n",
    "y_pred_prunc_cd_t1to10 = np.load('data/y_pred_test_prunc_cd_trange1to10_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_cd_t1to20 = np.load('data/y_pred_test_cd_trange1to20_200ep_001.npy')\n",
    "y_pred_alunc_cd_t1to20 = np.load('data/y_pred_test_alunc_cd_trange1to20_200ep_001.npy')\n",
    "y_pred_epunc_cd_t1to20 = np.load('data/y_pred_test_epunc_cd_trange1to20_200ep_001.npy')\n",
    "y_pred_prunc_cd_t1to20 = np.load('data/y_pred_test_prunc_cd_trange1to20_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_cd_t1to5 = np.load('data/y_pred_test_cd_trange1to5_200ep_001.npy')\n",
    "y_pred_alunc_cd_t1to5 = np.load('data/y_pred_test_alunc_cd_trange1to5_200ep_001.npy')\n",
    "y_pred_epunc_cd_t1to5 = np.load('data/y_pred_test_epunc_cd_trange1to5_200ep_001.npy')\n",
    "y_pred_prunc_cd_t1to5 = np.load('data/y_pred_test_prunc_cd_trange1to5_200ep_001.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "analytic_t1to10, median_de_t1to10, low_de_t1to10, upp_de_t1to10 = medians_from_scatter(delta_y_test_t1to10/calc_y_test_t1to10, \n",
    "                                                                                       y_pred_alunc_de_t1to10/y_pred_de_t1to10)\n",
    "analytic_t1to20, median_de_t1to20, low_de_t1to20, upp_de_t1to20 = medians_from_scatter(delta_y_test_t1to20/calc_y_test_t1to20, \n",
    "                                                                                       y_pred_alunc_de_t1to20/y_pred_de_t1to20)\n",
    "analytic_t1to5, median_de_t1to5, low_de_t1to5, upp_de_t1to5 = medians_from_scatter(delta_y_test_t1to5/calc_y_test_t1to5, \n",
    "                                                                                   y_pred_alunc_de_t1to5/y_pred_de_t1to5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "_, median_bnn_t1to10, low_bnn_t1to10, upp_bnn_t1to10 = medians_from_scatter(delta_y_test_t1to10/calc_y_test_t1to10, \n",
    "                                                                            y_pred_alunc_bnn_t1to10/y_pred_bnn_t1to10)\n",
    "_, median_bnn_t1to20, low_bnn_t1to20, upp_bnn_t1to20 = medians_from_scatter(delta_y_test_t1to20/calc_y_test_t1to20, \n",
    "                                                                            y_pred_alunc_bnn_t1to20/y_pred_bnn_t1to20)\n",
    "_, median_bnn_t1to5, low_bnn_t1to5, upp_bnn_t1to5 = medians_from_scatter(delta_y_test_t1to5/calc_y_test_t1to5, \n",
    "                                                                         y_pred_alunc_bnn_t1to5/y_pred_bnn_t1to5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "_, median_cd_t1to10, low_cd_t1to10, upp_cd_t1to10 = medians_from_scatter(delta_y_test_t1to10/calc_y_test_t1to10, \n",
    "                                                                         y_pred_alunc_cd_t1to10/y_pred_cd_t1to10)\n",
    "_, median_cd_t1to20, low_cd_t1to20, upp_cd_t1to20 = medians_from_scatter(delta_y_test_t1to20/calc_y_test_t1to20, \n",
    "                                                                         y_pred_alunc_cd_t1to20/y_pred_cd_t1to20)\n",
    "_, median_cd_t1to5, low_cd_t1to5, upp_cd_t1to5 = medians_from_scatter(delta_y_test_t1to5/calc_y_test_t1to5, \n",
    "                                                                      y_pred_alunc_cd_t1to5/y_pred_cd_t1to5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x324 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fontsize=14\n",
    "fontsize_title=16\n",
    "\n",
    "plt.figure(figsize=(14,4.5))\n",
    "\n",
    "plt.subplot(131)\n",
    "plt.plot(analytic_t1to5, median_de_t1to5, alpha=0.5, c='#000080', label='DE')\n",
    "plt.fill_between(analytic_t1to5, low_de_t1to5, upp_de_t1to5,\n",
    "                facecolor='#000080', alpha=0.2)\n",
    "plt.plot(analytic_t1to5, median_cd_t1to5, alpha=0.5, c='#800080', label='CD')\n",
    "plt.fill_between(analytic_t1to5, low_cd_t1to5, upp_cd_t1to5,\n",
    "                facecolor='#800080', alpha=0.2)\n",
    "plt.plot(analytic_t1to5, median_bnn_t1to5, alpha=0.5, c='#FF0000', label='BNN')\n",
    "plt.fill_between(analytic_t1to5, low_bnn_t1to5, upp_bnn_t1to5,\n",
    "                facecolor='#FF0000', alpha=0.2)\n",
    "start_pt = 0.02\n",
    "end_pt = 0.05\n",
    "plt.plot([start_pt, end_pt], [start_pt, end_pt], '--', color=(0,0,0))\n",
    "plt.xlim([start_pt, end_pt])\n",
    "plt.ylim([start_pt, end_pt])\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('Trained with T noise of 1-5\\%', fontsize=fontsize_title)\n",
    "plt.xlabel('Relative analytic uncertainty estimates', fontsize=fontsize)\n",
    "plt.ylabel('Relative aleatoric uncertainty estimates', fontsize=fontsize)\n",
    "\n",
    "ax0 = plt.subplot(132)\n",
    "plt.plot(analytic_t1to10, median_de_t1to10, alpha=0.5, c='#000080', label='DE')\n",
    "plt.fill_between(analytic_t1to10, low_de_t1to10, upp_de_t1to10,\n",
    "                facecolor='#000080', alpha=0.2)\n",
    "plt.plot(analytic_t1to10, median_cd_t1to10, alpha=0.5, c='#800080', label='CD')\n",
    "plt.fill_between(analytic_t1to10, low_cd_t1to10, upp_cd_t1to10,\n",
    "                facecolor='#800080', alpha=0.2)\n",
    "plt.plot(analytic_t1to10, median_bnn_t1to10, alpha=0.5, c='#FF0000', label='BNN')\n",
    "plt.fill_between(analytic_t1to10, low_bnn_t1to10, upp_bnn_t1to10,\n",
    "                facecolor='#FF0000', alpha=0.2)\n",
    "start_pt = 0.025\n",
    "end_pt = 0.07\n",
    "plt.plot([start_pt, end_pt], [start_pt, end_pt], '--', color=(0,0,0))\n",
    "plt.xlim([start_pt, end_pt])\n",
    "plt.ylim([start_pt, end_pt])\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('Trained with T noise of 1-10\\%', fontsize=fontsize_title)\n",
    "plt.xlabel('Relative analytic uncertainty estimates', fontsize=fontsize)\n",
    "\n",
    "ax0 = plt.subplot(133)\n",
    "plt.plot(analytic_t1to20, median_de_t1to20, alpha=0.5, c='#000080', label='DE')\n",
    "plt.fill_between(analytic_t1to20, low_de_t1to20, upp_de_t1to20,\n",
    "                facecolor='#000080', alpha=0.2)\n",
    "plt.plot(analytic_t1to20, median_cd_t1to20, alpha=0.5, c='#800080', label='CD')\n",
    "plt.fill_between(analytic_t1to20, low_cd_t1to20, upp_cd_t1to20,\n",
    "                facecolor='#800080', alpha=0.2)\n",
    "plt.plot(analytic_t1to20, median_bnn_t1to20, alpha=0.5, c='#FF0000', label='BNN')\n",
    "plt.fill_between(analytic_t1to20, low_bnn_t1to20, upp_bnn_t1to20,\n",
    "                facecolor='#FF0000', alpha=0.2)\n",
    "start_pt = 0.025\n",
    "end_pt = 0.14\n",
    "plt.plot([start_pt, end_pt], [start_pt, end_pt], '--', color=(0,0,0))\n",
    "plt.xlim([start_pt, end_pt])\n",
    "plt.ylim([start_pt, end_pt])\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('Trained with T noise of 1-20\\%', fontsize=fontsize_title)\n",
    "plt.xlabel('Relative analytic uncertainty estimates', fontsize=fontsize)\n",
    "plt.tight_layout()\n",
    "plt.savefig('comparisons_diffnoise_perc.pdf', fmt='pdf', dpi=500)\n",
    "plt.savefig('comparisons_diffnoise_perc.png', fmt='png', dpi=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_bnn_oodl = []\n",
    "y_pred_alunc_bnn_oodl = []\n",
    "y_pred_epunc_bnn_oodl = []\n",
    "y_pred_prunc_bnn_oodl = []\n",
    "for i in range(2, 16):\n",
    "    y_pred_bnn_oodl.append(np.load(f'data/y_pred_test_bnn_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_bnn_oodl.append(np.load(f'data/y_pred_test_alunc_bnn_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_bnn_oodl.append(np.load(f'data/y_pred_test_epunc_bnn_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_bnn_oodl.append(np.load(f'data/y_pred_test_prunc_bnn_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    \n",
    "y_pred_de_oodl = []\n",
    "y_pred_alunc_de_oodl = []\n",
    "y_pred_epunc_de_oodl = []\n",
    "y_pred_prunc_de_oodl = []\n",
    "for i in range(2, 16):\n",
    "    y_pred_de_oodl.append(np.load(f'data/y_pred_test_de_trange1to20_40ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_de_oodl.append(np.load(f'data/y_pred_test_alunc_de_trange1to20_40ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_de_oodl.append(np.load(f'data/y_pred_test_epunc_de_trange1to20_40ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_de_oodl.append(np.load(f'data/y_pred_test_prunc_de_trange1to20_40ep_001_ell{i}to{i+1}.npy'))\n",
    "    \n",
    "y_pred_cd_oodl = []\n",
    "y_pred_alunc_cd_oodl = []\n",
    "y_pred_epunc_cd_oodl = []\n",
    "y_pred_prunc_cd_oodl = []\n",
    "for i in range(2, 16):\n",
    "    y_pred_cd_oodl.append(np.load(f'data/y_pred_test_cd_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_cd_oodl.append(np.load(f'data/y_pred_test_alunc_cd_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_cd_oodl.append(np.load(f'data/y_pred_test_epunc_cd_trange1to20_200ep_001_ell{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_cd_oodl.append(np.load(f'data/y_pred_test_prunc_cd_trange1to20_200ep_001_ell{i}to{i+1}.npy'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_epunc_bnn_oodl_50 = np.percentile(y_pred_epunc_bnn_oodl, 50, axis=1)\n",
    "y_pred_epunc_cd_oodl_50 = np.percentile(y_pred_epunc_cd_oodl, 50, axis=1)\n",
    "y_pred_epunc_de_oodl_50 = np.percentile(y_pred_epunc_de_oodl, 50, axis=1)\n",
    "y_pred_epunc_bnn_oodl_16 = np.percentile(y_pred_epunc_bnn_oodl, 16, axis=1)\n",
    "y_pred_epunc_cd_oodl_16 = np.percentile(y_pred_epunc_cd_oodl, 16, axis=1)\n",
    "y_pred_epunc_de_oodl_16 = np.percentile(y_pred_epunc_de_oodl, 16, axis=1)\n",
    "y_pred_epunc_bnn_oodl_84 = np.percentile(y_pred_epunc_bnn_oodl, 84, axis=1)\n",
    "y_pred_epunc_cd_oodl_84 = np.percentile(y_pred_epunc_cd_oodl, 84, axis=1)\n",
    "y_pred_epunc_de_oodl_84 = np.percentile(y_pred_epunc_de_oodl, 84, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fontsize=14\n",
    "fontsize_title=16\n",
    "\n",
    "x = np.arange(0.25, 1.6, 0.1)\n",
    "plt.plot(x, y_pred_epunc_de_oodl_50, alpha=0.5, c='#000080', label='DE')\n",
    "plt.fill_between(x, y_pred_epunc_de_oodl_16, y_pred_epunc_de_oodl_84,\n",
    "                facecolor='#000080', alpha=0.2)\n",
    "plt.plot(x, y_pred_epunc_cd_oodl_50, alpha=0.5, c='#800080', label='CD')\n",
    "plt.fill_between(x, y_pred_epunc_cd_oodl_16, y_pred_epunc_cd_oodl_84,\n",
    "                facecolor='#800080', alpha=0.2)\n",
    "plt.plot(x, y_pred_epunc_bnn_oodl_50, alpha=0.5, c='#FF0000', label='BNN')\n",
    "plt.fill_between(x, y_pred_epunc_bnn_oodl_16, y_pred_epunc_bnn_oodl_84,\n",
    "                facecolor='#FF0000', alpha=0.2)\n",
    "plt.axvline(0.2, ls='--', c='black', linewidth=1)\n",
    "plt.axvline(0.8, ls='--', c='black', linewidth=1, label='Training range')\n",
    "\n",
    "plt.legend(fontsize=fontsize, loc=2)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.xlabel('Input $L$ (m)', fontsize=fontsize)\n",
    "plt.ylabel('Epistemic uncertainty (m/s$^2$)', fontsize=fontsize)\n",
    "plt.tight_layout()\n",
    "plt.savefig('oodl.pdf', fmt='pdf', dpi=500)\n",
    "plt.savefig('oodl.png', fmt='png', dpi=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "residuals_de = y_pred_de_t1to20 - y_test_t1to20\n",
    "residuals_cd = y_pred_cd_t1to20 - y_test_t1to20\n",
    "residuals_bnn = y_pred_bnn_t1to20 - y_test_t1to20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "coverage_de = find_coverage(residuals_de, y_pred_prunc_de_t1to20)\n",
    "coverage_cd = find_coverage(residuals_cd, y_pred_prunc_cd_t1to20)\n",
    "coverage_bnn = find_coverage(residuals_bnn, y_pred_prunc_bnn_t1to20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_de_t1to20_8to12 = np.load('data/y_pred_test_de_trange1to20_40ep_001_ell8to12.npy')\n",
    "y_pred_prunc_de_t1to20_8to12 = np.load('data/y_pred_test_prunc_de_trange1to20_40ep_001_ell8to12.npy')\n",
    "\n",
    "y_pred_cd_t1to20_8to12 = np.load('data/y_pred_test_cd_trange1to20_200ep_001_ell8to12.npy')\n",
    "y_pred_prunc_cd_t1to20_8to12 = np.load('data/y_pred_test_prunc_cd_trange1to20_200ep_001_ell8to12.npy')\n",
    "\n",
    "y_pred_bnn_t1to20_8to12 = np.load('data/y_pred_test_bnn_trange1to20_200ep_001_ell8to12.npy')\n",
    "y_pred_prunc_bnn_t1to20_8to12 = np.load('data//y_pred_test_prunc_bnn_trange1to20_200ep_001_ell8to12.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_de_t1to20_12to16 = np.load('data/y_pred_test_de_trange1to20_40ep_001_ell12to16.npy')\n",
    "y_pred_prunc_de_t1to20_12to16 = np.load('data/y_pred_test_prunc_de_trange1to20_40ep_001_ell12to16.npy')\n",
    "\n",
    "y_pred_cd_t1to20_12to16 = np.load('data/y_pred_test_cd_trange1to20_200ep_001_ell12to16.npy')\n",
    "y_pred_prunc_cd_t1to20_12to16 = np.load('data/y_pred_test_prunc_cd_trange1to20_200ep_001_ell12to16.npy')\n",
    "\n",
    "y_pred_bnn_t1to20_12to16 = np.load('data/y_pred_test_bnn_trange1to20_200ep_001_ell12to16.npy')\n",
    "y_pred_prunc_bnn_t1to20_12to16 = np.load('data//y_pred_test_prunc_bnn_trange1to20_200ep_001_ell12to16.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_test_t1to20_8to12, y_test_t1to20_8to12, calc_y_test_t1to20_8to12, delta_y_test_t1to20_8to12 = pendulum(n=nobject, t_spread=[0.01, 0.2],\n",
    "                                                  ell_spread=[0.02, 0.02], seed=666, ell_range=[0.8, 1.2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "feat_test_t1to20_12to16, y_test_t1to20_12to16, calc_y_test_t1to20_12to16, delta_y_test_t1to20_12to16 = pendulum(n=nobject, t_spread=[0.01, 0.2],\n",
    "                                                  ell_spread=[0.02, 0.02], seed=666, ell_range=[1.2, 1.6])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "residuals_de_8to12 = y_pred_de_t1to20_8to12 - y_test_t1to20_8to12\n",
    "residuals_cd_8to12 = y_pred_cd_t1to20_8to12 - y_test_t1to20_8to12\n",
    "residuals_bnn_8to12 = y_pred_bnn_t1to20_8to12 - y_test_t1to20_8to12\n",
    "\n",
    "residuals_de_12to16 = y_pred_de_t1to20_12to16 - y_test_t1to20_12to16\n",
    "residuals_cd_12to16 = y_pred_cd_t1to20_12to16 - y_test_t1to20_12to16\n",
    "residuals_bnn_12to16 = y_pred_bnn_t1to20_12to16 - y_test_t1to20_12to16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "coverage_de_8to12 = find_coverage(residuals_de_8to12, y_pred_prunc_de_t1to20_8to12)\n",
    "coverage_cd_8to12 = find_coverage(residuals_cd_8to12, y_pred_prunc_cd_t1to20_8to12)\n",
    "coverage_bnn_8to12 = find_coverage(residuals_bnn_8to12, y_pred_prunc_bnn_t1to20_8to12)\n",
    "\n",
    "coverage_de_12to16 = find_coverage(residuals_de_12to16, y_pred_prunc_de_t1to20_12to16)\n",
    "coverage_cd_12to16 = find_coverage(residuals_cd_12to16, y_pred_prunc_cd_t1to20_12to16)\n",
    "coverage_bnn_12to16 = find_coverage(residuals_bnn_12to16, y_pred_prunc_bnn_t1to20_12to16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = np.linspace(0.01, 1, 100)\n",
    "\n",
    "fontsize=14\n",
    "fontsize_title=16\n",
    "\n",
    "plt.figure(figsize=(13,4))\n",
    "\n",
    "ax0 = plt.subplot(131)\n",
    "plt.plot(x, x, ls='--', c=(0,0,0))\n",
    "plt.plot(x, coverage_de, label='DE', c='#000080', alpha=0.5)\n",
    "plt.plot(x, coverage_cd, label='CD', c='#800080', alpha=0.5)\n",
    "plt.plot(x, coverage_bnn, label='BNN', c='#FF0000', alpha=0.5)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xlabel('Conf. int. size assuming normal dist.', fontsize=fontsize)\n",
    "plt.ylabel('Proportion of points inside interval', fontsize=fontsize)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('L in 0.2 to 0.8 m', fontsize=fontsize_title)\n",
    "\n",
    "ax0 = plt.subplot(132)\n",
    "plt.plot(x, x, ls='--', c=(0,0,0))\n",
    "plt.plot(x, coverage_de_8to12, label='DE', c='#000080', alpha=0.5)\n",
    "plt.plot(x, coverage_cd_8to12, label='CD', c='#800080', alpha=0.5)\n",
    "plt.plot(x, coverage_bnn_8to12, label='BNN', c='#FF0000', alpha=0.5)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xlabel('Conf. int. size assuming normal dist.', fontsize=fontsize)\n",
    "plt.setp(ax0.get_yticklabels(), visible=False)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('L in 0.8 to 1.2 m', fontsize=fontsize_title)\n",
    "plt.subplots_adjust(wspace=0., hspace=0.)\n",
    "\n",
    "ax0 = plt.subplot(133)\n",
    "plt.plot(x, x, ls='--', c=(0,0,0))\n",
    "plt.plot(x, coverage_de_12to16, label='DE', c='#000080', alpha=0.5)\n",
    "plt.plot(x, coverage_cd_12to16, label='CD', c='#800080', alpha=0.5)\n",
    "plt.plot(x, coverage_bnn_12to16, label='BNN', c='#FF0000', alpha=0.5)\n",
    "plt.legend(fontsize=fontsize)\n",
    "plt.xlabel('Conf. int. size assuming normal dist.', fontsize=fontsize)\n",
    "plt.setp(ax0.get_yticklabels(), visible=False)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.title('L in 1.2 to 1.6 m', fontsize=fontsize_title)\n",
    "plt.tight_layout()\n",
    "plt.savefig('calibration_changingell.pdf', fmt='pdf', dpi=500)\n",
    "plt.savefig('calibration_changingell.png', fmt='png', dpi=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_bnn_oodg = []\n",
    "y_pred_alunc_bnn_oodg = []\n",
    "y_pred_epunc_bnn_oodg = []\n",
    "y_pred_prunc_bnn_oodg = []\n",
    "for i in range(10, 25):\n",
    "    y_pred_bnn_oodg.append(np.load(f'data/y_pred_test_bnn_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_bnn_oodg.append(np.load(f'data/y_pred_test_alunc_bnn_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_bnn_oodg.append(np.load(f'data/y_pred_test_epunc_bnn_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_bnn_oodg.append(np.load(f'data/y_pred_test_prunc_bnn_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    \n",
    "y_pred_de_oodg = []\n",
    "y_pred_alunc_de_oodg = []\n",
    "y_pred_epunc_de_oodg = []\n",
    "y_pred_prunc_de_oodg = []\n",
    "for i in range(10, 25):\n",
    "    y_pred_de_oodg.append(np.load(f'data/y_pred_test_de_trange1to20_40ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_de_oodg.append(np.load(f'data/y_pred_test_alunc_de_trange1to20_40ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_de_oodg.append(np.load(f'data/y_pred_test_epunc_de_trange1to20_40ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_de_oodg.append(np.load(f'data/y_pred_test_prunc_de_trange1to20_40ep_001_g{i}to{i+1}.npy'))\n",
    "    \n",
    "y_pred_cd_oodg = []\n",
    "y_pred_alunc_cd_oodg = []\n",
    "y_pred_epunc_cd_oodg = []\n",
    "y_pred_prunc_cd_oodg = []\n",
    "for i in range(10, 25):\n",
    "    y_pred_cd_oodg.append(np.load(f'data/y_pred_test_cd_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_alunc_cd_oodg.append(np.load(f'data/y_pred_test_alunc_cd_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_epunc_cd_oodg.append(np.load(f'data/y_pred_test_epunc_cd_trange1to20_200ep_001_g{i}to{i+1}.npy'))\n",
    "    y_pred_prunc_cd_oodg.append(np.load(f'data/y_pred_test_prunc_cd_trange1to20_200ep_001_g{i}to{i+1}.npy'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_epunc_bnn_oodg_50 = np.percentile(y_pred_epunc_bnn_oodg, 50, axis=1)\n",
    "y_pred_epunc_cd_oodg_50 = np.percentile(y_pred_epunc_cd_oodg, 50, axis=1)\n",
    "y_pred_epunc_de_oodg_50 = np.percentile(y_pred_epunc_de_oodg, 50, axis=1)\n",
    "y_pred_epunc_bnn_oodg_16 = np.percentile(y_pred_epunc_bnn_oodg, 16, axis=1)\n",
    "y_pred_epunc_cd_oodg_16 = np.percentile(y_pred_epunc_cd_oodg, 16, axis=1)\n",
    "y_pred_epunc_de_oodg_16 = np.percentile(y_pred_epunc_de_oodg, 16, axis=1)\n",
    "y_pred_epunc_bnn_oodg_84 = np.percentile(y_pred_epunc_bnn_oodg, 84, axis=1)\n",
    "y_pred_epunc_cd_oodg_84 = np.percentile(y_pred_epunc_cd_oodg, 84, axis=1)\n",
    "y_pred_epunc_de_oodg_84 = np.percentile(y_pred_epunc_de_oodg, 84, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fontsize=14\n",
    "fontsize_title=16\n",
    "\n",
    "x = np.arange(10.5, 25, 1.)\n",
    "plt.plot(x, y_pred_epunc_de_oodg_50, alpha=0.5, c='#000080', label='DE')\n",
    "plt.fill_between(x, y_pred_epunc_de_oodg_16, y_pred_epunc_de_oodg_84,\n",
    "                facecolor='#000080', alpha=0.2)\n",
    "plt.plot(x, y_pred_epunc_cd_oodg_50, alpha=0.5, c='#800080', label='CD')\n",
    "plt.fill_between(x, y_pred_epunc_cd_oodg_16, y_pred_epunc_cd_oodg_84,\n",
    "                facecolor='#800080', alpha=0.2)\n",
    "plt.plot(x, y_pred_epunc_bnn_oodg_50, alpha=0.5, c='#FF0000', label='BNN')\n",
    "plt.fill_between(x, y_pred_epunc_bnn_oodg_16, y_pred_epunc_bnn_oodg_84,\n",
    "                facecolor='#FF0000', alpha=0.2)\n",
    "plt.axvline(15, ls='--', c='black', linewidth=1, label='Training range')\n",
    "\n",
    "plt.legend(fontsize=fontsize, loc=2)\n",
    "plt.xticks(fontsize=fontsize)\n",
    "plt.yticks(fontsize=fontsize)\n",
    "plt.xlabel('Output $g$ (m/s$^2$)', fontsize=fontsize)\n",
    "plt.ylabel('Epistemic uncertainty (m/s$^2$)', fontsize=fontsize)\n",
    "plt.tight_layout()\n",
    "plt.savefig('oodg.pdf', fmt='pdf', dpi=500)\n",
    "plt.savefig('oodg.png', fmt='png', dpi=500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigmoid(x):\n",
    "    return 1/(1+np.exp(-x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from models.cd import make_model\n",
    "model = make_model(13, 1, 100)\n",
    "model.load_weights('data/model_cd_trange1to20_200ep_001.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dropout probabilities:  [5.7058816e-05 1.2645748e-06 2.5880211e-06 1.3074545e-06 1.9308177e-06]\n"
     ]
    }
   ],
   "source": [
    "print('Dropout probabilities: ', np.array([sigmoid(layer.p_logit.numpy()) for layer in model.layers if hasattr(layer, 'p_logit')]).flatten())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "corr_cd_1to5 = np.corrcoef(delta_y_test_t1to5/calc_y_test_t1to5, y_pred_alunc_cd_t1to5/y_pred_cd_t1to5)[0, 1]\n",
    "corr_cd_1to10 = np.corrcoef(delta_y_test_t1to10/calc_y_test_t1to10, y_pred_alunc_cd_t1to10/y_pred_cd_t1to10)[0, 1]\n",
    "corr_cd_1to20 = np.corrcoef(delta_y_test_t1to20/calc_y_test_t1to20, y_pred_alunc_cd_t1to20/y_pred_cd_t1to20)[0, 1]\n",
    "corr_de_1to5 = np.corrcoef(delta_y_test_t1to5/calc_y_test_t1to5, y_pred_alunc_de_t1to5/y_pred_de_t1to5)[0, 1]\n",
    "corr_de_1to10 = np.corrcoef(delta_y_test_t1to10/calc_y_test_t1to10, y_pred_alunc_de_t1to10/y_pred_de_t1to10)[0, 1]\n",
    "corr_de_1to20 = np.corrcoef(delta_y_test_t1to20/calc_y_test_t1to20, y_pred_alunc_de_t1to20/y_pred_de_t1to20)[0, 1]\n",
    "corr_bnn_1to5 = np.corrcoef(delta_y_test_t1to5/calc_y_test_t1to5, y_pred_alunc_bnn_t1to5/y_pred_bnn_t1to5)[0, 1]\n",
    "corr_bnn_1to10 = np.corrcoef(delta_y_test_t1to10/calc_y_test_t1to10, y_pred_alunc_bnn_t1to10/y_pred_bnn_t1to10)[0, 1]\n",
    "corr_bnn_1to20 = np.corrcoef(delta_y_test_t1to20/calc_y_test_t1to20, y_pred_alunc_bnn_t1to20/y_pred_bnn_t1to20)[0, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.linspace(0.01, 1, 100)\n",
    "coverage_err_cd_2to8 = np.sum(np.abs([coverage_cd-x]))/100\n",
    "coverage_err_de_2to8 = np.sum(np.abs([coverage_de-x]))/100\n",
    "coverage_err_bnn_2to8 = np.sum(np.abs([coverage_bnn-x]))/100\n",
    "coverage_err_cd_8to12 = np.sum(np.abs([coverage_cd_8to12-x]))/100\n",
    "coverage_err_de_8to12 = np.sum(np.abs([coverage_de_8to12-x]))/100\n",
    "coverage_err_bnn_8to12 = np.sum(np.abs([coverage_bnn_8to12-x]))/100\n",
    "coverage_err_cd_12to16 = np.sum(np.abs([coverage_cd_12to16-x]))/100\n",
    "coverage_err_de_12to16 = np.sum(np.abs([coverage_de_12to16-x]))/100\n",
    "coverage_err_bnn_12to16 = np.sum(np.abs([coverage_bnn_12to16-x]))/100"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "mse_cd = mean_squared_error(y_test_t1to20, y_pred_cd_t1to20)\n",
    "mse_de = mean_squared_error(y_test_t1to20, y_pred_de_t1to20)\n",
    "mse_bnn = mean_squared_error(y_test_t1to20, y_pred_bnn_t1to20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " & 1-5\\% & 1-10\\% & 1-20\\% &  & (0.2, 0.8) m & (0.8, 1.2) m & (1.2, 1.6) m & \\\\\n",
      "DE & 0.00 & 0.79 & 0.84 &  & 0.031 & 0.019 & 0.055 & 0.49 \\\\\n",
      "CD & 0.00 & 0.83 & 0.84 &  & 0.011 & 0.20 & 0.39 & 0.51 \\\\\n",
      "BNN & 0.00 & -0.00 & 0.76 &  & 0.023 & 0.014 & 0.137 & 0.51 \\\\\n"
     ]
    }
   ],
   "source": [
    "print(' & '.join(['', '1-5\\%', '1-10\\%', '1-20\\%', '', '(0.2, 0.8) m', '(0.8, 1.2) m', '(1.2, 1.6) m', '\\\\\\\\']))\n",
    "print(' & '.join(['DE', '%.2f' % corr_de_1to5, '%.2f' % corr_de_1to10, '%.2f' % corr_de_1to20, '', '%.3f' % coverage_err_de_2to8, \n",
    "                  '%.3f' % coverage_err_de_8to12, '%.3f' % coverage_err_de_12to16, '%.2f \\\\\\\\' % mse_de]))\n",
    "print(' & '.join(['CD', '%.2f' % corr_cd_1to5, '%.2f' % corr_cd_1to10, '%.2f' % corr_cd_1to20, '', '%.3f' % coverage_err_cd_2to8, \n",
    "                  '%.2f' % coverage_err_cd_8to12, '%.2f' % coverage_err_cd_12to16, '%.2f \\\\\\\\' % mse_cd]))\n",
    "print(' & '.join(['BNN', '%.2f' % corr_bnn_1to5, '%.2f' % corr_bnn_1to10, '%.2f' % corr_bnn_1to20, '', '%.3f' % coverage_err_bnn_2to8, \n",
    "                  '%.3f' % coverage_err_bnn_8to12, '%.3f' % coverage_err_bnn_12to16, '%.2f \\\\\\\\' % mse_bnn]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}