Simulation/run_3D_simulation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run Simulation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import os.path\n",
    "from os.path import dirname\n",
    "import numpy as np\n",
    "from ROIAL_simulation import run_simulation\n",
    "file_dir = dirname(os.getcwd())\n",
    "root_directory = dirname(dirname(dirname(file_dir))) \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## online learning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# initialize hyperparameters\n",
    "feedback_type = ['ord','pref']\n",
    "model_params = {'ord_nos':0,'pref_nos':0.06,'signal_variance': 1.0, 'lengthscales': [0.15, 0.15,0.15],'GP_noise_var': 0.01,\n",
    "             'IG_it':1000,'query_type':'IG','lambda': 0.0, 'saf_method': 'none',\n",
    "             'pref_noise':0.015,'num_category':5, 'ordinal_noise':0.1}\n",
    "\n",
    "#define the ordinal threshold the algorithm uses for the simulation\n",
    "delta_int = 2/model_params['num_category'] \n",
    "ordinal_threshold_estimate = np.array([delta_int]*(model_params['num_category']-1))\n",
    "ordinal_threshold_estimate[0] = -0.5\n",
    "model_params['ord_threshold_estimate'] = ordinal_threshold_estimate\n",
    "\n",
    "model_params['saf_method'] = 'IG_ucb' #default is none and 0, which means the ROI is the whole action space\n",
    "model_params['lambda'] = 1\n",
    "\n",
    "sub_params = {'model_params':model_params,'rd_sz':500,'run_nums': range(1),'num_trials':80,'D':3,'num_category':5,'ord_perct':False,\n",
    "               'ord_b1':0.17,'ord_delta':[0.2,0.15,0.13],'file_part': file_dir + '/Simulation/Sampled_functions_3D/Sampled_objective_'}\n",
    "save_folder = root_directory + '/Results/' + '_'.join(feedback_type) + '_RD_' + str(sub_params['rd_sz']) +'/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ordinal threshold true:  [-0.75498344 -0.40224755  0.01273585  0.3239734   0.59371261  1.31993357]\n",
      "8000\n",
      "iteration:  1\n",
      "4493\n",
      "start rd\n",
      "(501, 3)\n",
      "safety set size 501\n",
      "sampled action:  [ 7  0 14]\n",
      "iteration:  2\n",
      "2814\n",
      "start rd\n",
      "(502, 3)\n",
      "safety set size 502\n",
      "sampled action:  [17  3 14]\n",
      "iteration:  3\n",
      "6874\n",
      "start rd\n",
      "(503, 3)\n",
      "safety set size 503\n",
      "sampled action:  [12  7 17]\n",
      "iteration:  4\n",
      "4957\n",
      "start rd\n",
      "(504, 3)\n",
      "safety set size 504\n",
      "sampled action:  [ 8  9 13]\n",
      "iteration:  5\n",
      "3393\n",
      "start rd\n",
      "(505, 3)\n",
      "safety set size 505\n",
      "sampled action:  [ 5 12 12]\n",
      "iteration:  6\n",
      "2252\n",
      "start rd\n",
      "(506, 3)\n",
      "safety set size 506\n",
      "sampled action:  [19  1 11]\n",
      "iteration:  7\n",
      "7631\n",
      "start rd\n",
      "(507, 3)\n",
      "safety set size 507\n",
      "sampled action:  [2 1 1]\n",
      "iteration:  8\n",
      "821\n",
      "start rd\n",
      "(508, 3)\n",
      "safety set size 508\n",
      "sampled action:  [19  7 12]\n",
      "iteration:  9\n",
      "7752\n",
      "start rd\n",
      "(509, 3)\n",
      "safety set size 509\n",
      "sampled action:  [0 9 1]\n",
      "iteration:  10\n",
      "181\n",
      "start rd\n",
      "(510, 3)\n",
      "safety set size 510\n",
      "sampled action:  [0 6 3]\n",
      "iteration:  11\n",
      "123\n",
      "start rd\n",
      "(511, 3)\n",
      "safety set size 511\n",
      "sampled action:  [6 9 8]\n",
      "iteration:  12\n",
      "2588\n",
      "start rd\n",
      "(512, 3)\n",
      "safety set size 512\n",
      "sampled action:  [5 4 8]\n",
      "iteration:  13\n",
      "2088\n",
      "start rd\n",
      "(513, 3)\n",
      "safety set size 513\n",
      "sampled action:  [5 9 3]\n",
      "iteration:  14\n",
      "2183\n",
      "start rd\n",
      "(514, 3)\n",
      "safety set size 514\n",
      "sampled action:  [ 2  3 12]\n",
      "iteration:  15\n",
      "872\n",
      "start rd\n",
      "(515, 3)\n",
      "safety set size 515\n",
      "sampled action:  [ 1  5 16]\n",
      "iteration:  16\n",
      "516\n",
      "start rd\n",
      "(516, 3)\n",
      "safety set size 516\n",
      "sampled action:  [11  6  9]\n",
      "iteration:  17\n",
      "4529\n",
      "start rd\n",
      "(517, 3)\n",
      "safety set size 517\n",
      "sampled action:  [14  6  3]\n",
      "iteration:  18\n",
      "5723\n",
      "start rd\n",
      "(518, 3)\n",
      "safety set size 518\n",
      "sampled action:  [10  3  5]\n",
      "iteration:  19\n",
      "4065\n",
      "start rd\n",
      "(519, 3)\n",
      "safety set size 519\n",
      "sampled action:  [ 5  3 18]\n",
      "iteration:  20\n",
      "2078\n",
      "start rd\n",
      "(520, 3)\n",
      "safety set size 520\n",
      "sampled action:  [ 1 12  8]\n",
      "iteration:  21\n",
      "648\n",
      "start rd\n",
      "(521, 3)\n",
      "safety set size 521\n",
      "sampled action:  [ 3 14  4]\n",
      "iteration:  22\n",
      "1484\n",
      "start rd\n",
      "(522, 3)\n",
      "safety set size 522\n",
      "sampled action:  [ 1  9 12]\n",
      "iteration:  23\n",
      "592\n",
      "start rd\n",
      "(523, 3)\n",
      "safety set size 523\n",
      "sampled action:  [13  0  8]\n",
      "iteration:  24\n",
      "5208\n",
      "start rd\n",
      "(524, 3)\n",
      "safety set size 524\n",
      "sampled action:  [ 6 14  9]\n",
      "iteration:  25\n",
      "2689\n",
      "start rd\n",
      "(525, 3)\n",
      "safety set size 525\n",
      "sampled action:  [ 9 13  3]\n",
      "iteration:  26\n",
      "3863\n",
      "start rd\n",
      "(526, 3)\n",
      "safety set size 526\n",
      "sampled action:  [11 12  8]\n",
      "iteration:  27\n",
      "4648\n",
      "start rd\n",
      "(527, 3)\n",
      "safety set size 527\n",
      "sampled action:  [17 15  5]\n",
      "iteration:  28\n",
      "7105\n",
      "start rd\n",
      "(528, 3)\n",
      "safety set size 528\n",
      "sampled action:  [ 2  9 18]\n",
      "iteration:  29\n",
      "998\n",
      "start rd\n",
      "(529, 3)\n",
      "safety set size 529\n",
      "sampled action:  [ 6  9 19]\n",
      "iteration:  30\n",
      "2599\n",
      "start rd\n",
      "(530, 3)\n",
      "safety set size 530\n",
      "sampled action:  [ 0 19  7]\n",
      "iteration:  31\n",
      "387\n",
      "start rd\n",
      "(531, 3)\n",
      "safety set size 531\n",
      "sampled action:  [15  2  0]\n",
      "iteration:  32\n",
      "6040\n",
      "start rd\n",
      "(532, 3)\n",
      "safety set size 532\n",
      "sampled action:  [19 19  2]\n",
      "iteration:  33\n",
      "7982\n",
      "start rd\n",
      "(533, 3)\n",
      "safety set size 533\n",
      "sampled action:  [ 7 18  4]\n",
      "iteration:  34\n",
      "3164\n",
      "start rd\n",
      "(534, 3)\n",
      "safety set size 534\n",
      "sampled action:  [18 13  8]\n",
      "iteration:  35\n",
      "7468\n",
      "start rd\n",
      "(535, 3)\n",
      "safety set size 535\n",
      "sampled action:  [13 16  2]\n",
      "iteration:  36\n",
      "5522\n",
      "start rd\n",
      "(536, 3)\n",
      "safety set size 536\n",
      "sampled action:  [18 19 17]\n",
      "iteration:  37\n",
      "7597\n",
      "start rd\n",
      "(537, 3)\n",
      "safety set size 537\n",
      "sampled action:  [16 15 17]\n",
      "iteration:  38\n",
      "6717\n",
      "start rd\n",
      "(538, 3)\n",
      "safety set size 538\n",
      "sampled action:  [ 8 19 15]\n",
      "iteration:  39\n",
      "3595\n",
      "start rd\n",
      "(539, 3)\n",
      "safety set size 539\n",
      "sampled action:  [ 4 13 17]\n",
      "iteration:  40\n",
      "1877\n",
      "start rd\n",
      "(540, 3)\n",
      "safety set size 540\n",
      "sampled action:  [ 7 19 19]\n",
      "iteration:  41\n",
      "3199\n",
      "start rd\n",
      "(541, 3)\n",
      "safety set size 541\n",
      "sampled action:  [19 18  7]\n",
      "iteration:  42\n",
      "7967\n",
      "start rd\n",
      "(542, 3)\n",
      "safety set size 542\n",
      "sampled action:  [ 8 16 12]\n",
      "iteration:  43\n",
      "3532\n",
      "start rd\n",
      "(543, 3)\n",
      "safety set size 543\n",
      "sampled action:  [ 2 16  8]\n",
      "iteration:  44\n",
      "1128\n",
      "start rd\n",
      "(544, 3)\n",
      "safety set size 544\n",
      "sampled action:  [18  1  6]\n",
      "iteration:  45\n",
      "7226\n",
      "start rd\n",
      "(545, 3)\n",
      "safety set size 545\n",
      "sampled action:  [6 2 0]\n",
      "iteration:  46\n",
      "2440\n",
      "start rd\n",
      "(546, 3)\n",
      "safety set size 546\n",
      "sampled action:  [ 2 16  0]\n",
      "iteration:  47\n",
      "1120\n",
      "start rd\n",
      "(547, 3)\n",
      "safety set size 547\n",
      "sampled action:  [17  8  0]\n",
      "iteration:  48\n",
      "6960\n",
      "start rd\n",
      "(548, 3)\n",
      "safety set size 548\n",
      "sampled action:  [17  6 19]\n",
      "iteration:  49\n",
      "6939\n",
      "start rd\n",
      "(549, 3)\n",
      "safety set size 549\n",
      "sampled action:  [ 2 18 19]\n",
      "iteration:  50\n",
      "1179\n",
      "start rd\n",
      "(550, 3)\n",
      "safety set size 550\n",
      "sampled action:  [ 9 18  8]\n",
      "iteration:  51\n",
      "3968\n",
      "start rd\n",
      "(551, 3)\n",
      "safety set size 551\n",
      "sampled action:  [18  6  5]\n",
      "iteration:  52\n",
      "7325\n",
      "start rd\n",
      "(552, 3)\n",
      "safety set size 552\n",
      "sampled action:  [12  9  0]\n",
      "iteration:  53\n",
      "4980\n",
      "start rd\n",
      "(553, 3)\n",
      "safety set size 553\n",
      "sampled action:  [13 18 13]\n",
      "iteration:  54\n",
      "5573\n",
      "start rd\n",
      "(554, 3)\n",
      "safety set size 554\n",
      "sampled action:  [12 11 15]\n",
      "iteration:  55\n",
      "5035\n",
      "start rd\n",
      "(555, 3)\n",
      "safety set size 555\n",
      "sampled action:  [16  0 18]\n",
      "iteration:  56\n",
      "6418\n",
      "start rd\n",
      "(556, 3)\n",
      "safety set size 556\n",
      "sampled action:  [1 1 5]\n",
      "iteration:  57\n",
      "425\n",
      "start rd\n",
      "(557, 3)\n",
      "safety set size 557\n",
      "sampled action:  [19 15 11]\n",
      "iteration:  58\n",
      "7911\n",
      "start rd\n",
      "(558, 3)\n",
      "safety set size 558\n",
      "sampled action:  [13 12 19]\n",
      "iteration:  59\n",
      "5459\n",
      "start rd\n",
      "(559, 3)\n",
      "safety set size 559\n",
      "sampled action:  [7 8 0]\n",
      "iteration:  60\n",
      "2960\n",
      "start rd\n",
      "(560, 3)\n",
      "safety set size 560\n",
      "sampled action:  [ 9 19  0]\n",
      "iteration:  61\n",
      "3980\n",
      "start rd\n",
      "(561, 3)\n",
      "safety set size 561\n",
      "sampled action:  [18  9 17]\n",
      "iteration:  62\n",
      "7397\n",
      "start rd\n",
      "(562, 3)\n",
      "safety set size 562\n",
      "sampled action:  [ 5 19  8]\n",
      "iteration:  63\n",
      "2388\n",
      "start rd\n",
      "(563, 3)\n",
      "safety set size 563\n",
      "sampled action:  [11  1 19]\n",
      "iteration:  64\n",
      "4439\n",
      "start rd\n",
      "(564, 3)\n",
      "safety set size 564\n",
      "sampled action:  [ 0 14 14]\n",
      "iteration:  65\n",
      "294\n",
      "start rd\n",
      "(565, 3)\n",
      "safety set size 565\n",
      "sampled action:  [14 12  4]\n",
      "iteration:  66\n",
      "5844\n",
      "start rd\n",
      "(566, 3)\n",
      "safety set size 566\n",
      "sampled action:  [12 16 18]\n",
      "iteration:  67\n",
      "5138\n",
      "start rd\n",
      "(567, 3)\n",
      "safety set size 567\n",
      "sampled action:  [ 4 18 12]\n",
      "iteration:  68\n",
      "1972\n",
      "start rd\n",
      "(568, 3)\n",
      "safety set size 568\n",
      "sampled action:  [15 12 12]\n",
      "iteration:  69\n",
      "6252\n",
      "start rd\n",
      "(569, 3)\n",
      "safety set size 569\n",
      "sampled action:  [0 1 9]\n",
      "iteration:  70\n",
      "29\n",
      "start rd\n",
      "(570, 3)\n",
      "safety set size 570\n",
      "sampled action:  [17 16  1]\n",
      "iteration:  71\n",
      "7121\n",
      "start rd\n",
      "(571, 3)\n",
      "safety set size 571\n",
      "sampled action:  [6 0 8]\n",
      "iteration:  72\n",
      "2408\n",
      "start rd\n",
      "(572, 3)\n",
      "safety set size 572\n",
      "sampled action:  [14 14 10]\n",
      "iteration:  73\n",
      "5890\n",
      "start rd\n",
      "(573, 3)\n",
      "safety set size 573\n",
      "sampled action:  [ 0 19 17]\n",
      "iteration:  74\n",
      "397\n",
      "start rd\n",
      "(574, 3)\n",
      "safety set size 574\n",
      "sampled action:  [ 0 19  2]\n",
      "iteration:  75\n",
      "382\n",
      "start rd\n",
      "(575, 3)\n",
      "safety set size 575\n",
      "sampled action:  [19 12  0]\n",
      "iteration:  76\n",
      "7840\n",
      "start rd\n",
      "(576, 3)\n",
      "safety set size 576\n",
      "sampled action:  [19  0 14]\n",
      "iteration:  77\n",
      "7614\n",
      "start rd\n",
      "(577, 3)\n",
      "safety set size 577\n",
      "sampled action:  [19  1  2]\n",
      "iteration:  78\n",
      "7622\n",
      "start rd\n",
      "(578, 3)\n",
      "safety set size 578\n",
      "sampled action:  [15  9  7]\n",
      "iteration:  79\n",
      "6187\n",
      "start rd\n",
      "(579, 3)\n",
      "safety set size 579\n",
      "sampled action:  [10  0  0]\n"
     ]
    }
   ],
   "source": [
    "\n",
    "run_simulation(root_directory,feedback_type,sub_params,save_folder)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## post-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "postprocessing run 0\n",
      "initiate cov matrix\n",
      "80\n"
     ]
    }
   ],
   "source": [
    "from post_eval_sub_GP import eval_posterior\n",
    "num_trials = 80\n",
    "# you can also pass path = covariance to reuse a covariance matrix file you pregenerated\n",
    "eval_posterior(save_folder,model_params,state_dim =3,num_trials = num_trials,num_runs = range(1),feedback_type = ['ord','pref'],spec_trial_only = True)\n",
    "#set spec_trial_only = True means only the iteration num_trials would be post-processed"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualize Results\n",
    "To visualize 3D function, we average the values over the third dimension"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 612x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from plot_3d import plot_3d_function\n",
    "%matplotlib inline\n",
    "obj_folder = '/Users/amyli/Desktop/CalTech/gitrepo/ROIAL/Simulation/Sampled_functions_3D/'\n",
    "plot_3d_function(D = 3, run_num = 0, iter_list = [num_trials], save_folder = save_folder, obj_folder = sub_params['file_part'])\n",
    "#iter_list should match the specific iterations being post_processed\n",
    "# for instance, if posterior was evaluated at iterations 1, 10, 20 at the previous step, you should pass iter_list = [1,10,20]"
   ]
  },
  {
   "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
}