Demo1 human feedback for battery experiments.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6095676d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import sys\n",
    "sys.path.append(os.path.join(os.getcwd(), '..'))\n",
    "import time\n",
    "import torch\n",
    "import pickle\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "321a7ddd",
   "metadata": {},
   "source": [
    "# 1. Problem definition\n",
    "## features (dimensions)\n",
    "1. Molarity of LiPF6, from 0.2 to 2.1 (M)\n",
    "2. Mass fraction of Ethylene Carbonate (EC) cosolvents of all solvents, from 0.3 to 0.52\n",
    "3. DMC/EMC cosolvents ratio. from 0 to 1. 0 is pure EMC, 1 is pure DMC.\n",
    "\n",
    "## Objective\n",
    "- Lithium-ion conductivity (mS/cm). Larger is better"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6223d252",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[KeOps] Warning : Cuda libraries were not detected on the system ; using cpu only mode\n"
     ]
    }
   ],
   "source": [
    "from experiments._electrolyte import ElectrolyteSearch\n",
    "from CoExBO._prior import Uniform\n",
    "\n",
    "path_data = \"./experiments/AEM_training_data.csv\"\n",
    "path_prior = \"./experiments/electrolyte_prior.pickle\"\n",
    "true_function = ElectrolyteSearch(path_data, path_prior)\n",
    "domain = Uniform(true_function.bounds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "0938db78",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['LiPF6', 'EC', 'DMC']"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_function.feature_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "ef8cca54",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0.2000, 0.3000, 0.0000],\n",
       "        [2.1000, 0.5200, 1.0000]], dtype=torch.float64)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "true_function.bounds"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23b24407",
   "metadata": {},
   "source": [
    "## 2. Define experimental conditions\n",
    "Next, we define the experimental conditions.<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "cbafa1d2",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_iterations = 5      # number of iterations\n",
    "n_init_obj = 5        # number of initial random samples for objective function\n",
    "n_init_pref = 100     # number of initial random samples for preferential learning\n",
    "seed = 0              # random seed\n",
    "file_save_name = 'electrolyte_masaki_trial1.pickle' # file save name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "802fee2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "from CoExBO._coexbo import CoExBO, StateManager\n",
    "\n",
    "#>>> initial setting\n",
    "torch.manual_seed(seed)          # random seed\n",
    "state = StateManager(n_dims=3)   # state parameter manager\n",
    "coexbo = CoExBO(\n",
    "    domain,                      # domain\n",
    "    true_function,               # true function (In reality, we run physical experiments.)\n",
    "    true_function.feature_names, # feature names\n",
    "    noisy=True,                  # Is observation noisy?\n",
    "    explanation=True,\n",
    ")\n",
    "dataset_duel = true_function.prior_duel\n",
    "\n",
    "# Initial run\n",
    "X = coexbo.domain.sample(n_init_obj)\n",
    "Y, Y_true = coexbo.true_function(X.squeeze())\n",
    "dataset_obj = (X, Y, Y_true)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "012666d8",
   "metadata": {},
   "source": [
    "# 3. Collaborative Experiments!\n",
    "\n",
    "### Rules\n",
    "- You will get the pair of points (X0, X1).\n",
    "- X0 is generated by the preference-based model.\n",
    "- X1 is generated by the standard UCB. Namely, normal BO.\n",
    "- You need to answer which you think larger.\n",
    "- Type \"0\" if you think f_true(X0) > f_true(X1).\n",
    "- Otherwise type \"1\".\n",
    "- If you are unsure, type \"2\". Then, X1 is selected but it makes preference model remain unsure at the given pairwise candidate location. \n",
    "- We will iterate this back and forth at $n_{iterations}$ times\n",
    "\n",
    "### Role of explanations\n",
    "- The explanation is for demystifying the GP acquisition process, rather than giving \"oracle\" explanation guiding to the true global optimum.\n",
    "- As such, YOU need to know the blackbox function better than GP. You should keep in mind the explanation is just advice, and you need to select based on your internal belief.\n",
    "- Note that the early-stage GP is unreliable. All explanation is based on GP, so the early-stage explanation is not a good adviser. It can be useful to rectify your over-trust over GP. \n",
    "- Conversely, the later round GP is reliable. You can use the explanation for fine-tune your understanding.\n",
    "\n",
    "### Explanation features\n",
    "- You can ask more information by answering \"y\" when system asked if you need more explanation. Otherwise type\"n\" so that you can skip the explanation.\n",
    "- The first feature is the simple GP prediction. You can check how GP predictive mean, variance, and UCB are different for each candidate. In early stage, mean is just unreliable. Uncertainty will give insights how likely you can get the information.\n",
    "- The second feature is the Shapley value. This gives the importance of input dimension to the prediction at the given candidate. This example, Branin function, is a uncorrelated dimensions so the two dimensions should be equal contribution. If the Shapley values give biased feature importance, GP is likely to be misspecified (particularly in the early stages.).\n",
    "- The third feature is the visualisation of each candidate. It will automatically visualise the 2 dimensional plot of GP predictive mean and variance based on top 2 largest Shapley values. You can check the geographical relationship between the candidates and the best observed point.\n",
    "- The fourth feature is the visualisation of preference-based GP predictive mean and variance. This is the mixture of human preference and surroagte GP model so it is not the pure preference-based plot.\n",
    "- The last feature is the feeback of your selection. The post-hoc analysis on the probability of correct answer will be returned."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b83ca3f0",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0) parameters: beta 8.485e-01 gamma 0.000e+00\n",
      "training models...\n",
      "generating candidates...\n",
      "X0 (preference): tensor([1.0298, 0.4718, 1.0000])\n",
      "X1 (normal UCB): tensor([1.0661, 0.3000, 0.6981])\n",
      "Do you need more explanation on the two candidates?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAukAAAF0CAYAAACe1KcLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAnR0lEQVR4nO3deXRU9f3/8deEJJMETGTNAoGEqAEVQUItBGvYORVpccPdACniAiIiNLF82VqlWqm1ngCCIi4sLqBWVEAxQStpCTWoFKxgQUNLVFCSlLIk4fP7g99MHROSTDKT+UzyfJwz5zB37tx53w/3/ckrM3duHMYYIwAAAADWCAl0AQAAAAA8EdIBAAAAyxDSAQAAAMsQ0gEAAADLENIBAAAAyxDSAQAAAMsQ0gEAAADLENIBAAAAyxDSAQAAAMsQ0puBjz/+WFlZWUpJSVFkZKQiIyN17rnnatKkSdq+fbt7vblz58rhcLhv4eHhSk5O1tSpU3XkyJE6X2fQoEG68MILa3zs0KFDcjgcmjt3bo31jR8/XsnJyYqIiFCbNm3Ut29fPfzww/r22289tv/9+sLCwpSUlKSsrCx98cUXXo8LWp769oLU+H6wQVJSksaNG+e+n5+fL4fDofz8fK+2s3XrVs2dO7fG/R40aJAGDRrUqDqBpsIc0LLmgAcffFCvvvpqvddfsWKFHA6H9u/f77eafCk00AWgcZ544glNnjxZqampmjp1qi644AI5HA7t3r1bq1ev1o9+9CPt3btXKSkp7uds2LBBMTExKi8v15tvvqnHHntM27Zt09atW+VwOHxa37Jly3TnnXcqNTVVM2bM0Pnnn6+Kigpt375dS5YsUUFBgV555RX3+t27d9fKlSslSSdPntTOnTs1b948vf322/r0008VFRXl0/rQfDSkF6Sm7Qd/69u3rwoKCnT++ed79bytW7dq3rx5GjdunM4++2yPxxYtWuTDCgH/YQ5oeXPAgw8+qGuuuUZjxowJdCl+QUgPYh988IHuvPNOjRo1Si+//LLCw8Pdjw0ZMkR33XWXXnrpJUVGRno8Ly0tTR06dJAkDR8+XIcPH9Zzzz2nrVu3auDAgT6rr6CgQHfccYeGDx+uV199VU6n0/3Y8OHDNX36dG3YsMHjOZGRkerfv7/7/mWXXaaIiAhlZWXpz3/+s0aMGOGz+tB8NLQXpKbrh+/773//65dfOKOjoz36xxe8/WEPBAJzwGnMAc0Lp7sEsQcffFCtWrXSE0884TEhfd+1116rhISEWrfjamhfn1Ly4IMPyuFwaOnSpR4B3SU8PFw/+9nP6txOTEyMJCksLMyn9aH58FUvSPXrB9dHys8//7zuvfdexcXFKTIyUhkZGSoqKvJYd9y4cWrTpo0++eQTjRgxQmeddZaGDh0q6fSnRb/5zW/Uo0cPOZ1OdezYUePHj9c333zjsY2KigrNnDlTcXFxioqK0qWXXqpt27adsa4fftT917/+VaNHj1b79u0VERGhlJQU3XPPPZJOf+Q/Y8YMSVJycrL7o3/XNr7/UXdFRYU6deqkW265pdprHzlyRJGRkbr33nvdy8rKynTfffcpOTlZ4eHh6ty5s+655x4dPXr0jGPr4jq9rqCgQOnp6YqMjFRSUpKefvppSdIbb7yhvn37KioqSr169ar2C78k7dmzRzfeeKM6deokp9Opnj17Kjc312Od48ePa/r06erTp49iYmLUrl07DRgwQK+99lq17TkcDk2ePFnPPfecevbsqaioKPXu3Vvr16+vc3/gX8wBnnU1hzmgqKhIV1xxhbt/ExISNGrUKB04cEDS6X48evSonnnmGXfN3z8t5y9/+YsGDhyoiIgIJSQkKCcnRxUVFXW+rk14Jz1IVVVVKS8vT/369VN8fHyjtrV3715JUseOHX1RmqTT9b377rtKS0tTYmKiV8+trKyU9L/TXebPn6/u3bsrPT3dZ/Wh+fBlL0je9cP999+vvn376sknn1Rpaanmzp2rQYMGqaioSN27d3evd/LkSf3sZz/TpEmTlJ2drcrKSp06dUo///nP9f7772vmzJlKT0/XF198oTlz5mjQoEHavn27+12/iRMn6tlnn9V9992n4cOHa+fOnbrqqqtUXl5eZ40bN27U6NGj1bNnT/3+979X165dtX//fm3atEmS9Itf/ELffvutHn/8ca1bt849hjW9exYWFqabb75ZS5YsUW5urqKjo92PrV69WsePH9f48eMlnX6nMCMjQwcOHND999+viy66SH//+981e/ZsffLJJ3rnnXfqPJWgpKRE48eP18yZM9WlSxc9/vjjmjBhgoqLi/Xyyy/r/vvvV0xMjObPn68xY8bon//8pzuE7dq1S+np6eratasWLlyouLg4bdy4UXfffbcOHTqkOXPmSJJOnDihb7/9Vvfdd586d+6skydP6p133tFVV12lp59+WrfeeqtHTW+88YYKCws1f/58tWnTRg8//LCuvPJK/eMf//D4P0fTYQ6oXTDOAUePHtXw4cOVnJys3NxcxcbGqqSkRHl5ee59Ligo0JAhQzR48GD93//9nyS569m1a5eGDh2qpKQkrVixQlFRUVq0aJFWrVpV53hZxSAolZSUGEnm+uuvr/ZYZWWlqaiocN9OnTpljDFmzpw5RpIpKSkxFRUV5rvvvjPPP/+8iYyMNImJiebYsWO1vmZGRoa54IILanzsm2++MZLMnDlz6qyvtu1LqnY777zzzO7du+u9HbQsDekFYxrXD3l5eUaS6du3r8c29+/fb8LCwswvfvEL97LMzEwjySxfvtxjG6tXrzaSzNq1az2WFxYWGklm0aJFxhhjdu/ebSSZadOmeay3cuVKI8lkZmZWqysvL8+9LCUlxaSkpNS6P7/73e+MJLNv375qj2VkZJiMjAz3/Y8//thIMkuXLvVY75JLLjFpaWnu+wsWLDAhISGmsLDQY72XX37ZSDJvvvnmGetxva4ks337dveyw4cPm1atWpnIyEjzr3/9y718x44dRpL54x//6F42cuRI06VLF1NaWuqx3cmTJ5uIiAjz7bff1vi6rmMmKyvLXHzxxR6PSTKxsbGmrKzMvaykpMSEhISYBQsW1Lo/8B/mgMxqdQX7HLB9+3Yjybz66qtnXMcYY1q3bu2x/y7XXXediYyMNCUlJe5llZWVpkePHmfcTxtxukszlJaWprCwMPdt4cKFHo/HxcUpLCxMbdu21c0336y+fftqw4YNioiI0KlTp1RZWem+VVVVNWntKSkpKiwsVGFhoQoKCrRq1SpFRkZq6NCh2rNnT5PWguBXVy9ItfdDXW688UaPd4K6deum9PR05eXlVVv36quv9ri/fv16nX322Ro9erRHz/Xp00dxcXHuj5pd27rppps8nj927FiFhtb+Yehnn32mzz//XFlZWfXan/ro1auX0tLS3KedSNLu3bu1bds2TZgwwWP/LrzwQvXp08dj/0aOHFnvq0/Ex8crLS3Nfb9du3bq1KmT+vTp43HaQs+ePSX97/SE48ePa/PmzbryyisVFRXl8fqXX365jh8/rr/85S/u57/00ksaOHCg2rRpo9DQUIWFhempp57S7t27q9U0ePBgnXXWWe77sbGx6tSpE1egshRzQHDOAeecc47atm2rX/7yl1qyZIl27drlVY15eXkaOnSoYmNj3ctatWql6667zqvtBBqnuwSpDh06KDIyssYfDKtWrdJ///tfHTx4sMZzvt955x3FxMQoLCxMXbp0Ufv27d2PzZ8/X/PmzXPf79atm/tSRaGhoWcM7a5TVFznjXfo0EFRUVHat2+fV/sVERGhfv36ue/3799fgwYNUufOnTV79mytXr3aq+2h+WtML0i190Nd4uLialz20UcfeSyLiory+FhYkr766isdOXLkjOfPHjp0SJJ0+PDhGl8rNDS0zlpd57V26dKl1vW8NWHCBN1111369NNP1aNHDz399NNyOp264YYb3Ot89dVX2rt37xm/S+Lav9q0a9eu2rLw8PBqy11jePz4cUmnx6yyslKPP/64Hn/88Vpff926dRo7dqyuvfZazZgxQ3FxcQoNDdXixYu1fPnyas+racydTqeOHTtW5/7AP5gDzixY54CYmBht2bJFDzzwgO6//3599913io+P18SJEzVr1qw6v6N2+PDhM/7fBBNCepBq1aqVhgwZok2bNungwYMe5+G5ziM703VAe/fu7f4m+w/ddtttuuKKK9z3v/+Fz9jYWBUWFsoYU+08sn/961/udVz1DR06VG+99ZYOHDjQqAkiPj5eHTp0qDbpAVLjekGqvR/qUlJSUuOyH/7grOm8yw4dOqh9+/Y1fuFRkvvdWte2SkpK1LlzZ/fjlZWV7h/eZ+I6p9b1RStfueGGG3TvvfdqxYoVeuCBB/Tcc89pzJgxatu2rXsdV3CqKei6HveXtm3bqlWrVrrlllt011131bhOcnKyJOn5559XcnKyXnjhBY//pxMnTvitPvgWc8CZBfMc0KtXL61Zs0bGGH388cdasWKF5s+fr8jISGVnZ9f63Pbt25/x/yaYcLpLEMvJyVFVVZVuv/12n31jOSEhQf369XPfevXq5X5s2LBhKisrq3FCefHFFxUSEqIhQ4Z41GeM0cSJE3Xy5Mlqz6moqNDrr79eZ00HDhzQoUOH1KlTpwbuFZo7f/RCfaxevVrGGPf9L774Qlu3bq3XH/644oordPjwYVVVVXn0nOuWmpoqSe5tuf5+gMuLL77o/gTrTM477zylpKRo+fLltYZO1y/j9X03uG3bthozZoyeffZZrV+/XiUlJR4fc7v27/PPP1f79u1r3L+kpKR6vVZDREVFafDgwSoqKtJFF11U4+u7go/rj9h8P0SVlJTUeHUX2Is5oGbNYQ5wOBzq3bu3Hn30UZ199tn68MMPPequqebBgwdr8+bN+uqrr9zLqqqq9MILL9TrNW3BO+lBbODAgcrNzdWUKVPUt29f3XbbbbrgggsUEhKigwcPau3atZJU7SO2hrrpppu0aNEijR07VtnZ2frRj36kY8eO6c0339SyZcs0ZcoUj2+zDxgwQIsXL9add96ptLQ03XHHHbrgggtUUVGhoqIiLV26VBdeeKFGjx7tfs6xY8fc54pWVVVp3759evjhhyXJfbko4Ieauhdcvv76a1155ZWaOHGiSktLNWfOHEVERCgnJ6fO515//fVauXKlLr/8ck2dOlWXXHKJwsLCdODAAeXl5ennP/+5rrzySvXs2VM333yz/vCHPygsLEzDhg3Tzp079cgjj9Rrf3JzczV69Gj1799f06ZNU9euXfXll19q48aN7h/6rl/GH3vsMWVmZiosLEypqake517/0IQJE/TCCy9o8uTJ6tKli4YNG+bx+D333KO1a9fqsssu07Rp03TRRRfp1KlT+vLLL7Vp0yZNnz5dP/7xj+usv6Eee+wxXXrppfrJT36iO+64Q0lJSSovL9fevXv1+uuv691335V0OkisW7dOd955p6655hoVFxfr17/+teLj4/keTBBhDjizYJwD1q9fr0WLFmnMmDHq3r27jDFat26djhw5ouHDh7vX69Wrl/Lz8/X6668rPj5eZ511llJTUzVr1iz96U9/0pAhQzR79mxFRUUpNze3Xpd+tEogv7UK39ixY4cZP368SU5ONk6n00RERJhzzjnH3HrrrWbz5s3u9VzfZP/mm28a/FplZWVm5syZ5txzzzXh4eEmKirK9OvXzyxZssTjG+4/rC8zM9N07drVhIeHm9atW5uLL77YzJ4923z99dfu9X54dZeQkBCTkJBgfvrTn5r8/PwG14yWo769YEzj+sF1BYXnnnvO3H333aZjx47G6XSan/zkJx5XIzHm9JUdWrduXeN2KioqzCOPPGJ69+5tIiIiTJs2bUyPHj3MpEmTzJ49e9zrnThxwkyfPt106tTJREREmP79+5uCggLTrVu3Oq/sYIwxBQUF5qc//amJiYkxTqfTpKSkVLtSRE5OjklISDAhISEe2/jhlR1cqqqqTGJiopFkfvWrX9W4f//5z3/MrFmzTGpqqgkPDzcxMTGmV69eZtq0aR5XXajJma4m1a1bNzNq1KhqyyWZu+66y2PZvn37zIQJE0znzp1NWFiY6dixo0lPTze/+c1vPNb77W9/a5KSkozT6TQ9e/Y0y5Ytcx8fdb2Gq6aarjCBpscc0DzmgE8//dTccMMNJiUlxURGRpqYmBhzySWXmBUrVnist2PHDjNw4EATFRVlJHnU+cEHH5j+/fsbp9Np4uLizIwZM8zSpUuD6uouDmO+9zkNAKBO+fn5Gjx4sF566SVdc801gS4HQBNjDkBT4Jx0AAAAwDKEdAAAAMAynO4CAAAAWIZ30gEAAADLENIBAAAAyxDSAQAAAMtY+ceMTp06pX//+98666yzavxTugA8GWNUXl6uhIQEhYTY+bs3fQ14h74Gmqf69raVIf3f//63EhMTA10GEHSKi4vVpUuXQJdRI/oaaBj6Gmie6uptK0O668/QFhcX+/xP+ALNUVlZmRITE2v9E86BRl8D3qGvgeapvr1tZUh3fWQWHR1N0wNesPnjZvoaaBj6Gmie6uptO09yAwAAAFowQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgmdBAF1CbBQskpzPQVQC1mzs30BUEF/oaLvRO80FfA6f5cl7jnXQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAy9Q7pVVVVSk9P19VXX+2xvLS0VImJiZo1a5Yk6csvv9To0aPVunVrdejQQXfffbdOnjzp26oB+Ay9DTQ/9DUQ/Ood0lu1aqVnnnlGGzZs0MqVK93Lp0yZonbt2mn27NmqqqrSqFGjdPToUf35z3/WmjVrtHbtWk2fPt0vxQNoPHobaH7oayD4hXqz8rnnnqsFCxZoypQpGjx4sAoLC7VmzRpt27ZN4eHheuutt7Rr1y4VFxcrISFBkrRw4UKNGzdODzzwgKKjo/2yEwAah94Gmh/6GghuXp+TPmXKFPXu3Vu33nqrbrvtNs2ePVt9+vSRJBUUFOjCCy90N7skjRw5UidOnNDf/vY3nxUNwPfobaD5oa+B4OXVO+mS5HA4tHjxYvXs2VO9evVSdna2+7GSkhLFxsZ6rN+2bVuFh4erpKTkjNs8ceKETpw44b5fVlbmbVkAGsnXvU1fA4FHXwPBq0FXd1m+fLmioqK0b98+HThwwOMxh8NRbX1jTI3LXRYsWKCYmBj3LTExsSFlAWgkX/Y2fQ3Ygb4GgpPXIb2goECPPvqoXnvtNQ0YMEBZWVkyxkiS4uLiqv32/d1336mioqLab+vfl5OTo9LSUvetuLjY27IANJKve5u+BgKPvgaCl1ch/dixY8rMzNSkSZM0bNgwPfnkkyosLNQTTzwhSRowYIB27typgwcPup+zadMmOZ1OpaWlnXG7TqdT0dHRHjcATccfvU1fA4FFXwPBzauQnp2drVOnTumhhx6SJHXt2lULFy7UjBkztH//fo0YMULnn3++brnlFhUVFWnz5s267777NHHiRBoZsBi9DTQ/9DUQ3Ood0rds2aLc3FytWLFCrVu3di+fOHGi0tPTlZWVpZCQEL3xxhuKiIjQwIEDNXbsWI0ZM0aPPPKIX4oH0Hj0NtD80NdA8Kv31V0yMjJUWVlZ42MbN250/7tr165av3594ysD0CTobaD5oa+B4Negq7sAAAAA8B9COgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGAZQjoAAABgGUI6AAAAYBlCOgAAAGCZ0EAXUJucHCk6OtBVAPAl+hpofuhrwPd4Jx0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALAMIR0AAACwDCEdAAAAsAwhHQAAALBMaKALqM2CBZLTGegqgMCaOzfQFfgWfQ2c1px6m74OvOZ0POE03kkHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsQ0gHAAAALENIBwAAACxDSAcAAAAsU++QXlVVpfT0dF199dUey0tLS5WYmKhZs2ZJkqZOnaq0tDQ5nU716dPHp8UC8D16G2h+6Gsg+NU7pLdq1UrPPPOMNmzYoJUrV7qXT5kyRe3atdPs2bMlScYYTZgwQdddd53vqwXgc/Q20PzQ10DwC/Vm5XPPPVcLFizQlClTNHjwYBUWFmrNmjXatm2bwsPDJUl//OMfJUnffPONPv74Y99XDMDn6G2g+aGvgeDmVUiXTv8W/sorr+jWW2/VJ598otmzZzf6I7ITJ07oxIkT7vtlZWWN2h4A7/m6t+lrIPDoayB4ef3FUYfDocWLF2vz5s2KjY1VdnZ2o4tYsGCBYmJi3LfExMRGbxOAd3zd2/Q1EHj0NRC8GnR1l+XLlysqKkr79u3TgQMHGl1ETk6OSktL3bfi4uJGbxOA93zZ2/Q1YAf6GghOXof0goICPfroo3rttdc0YMAAZWVlyRjTqCKcTqeio6M9bgCalq97m74GAo++BoKXVyH92LFjyszM1KRJkzRs2DA9+eSTKiws1BNPPOGv+gA0AXobaH7oayC4eRXSs7OzderUKT300EOSpK5du2rhwoWaMWOG9u/fL0nau3evduzYoZKSEh07dkw7duzQjh07dPLkSZ8XD8A36G2g+aGvgeDmMPX83GvLli0aOnSo8vPzdemll3o8NnLkSFVWVuqdd97R4MGDtWXLlmrP37dvn5KSkupVVFlZmWJiYpSdXSqnk4/S0LLNnVv3Oq6eKS0t9frj56bqbfoa8FRXb9PX8EZ9flbADvXt7XpfgjEjI0OVlZU1PrZx40b3v/Pz8+tfJYCAo7eB5oe+BoJfg67uAgAAAMB/COkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlCOkAAACAZQjpAAAAgGUI6QAAAIBlQgNdQG1ycqTo6EBXAcCX6Gug+aGvAd/jnXQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyhHQAAADAMoR0AAAAwDKEdAAAAMAyoYEuoCbGGElSWVlZgCsBgoOrV1y9YyP6GvAOfQ00T/XtbStD+uHDhyVJiYmJAa4ECC7l5eWKiYkJdBk1oq+BhqGvgeaprt62MqS3a9dOkvTll19aOzHZoqysTImJiSouLlZ0dHSgy7FWcx8nY4zKy8uVkJAQ6FLOiL72r+Z+jAdaIMaXvm556GPfsXks69vbVob0kJDTp8rHxMRYN7C2io6OZqzqoTmPk+0/IOnrptGcj3EbNPX40tctE33sO7aOZX16my+OAgAAAJYhpAMAAACWsTKkO51OzZkzR06nM9ClWI+xqh/GKfD4P/Avxte/GN+aMS6+xXj6TnMYS4ex+dpOAAAAQAtk5TvpAAAAQEtGSAcAAAAsQ0gHAAAALENIBwAAACwTsJC+aNEiJScnKyIiQmlpaXr//fdrXX/Lli1KS0tTRESEunfvriVLljRRpYHnzVjl5+fL4XBUu3366adNWHHTe++99zR69GglJCTI4XDo1VdfrfM5LfmY8hf62r+YC/yHOeTM6Gvf8WYs161bp+HDh6tjx46Kjo7WgAEDtHHjxias1m7eHpcuH3zwgUJDQ9WnTx//FugLJgDWrFljwsLCzLJly8yuXbvM1KlTTevWrc0XX3xR4/r//Oc/TVRUlJk6darZtWuXWbZsmQkLCzMvv/xyE1fe9Lwdq7y8PCPJ/OMf/zAHDx503yorK5u48qb15ptvml/96ldm7dq1RpJ55ZVXal2/JR9T/kJf+xdzgX8xh9SMvvYdb8dy6tSp5qGHHjLbtm0zn332mcnJyTFhYWHmww8/bOLK7ePtWLocOXLEdO/e3YwYMcL07t27aYpthICE9EsuucTcfvvtHst69OhhsrOza1x/5syZpkePHh7LJk2aZPr37++3Gm3h7Vi5fjB/9913TVCdnerzA7YlH1P+Ql/7F3NB02EO+R/62ne8HcuanH/++WbevHm+Li3oNHQsr7vuOjNr1iwzZ86coAjpTX66y8mTJ/W3v/1NI0aM8Fg+YsQIbd26tcbnFBQUVFt/5MiR2r59uyoqKvxWa6A1ZKxcLr74YsXHx2vo0KHKy8vzZ5lBqaUeU/5CX/sXc4F9WsLxS1/7TmN62OXUqVMqLy9Xu3bt/FFi0GjoWD799NP6/PPPNWfOHH+X6DNNHtIPHTqkqqoqxcbGeiyPjY1VSUlJjc8pKSmpcf3KykodOnTIb7UGWkPGKj4+XkuXLtXatWu1bt06paamaujQoXrvvfeaouSg0VKPKX+hr/2LucA+LeH4pa99pyFj+UMLFy7U0aNHNXbsWH+UGDQaMpZ79uxRdna2Vq5cqdDQ0KYo0ycCVqnD4fC4b4yptqyu9Wta3hx5M1apqalKTU113x8wYICKi4v1yCOP6LLLLvNrncGmJR9T/kJf+xdzgV1ayvFLX/uOt2Ppsnr1as2dO1evvfaaOnXq5K/ygkp9x7Kqqko33nij5s2bp/POO6+pyvOJJn8nvUOHDmrVqlW133a+/vrrar8VucTFxdW4fmhoqNq3b++3WgOtIWNVk/79+2vPnj2+Li+otdRjyl/oa/9iLrBPSzh+6WvfaUwPv/DCC8rKytKLL76oYcOG+bPMoODtWJaXl2v79u2aPHmyQkNDFRoaqvnz5+ujjz5SaGio3n333aYq3WtNHtLDw8OVlpamt99+22P522+/rfT09BqfM2DAgGrrb9q0Sf369VNYWJjfag20hoxVTYqKihQfH+/r8oJaSz2m/IW+9i/mAvu0hOOXvvadhvbw6tWrNW7cOK1atUqjRo3yd5lBwduxjI6O1ieffKIdO3a4b7fffrtSU1O1Y8cO/fjHP26q0r0XiG+rui6d89RTT5ldu3aZe+65x7Ru3drs37/fGGNMdna2ueWWW9zruy7pNG3aNLNr1y7z1FNPtZhLOnk7Vo8++qh55ZVXzGeffWZ27txpsrOzjSSzdu3aQO1CkygvLzdFRUWmqKjISDK///3vTVFRkftyTBxT/kdf+xdzgX8xh9SMvvYdb8dy1apVJjQ01OTm5npcRvXIkSOB2gVreDuWPxQsV3cJSEg3xpjc3FzTrVs3Ex4ebvr27Wu2bNnifiwzM9NkZGR4rJ+fn28uvvhiEx4ebpKSkszixYubuOLA8WasHnroIZOSkmIiIiJM27ZtzaWXXmreeOONAFTdtFyXm/vhLTMz0xjDMdVU6Gv/Yi7wH+aQM6OvfcebsczIyKj1mGzpvD0uvy9YQrrDmP//jQ4AAAAAVmjyc9IBAAAA1I6QDgAAAFiGkA4AAABYhpAOAAAAWIaQDgAAAFiGkA4AAABYhpAOAAAAWIaQDgAAAFiGkA4AAABYhpAOAAAAWIaQDgAAAFiGkA4AAABY5v8Bnsc1qGKFE2kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 900x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to know the attribution to each feature?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████| 7/7 [00:00<00:00, 83.14it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to visualise the GP on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Do you want to visualise the learnt preference model on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Type 0 or 1 which you think larger. Type 2 if unsure.0\n",
      "You chose X0\n",
      "Probability of correct selection: 7.81e-03 ± 8.82e-02\n",
      "Estimated total correct selection rate amongst sure samples: 7.79e-01 ± 1.95e-01\n",
      "Is your selection correct? Yes if 1: 1.0\n",
      "Is your selection sure? Yes if 1: 1\n",
      "6) Best value: 1.68844e+01\n",
      "1) parameters: beta 1.200e+00 gamma 1.000e-02\n",
      "training models...\n",
      "generating candidates...\n",
      "X0 (preference): tensor([1.1105, 0.3000, 0.7394])\n",
      "X1 (normal UCB): tensor([1.0714, 0.3000, 0.5757])\n",
      "Do you need more explanation on the two candidates?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to know the attribution to each feature?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 7/7 [00:00<00:00, 4030.22it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to visualise the GP on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as no0\n",
      "Do you want to visualise the learnt preference model on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Type 0 or 1 which you think larger. Type 2 if unsure.0\n",
      "You chose X0\n",
      "Probability of correct selection: 4.59e-01 ± 9.15e-02\n",
      "Estimated total correct selection rate amongst sure samples: 7.29e-01 ± 2.33e-01\n",
      "Is your selection correct? Yes if 1: 1.0\n",
      "Is your selection sure? Yes if 1: 1\n",
      "7) Best value: 1.68844e+01\n",
      "2) parameters: beta 1.470e+00 gamma 4.000e-02\n",
      "training models...\n",
      "generating candidates...\n",
      "X0 (preference): tensor([1.5066, 0.4785, 0.9567])\n",
      "X1 (normal UCB): tensor([1.0141, 0.5200, 0.3778])\n",
      "Do you need more explanation on the two candidates?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to know the attribution to each feature?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 7/7 [00:00<00:00, 3875.92it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to visualise the GP on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Do you want to visualise the learnt preference model on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Type 0 or 1 which you think larger. Type 2 if unsure.0\n",
      "You chose X0\n",
      "Probability of correct selection: 1.01e-01 ± 1.11e-01\n",
      "Estimated total correct selection rate amongst sure samples: 7.20e-01 ± 2.23e-01\n",
      "Is your selection correct? Yes if 1: 1.0\n",
      "Is your selection sure? Yes if 1: 1\n",
      "8) Best value: 1.68844e+01\n",
      "3) parameters: beta 1.697e+00 gamma 9.000e-02\n",
      "training models...\n",
      "generating candidates...\n",
      "X0 (preference): tensor([1.0908, 0.3243, 0.9956])\n",
      "X1 (normal UCB): tensor([1.0071, 0.5200, 0.3777])\n",
      "Do you need more explanation on the two candidates?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to know the attribution to each feature?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 7/7 [00:00<00:00, 3467.19it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAicAAAF0CAYAAAD4opkZAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA74klEQVR4nO3deVxU9d4H8M/IMiDLiI4C6ogEKoi4gLlkN5cUNRe6KaVpinjVjBSrG4o9udyXj3Rdyrq3TXsEczcjMx9T08SrYWmIhklJmkIKGBozajgI/J4/fJgaB3QGZjmH+bxfr/OqOb+zfM+cD9O3M8tRCCEEiIiIiCSiiaMLICIiIvozNidEREQkKWxOiIiISFLYnBAREZGksDkhIiIiSWFzQkRERJLC5oSIiIgkhc0JERERSQqbEyIiIpIUNieNzKJFi6BQKFBaWlrreJcuXTBgwACjeSUlJZg3bx4iIyPh7e0NDw8PdOjQAUlJScjPzzfZds3UpEkTBAYG4rHHHsNXX31ly8Miifvuu+8wdepUhISEwNPTE56enujQoQNmzJiBb7/91mjZu3Pk7u6O4OBgJCUloayszDEHQE5HSpnNysrCokWLLNpW+/btER8ff9/l0tPToVAocOHChXrX5wiuji6AHOvYsWMYOXIkhBB4/vnn0bdvX7i7u+PHH3/Ehg0b0KtXL/z2229G6+zZswcqlQrV1dUoKCjAsmXLMGDAAHzzzTeIiopy0JGQo7z//vt4/vnn0alTJyQlJSEiIgIKhQJ5eXnYvHkzHnzwQfz0008ICQkxWq8mR9evX8fu3bvx5ptv4tixY8jKyoJCoXDQ0ZAzkFpms7KysHjxYsTHx6NZs2YNPLrGgc2JE9PpdIiNjYWHhweysrLQtm1bw9iAAQMwY8YMbN++3WS96OhoqNVqAMBDDz2EXr16ISQkBNu3b2dz4mS++uorPPfccxgxYgS2b98Od3d3w9igQYOQmJiIjz76CJ6enibr/jlHQ4YMwdWrV7F+/XpkZWWhX79+djsGci7MrDzwbR0ntmbNGhQXF2PZsmVGjcmfjR079r7bUalUAAA3Nzer1kfSt3TpUri4uOD99983epH/s7i4OLRu3fq+2+rTpw8A4OLFi3Uuk5mZCYVCgU2bNmHu3LkIDAyEt7c3Ro0ahZKSEly/fh3Tp0+HWq2GWq3GlClTcOPGDaNtCCHwzjvvoHv37vD09ISfnx/Gjh2L8+fPGy33xRdfIDY2Fm3btoWHhwdCQ0MxY8YMk7dMay75f//99xg/fjxUKhX8/f2RkJAArVZ73+Mm+7J3Zqurq7FkyRJ06tQJnp6eaNasGbp27Yo333wTwJ38vPzyywCA4OBgw1tHmZmZAIDbt28jOTkZAQEBaNq0KR5++GEcO3as1n19/fXX6NevHzw8PNC6dWukpKTg9u3btS67detW9O3bF15eXvD29sbQoUORk5NjGF+1ahUUCgV++uknk3Xnzp0Ld3f3Oj8+YA1sTpzYvn374OLiglGjRlm0XlVVFSorK1FRUYGffvoJiYmJUCqVZjUy1HhUVVXh4MGD6NmzJwIDAxu8vZoXwZYtW9532fnz5+PKlStIT0/HypUrkZmZifHjx2PMmDFQqVTYvHkzkpOTsX79esyfP99o3RkzZmDOnDkYPHgwduzYgXfeeQfff/89HnroIZSUlBiWO3fuHPr27Yt3330X+/btw4IFC/DNN9/g4YcfrvUFf8yYMejYsSM+/vhjzJs3D5s2bcILL7zQwGeFrMkRmV22bBkWLVqE8ePH43//93+xdetWTJ061fD5kr/97W+YNWsWACAjIwNHjx7F0aNHDVehp02bhhUrVmDSpEn49NNPMWbMGDzxxBMmb7efOXMGjz76KMrKypCeno733nsPOTk5WLJkiUlNS5cuxfjx49G5c2ds27YN69evx/Xr1/GXv/wFZ86cAQBMnDgR7u7uSE9PN1q3qqoKGzZswKhRowxXkWxCUKOycOFCAUD8+uuvtY5HRESI/v37CyGECAsLEwEBARZv++7J19dXZGRkWKN8kpHi4mIBQIwbN85krLKyUty+fdswVVdXG8ZqclRcXCxu374tfvvtN7Fhwwbh6ekpNBqNKC8vr3OfBw8eFADEqFGjjObPmTNHABCzZ882mv/444+L5s2bGx4fPXpUABArV640Wq6wsFB4enqK5OTkWvdbXV0tbt++LS5evCgAiE8//dTkeJYtW2a0znPPPSc8PDyMjp0cyxGZHTlypOjevfs961q+fLkAIH7++Wej+Xl5eQKAeOGFF4zmb9y4UQAQkydPNsx76qmnhKenpyguLjY6prCwMKNtFxQUCFdXVzFr1iyjbV6/fl0EBASIJ5980jDviSeeEG3bthVVVVWGebt37xYAxGeffXbPY2ooXjkhi+3fvx/Hjx/HsWPHsGvXLgwePBjjxo3DJ5984ujSSCKio6Ph5uZmmFauXGmyTEBAANzc3ODn54eJEyciKioKe/bsgYeHx323P3LkSKPH4eHhAIARI0aYzL927ZrhrZ1du3ZBoVBg4sSJqKysNEwBAQHo1q2b4VI6AFy5cgXPPvssNBoNXF1d4ebmhqCgIABAXl6eSU2jR482ety1a1fcunULV65cue/xkOPZKrO9evXCqVOn8Nxzz2Hv3r3Q6XRm13Tw4EEAwIQJE4zmP/nkk3B1dTVZ9tFHH4W/v79hnouLC5566imj5fbu3YvKykpMmjTJ6G/Aw8MD/fv3N/obmDJlCn755Rfs37/fMC8tLQ0BAQEYPny42cdRH/xAbCNTE9iqqqpaxysrKw2fDWnXrh3y8/Nx8+ZNeHl5mb2Pbt26GV3OGz58OCIjI5GYmIi//vWvDaie5EStVsPT07PW99s3bdqE33//HUVFRSb/0a6xf/9+qFQquLm5oW3btmjRooXZ+27evLnR45rPDtQ1/9atW/D29kZJSQmEEEYv4H/2wAMPALjzOYGYmBhcvnwZr776KiIjI+Hl5YXq6mr06dMH5eXlJuveXb9SqQSAWpclx3BEZlNSUuDl5YUNGzbgvffeg4uLCx555BH885//RM+ePe+57tWrVwHcaYr+zNXV1WTfV69eNVmutnVr3rp88MEHa91nkyZ/XLMYPnw4AgMDkZaWhpiYGPz222/YuXMnkpKS4OLics/aG4rNSSNT86J76dIlkxdgIQSKiooMfxBDhw7Fvn378Nlnn2HcuHH13meTJk0QERGBjz76CFeuXEGrVq3qfwAkGy4uLhg0aBD27duHoqIio/fwO3fuDAD3/G2Fu5tce1Cr1VAoFDh8+LChefizmnmnT5/GqVOnkJ6ejsmTJxvGa/twIMmHIzLr6uqKF198ES+++CLKysqwf/9+zJ8/H0OHDkVhYSGaNm1a57o1DUhxcTHatGljmF9ZWWloXP68bHFxsck27p5XU//27dsNVwLr4uLigmeeeQZvvfUWysrKsGnTJuj1ekyZMuXeB20FfFunkRk0aBAUCgW2bt1qMrZnzx7odDoMHjwYADB16lQEBAQgOTkZly5dqnV7GRkZ991nVVUVcnNzoVQq4evr27ADIFlJSUlBVVUVnn322Tq/FSAlNb/pc+nSJfTs2dNkioyMBADDb1bc3cC8//77dq+ZrMuRmW3WrBnGjh2LxMREXLt2zdAI1XWVreYHMzdu3Gg0f9u2baisrDSaN3DgQBw4cMDoQ91VVVUm/y0YOnQoXF1dce7cuVr/Bu6+mjNlyhTcunULmzdvRnp6Ovr27YuwsLB6Pwfm4pWTRiYkJATPP/88li9fjrKyMjz22GPw9PTE8ePH8dprr6Fnz554+umnAdz5CvCnn36KkSNHokePHkY/wpafn48NGzbg1KlTeOKJJ4z2kZ2dbfj6cElJCdauXYsffvgBL7zwglmfF6DGo1+/fnj77bcxa9YsREVFYfr06YiIiECTJk1QVFSEjz/+GAAk07T269cP06dPx5QpU/Dtt9/ikUcegZeXF4qKinDkyBFERkZi5syZCAsLQ0hICObNmwchBJo3b47PPvsMX3zxhaMPgRrI3pkdNWoUunTpgp49e6Jly5a4ePEiVq1ahaCgIHTo0AEADE3xm2++icmTJ8PNzQ2dOnVCeHg4Jk6ciFWrVsHNzQ2DBw/G6dOnsWLFCpP6/uu//gs7d+7EoEGDsGDBAjRt2hRvv/02bt68abRc+/bt8Y9//AOvvPIKzp8/j2HDhsHPzw8lJSU4duwYvLy8sHjxYsPyYWFh6Nu3L1JTU1FYWIjVq1db5Xm5L5t+3JYcorq6Wrz77ruiZ8+eomnTpsLd3V106NBBzJ07V1y/ft1k+eLiYjF37lwREREhmjZtKpRKpQgNDRUzZswQubm5huVq+7ZO8+bNRe/evcXatWuNPtFNzuXkyZNiypQpIjg4WCiVSuHh4SFCQ0PFpEmTxIEDB4yWvd83yu6l5ts6H330kdH8tLQ0AUAcP37crH2tXbtW9O7dW3h5eQlPT08REhIiJk2aJL799lvDMmfOnBFDhgwRPj4+ws/PT8TFxYmCggIBQCxcuPC++6ip6e5vYJA02CuzK1euFA899JBQq9XC3d1dtGvXTkydOlVcuHDBaLmUlBTRunVr0aRJEwFAHDx4UAghhF6vFy+99JJo1aqV8PDwEH369BFHjx4VQUFBRt/WEUKIr776SvTp00colUoREBAgXn75ZbF69epac7hjxw4xcOBA4evrK5RKpQgKChJjx44V+/fvNzmGmm14enoKrVZr8XNQHwohhLBPG0RERER0f/zMCREREUkKmxMiIiKSFDYnREREJClsToiIiEhS2JwQERGRpLA5ISIiIknhj7BZoLq6GpcvX4aPj4/hFyRJHoQQuH79Olq3bm107whnwwzLFzP8B+ZYnizJMJsTC1y+fBkajcbRZVADFBYWom3bto4uw2GYYflz9gwDzLHcmZNhNicW8PHxAXDniZXKz3Hfz8mTJ9G/f38cOnQI3bt3d3Q5DqPT6aDRaAzn0Fkxw/LFDP9Bbjlmhu+wJMNsTixQc/nQ19dXFn8QAODt7W34p1xqtiVnvwTMDMufs2cYkF+OmWFj5mTYud+4JCIiIslhc0JERESSwuaEiIiIJIXNCREREUkKmxMiIiKSFDYnREREJCn8KrGTiI6OdnQJdiWEcHQJZGXWyDBzQY5ki9fhxpppXjkhIiIiSWFzQkRERJLC5oSIiIgkRZLNyZUrVzBjxgy0a9cOSqUSAQEBGDp0KI4ePQrgzk/f7tixw2S9OXPmYMCAAYbH8fHxUCgUePbZZ02Wfe6556BQKBAfH2+joyBnxxyT3DHD5CiSbE7GjBmDU6dOYd26dTh79ix27tyJAQMG4Nq1axZvS6PRYMuWLSgvLzfMu3XrFjZv3ox27dpZs2wiI8wxyR0zTI4iuW/rlJWV4ciRI8jMzET//v0BAEFBQejVq1e9thcVFYXz588jIyMDEyZMAABkZGRAo9HggQcesFrdRH/GHJPcMcPkSJK7cuLt7Q1vb2/s2LEDer3eKtucMmUK0tLSDI/Xrl2LhIQEq2ybqDbMMckdM0yOJLnmxNXVFenp6Vi3bh2aNWuGfv36Yf78+fjuu+/qvc1nnnkGR44cwYULF3Dx4kV89dVXmDhx4n3X0+v10Ol0RhOROaSSY2aY6ksqGQaYY2ckueYEuPM+5+XLl7Fz504MHToUmZmZiIqKQnp6er22p1arMWLECKxbtw5paWkYMWIE1Gr1fddLTU2FSqUyTBqNpl77J+ckhRwzw9QQUsgwwBw7I0k2JwDg4eGBIUOGYMGCBcjKykJ8fDwWLlwIAPDx8YFWqzVZp6ysDCqVqtbtJSQkGP4vwNzLiCkpKdBqtYapsLCw/gdETsnROWaGqaEcnWGAOXZGkm1O7ta5c2fcvHkTABAWFobjx48bjQshkJ2djU6dOtW6/rBhw1BRUYGKigoMHTrUrH0qlUr4+voaTUQNYe8cM8NkbXwtJnuQ3Ld1rl69iri4OCQkJKBr167w8fHBt99+i2XLliE2NhYA8Pe//x2TJ09GWFgYYmJiUF5ejtWrV+PcuXNITEysdbsuLi7Iy8sz/DuRLTHHJHfMMDmS5JoTb29v9O7dG2+88QbOnTuH27dvQ6PRYNq0aZg/fz4A4Mknn4QQAitWrMArr7wCDw8P9OjRA4cPH0ZQUFCd22a3TfbCHJPcMcPkSArRWG9paAM6nQ4qlQparVY2f1wnTpxwujsSA6Z36pTjubMFOT4P1sywnF/u5HjubEVuz4UtX4fllGlLzptsPnNCREREzoHNCREREUkKmxMiIiKSFDYnTiI7OxtCCKeZqPGxRoaJHMkWr8ONFZsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJYXNCREREkiK5e+vIwcmTJ+Ht7W147Ofnh+DgYNy6dQtnzpwxWT4qKgoA8OOPPxru5lmjffv2aN68OX799VeT24D7+PigQ4cOqKqqwqlTp0y2GxkZCTc3N5w7d87ktuVt2rSBv78/dDodABhutAUAnp6eCA8PBwDk5OSYfB0tPDwcnp6euHjxIq5evWo05u/vjzZt2uD69evIz883GnNzc0NkZCQAIDc3F7dv3zYa79ChA3x8fHDp0iWUlJQYjbVo0QJBQUEoLy83qhUAFAoFevToYTiO8vJyo/Hg4GD4+fmhpKQEly5dMhpTqVQICQkxqcXZMcPMcGMglxzXuDsXzpbjli1bwmyCzKbVagUAk2nChAlCCCHy8/NrHa/Rp08fk7H169cLIYT497//bTIWExNzz/1euXJFCCHEqFGjTMZWrlwphBDitddeMxnr0aOHoSZ3d3eT8dOnTwshhJg6darJ2Lx584QQQhw8eNBkrE2bNobttmnTxmT84MGDQggh5s2bZzI2depUIYQQp0+fNhlzd3c3bLdHjx4m49u2bRNCCLFy5UqTsVGjRgkhhDh37pwAILRarTWiIFvMMDPcGMgtx9nZ2bWu52w5rnn+zMkwb/xngZqbFh06dEgW3bq/vz8yMzMxcOBAbNiwwdChO1u3HhISgqtXr0KtVsvmRmG2wgwzw42B3HJ86dIlREdHG2UYcL4ct2zZ0uwb/7E5sYDc7oQJ/HE3zOzsbMMfpjOS47mzBTk+D8zwHXI8d7Yit+eCGb6DdyUmIiIi2WJzQkRERJLC5oSIiIgkhc0JERERSQqbEyIiIpIUNidEREQkKfyFWCcRHW3d7fEL6EREljH3dZivr7xyQkRERBLD5oSIiIgkRbLNiUKhwI4dOxxdBlG9McMkd8wwOYpDm5P4+Hg8/vjjtY4VFRVh+PDhhscKhcIw+fj4oGfPnsjIyDCML1q0yGiZmmn//v2GZcrKypCYmIjAwEB4eHggPDwcu3fvttnxUePHDJPcMcMkRZL9QGxAQIDJvLS0NAwbNgxlZWVYvnw54uLicOTIEfTt2xcAEBERYfRHAADNmzcHAFRUVGDIkCFo1aoVtm/fjrZt26KwsBA+Pj62PxhySswwyR0zTI4i2eZEoVDgk08+MeromzVrhoCAAAQEBOC9997Dli1bsHPnTsMfhaura61/TACwdu1aXLt2DVlZWXBzcwMABAUF2fw4yHkxwyR3zDA5imQ/c3I/bm5ucHV1NbkVdF1q/ngSExPh7++PLl26YOnSpaiqqrJxpUS1Y4ZJ7phhshXJXjm5F71ej+XLl0On0+HRRx81zM/NzYW3t7fhcefOnXHs2DEAwPnz5/Hll19iwoQJ2L17N/Lz85GYmIjKykosWLCgzv3o9XrDY51OZ6MjImfDDJPc2SvDNftijp2LrJqT8ePHw8XFBeXl5VCpVFixYoXRh7U6deqEnTt3Gh4rlUrDv1dXV6NVq1ZYvXo1XFxcEB0djcuXL2P58uV1/lGkpqZi8eLFtjsgcjrMMMmdvTMMMMfOSFbNyRtvvIHBgwfD19cXrVq1Mhl3d3dHaGhoresGBgbCzc0NLi4uhnnh4eEoLi5GRUUF3N3dTdZJSUnBiy++aHis0+mg0WiscCTkrJhhkjt7Zxhgjp2RrJqTgICAOkN/P/369cOmTZtQXV2NJk3ufNTm7NmzCAwMrPMPQqlUGnX9RA3FDJPc2TvDAHPsjBzenGi1Wpw8edJoXs3Xzqxp5syZ+Ne//oWkpCTMmjUL+fn5WLp0KWbPnm31fZFzYYZJ7phhkhqHNyeZmZno0aOH0bzJkydbfT8ajQb79u3DCy+8gK5du6JNmzZISkrC3Llzrb4vci7MMMkdM0xSoxCC9z80l06ng0qlglarha+vr6PLMcuJEycQHR0NIBtAlNW2K7fUyPHc2YIcn4eaDGdnZyMqynoZlhs5njtbkdtzYenrsNxeX81lyXmT7e+cEBERUePE5oSIiIgkhc0JERERSQqbEyeRnX3nfUxrTUREZBlzX4eJzQkRERFJjMVfJb5x4ways7NRXFwMhUIBf39/REdHG91LgYiIiKi+zG5OKisr8dJLL2HNmjW4desW3N3dIYTA7du34eHhgenTp2P58uWG22ATERER1YfZb+u89NJL+Pjjj5GWloZr167h1q1b0Ov1uHbtGtLS0pCRkYGXX37ZlrUSERGREzD7ysmmTZuwdetWDBo0yGh+s2bN8NRTT0GtVmPcuHFYtWqVtWskIiIiJ2L2lZPy8nKo1eo6x1u0aIHy8nKrFEVERETOy+zmZODAgXjxxRdRUlJiMlZSUoLk5GSTqypEREREljL7bZ133nkHjz32GNq2bYsuXbrA398fCoUCxcXFOH36NDp37oxdu3bZslYiIiJyAmY3JxqNBqdOncLevXvx9ddfo7i4GADQq1cvpKamIiYmBk2a8GdTiIiIqGHMbk4WLFiAhQsXYvjw4Rg+fLjJeEFBAaZOnYovvvjCqgUSERGRczH7Ukd6ejp69uyJ3Nxck7HVq1ejS5cucHW1+DfdiIiIiIyY3ZycPn0akZGRePDBB5Gamorq6moUFBRg8ODBSE5Oxuuvv47PP//clrUSERGREzD7Uoevry8+/PBDjBkzBjNmzMDWrVvx888/o2/fvsjNzYVGo7FlnUREROQkLP4Ea+/evREZGYnvvvsO1dXVSE5OZmNCREREVmNRc7J582ZERESguroaeXl5mDlzJoYPH46kpCT+ABsRERFZhdnNydixYzF9+nQsWrQIBw4cQKdOnbBs2TJkZmZiz5496NatG44ePWrLWomIiMgJmP2Zk6KiIuTk5CA0NNRoft++fXHq1CnMnTsX/fv3R0VFhdWLpIaLjrbdtoWw3baJaliaYeaSpKa2DDOntTO7OTl8+HCdP7Lm4eGBN998E2PGjLFaYUREROSczH5bx5xff33kkUcaVAwRERERf2+eiIiIJKVRNCfx8fFQKBQm07BhwwzL5OTkIC4uDv7+/vDw8EDHjh0xbdo0nD171oGVE/2BOSa5Y4bJWhpFcwIAw4YNQ1FRkdG0efNmAMCuXbvQp08f6PV6bNy4EXl5eVi/fj1UKhVeffVVB1dO9AfmmOSOGSZraDQ3w1EqlQgICDCZ//vvv2PKlCl47LHH8MknnxjmBwcHo3fv3igrK7NjlUT3xhyT3DHDZA2N5spJXfbu3YvS0lIkJyfXOt6sWTP7FkRUD8wxyR0zTJZoNFdOdu3aBW9vb6N5c+fOhVKpBACEhYVZvE29Xg+9Xm94rNPpGlYk0X1YO8fMMNkbX4vJGhpNczJw4EC8++67RvOaN2+ONWvW1HubqampWLx4cUNLIzKbtXPMDJO98bWYrKHRvK3j5eWF0NBQo6l58+bo2LEjAOCHH36weJspKSnQarWGqbCw0NplExmxdo6ZYbI3vhaTNTSa5qQuMTExUKvVWLZsWa3j9/oQllKphK+vr9FE5Aj1zTEzTFLB12KyRKN5W0ev16O4uNhonqurK9RqNT744APExcVh9OjRmD17NkJDQ1FaWopt27ahoKAAW7ZscVDVRMaYY5I7ZpisodFcOdmzZw8CAwONpocffhgAEBsbi6ysLLi5ueHpp59GWFgYxo8fD61WiyVLlji4cqI/MMckd8wwWYNCCN4T0Vw6nQ4qlQparVY2lxVPnDiB6OhoANkAomyyDzkkSI7nzhbk+DzUN8NyyKUl5HjubEVuz8W9MtzYcnovlpy3RnPlhIiIiBoHNidEREQkKWxOiIiISFLYnDiJ7Ow7723aYiKyB0szTCQ1tWWYasfmhIiIiCSFzQkRERFJCpsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJYXNCREREksLmhIiIiCSFzQkRERFJCpsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJYXNCREREkuLq6ALIPqKjG7a+ENapg6i+asswc0lywgybj1dOiIiISFLYnBAREZGksDkhIiIiSZFUcxIfHw+FQgGFQgE3Nzf4+/tjyJAhWLt2Laqrqw3LtW/fHgqFAlu2bDHZRkREBBQKBdLT043m5+TkIC4uDv7+/vDw8EDHjh0xbdo0nD171taHRU6GOSa5Y4bJ0STVnADAsGHDUFRUhAsXLuDzzz/HwIEDkZSUhJEjR6KystKwnEajQVpamtG6X3/9NYqLi+Hl5WU0f9euXejTpw/0ej02btyIvLw8rF+/HiqVCq+++qpdjoucC3NMcscMk0MJCZk8ebKIjY01mX/gwAEBQKxZs0YIIURQUJCYN2+eUCqVoqCgwLDctGnTxKxZs4RKpRJpaWlCCCFu3rwp1Gq1ePzxx2vd52+//WZ2fVqtVgAQWq3W7HUcLTs7WwAQQLa487nw+k1yZ89zJ+UcN7YMOxNm+A9yyzEzfIcl501yV05qM2jQIHTr1g0ZGRmGef7+/hg6dCjWrVsHAPj999+xdetWJCQkGK27d+9elJaWIjk5udZtN2vWzGZ1E/0Zc0xyxwyTvciiOQGAsLAwXLhwwWheQkIC0tPTIYTA9u3bERISgu7duxstk5+fb1jfUnq9Hjqdzmgiagh755gZJmvjazHZg2yaEyEEFAqF0bwRI0bgxo0b+M9//oO1a9eadOo169VXamoqVCqVYdJoNPXeFhFg/xwzw2RtfC0me5BNc5KXl4fg4GCjea6urnjmmWewcOFCfPPNN5gwYYLJeh07dgQA/PDDDxbvMyUlBVqt1jAVFhbWr3ii/2fvHDPDZG18LSZ7kEVz8uWXXyI3NxdjxowxGUtISMChQ4cQGxsLPz8/k/GYmBio1WosW7as1m2XlZXVuV+lUglfX1+jiai+HJFjZpisia/FZC+Su7eOXq9HcXExqqqqUFJSgj179iA1NRUjR47EpEmTTJYPDw9HaWkpmjZtWuv2vLy88MEHHyAuLg6jR4/G7NmzERoaitLSUmzbtg0FBQW1fkefqCGYY5I7ZpgcSXLNyZ49exAYGAhXV1f4+fmhW7dueOuttzB58mQ0aVL7hZ4WLVrcc5uxsbHIyspCamoqnn76aeh0Omg0GgwaNAhLliyxxWGQk2OOSe6YYXIkhWjIp5ScjE6ng0qlglarlc1lxRMnTiA6OhpANoCoem9H7imR47mzBTk+D/fKsNxzaQk5njtbkdtzwQzfYcl5k8VnToiIiMh5sDkhIiIiSWFzQkRERJLC5sRJZGc35M46jq6eqPYME8kJM2w+NidEREQkKWxOiIiISFLYnBAREZGksDkhIiIiSWFzQkRERJLC5oSIiIgkRXL31pGyml/61+l0Dq7EfDdu3DD8U051W1vNsTv73RqYYflihv8gtxwzw3dYkmHeW8cCv/zyCzQajaPLoAYoLCxE27ZtHV2GwzDD8ufsGQaYY7kzJ8NsTixQXV2Ny5cvw8fHBwqFwjC/5s6ahYWFkrsJFWu7QwiB69evo3Xr1nXeUdUZMMPWxQw7Rm05Zk7qz171WZJhvq1jgSZNmtyz2/P19ZVk8ADWBgAqlcrm+5A6Ztg2mGH7uleOmZP6s0d95mbYudtvIiIikhw2J0RERCQpbE6sQKlUYuHChVAqlY4uxQRrI3NI+VywNjKHlM+FlGsDpFkfPxBLREREksIrJ0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJYXNCREREksLmhIiIiCSFzQkRERFJCpsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJYXNCREREksLmhIiIiCSFzQkRERFJCpsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJcXV0AXJSXV2Ny5cvw8fHBwqFwtHlkAWEELh+/Tpat26NJk2ctydnhuWLGf4DcyxPlmSYzYkFLl++DI1G4+gyqAEKCwvRtm1bR5fhMMyw/Dl7hgHmWO7MyTCbEwv4+PgAuPPE+vr6Orga85w8eRL9+/fHoUOH0L17d0eX4zA6nQ4ajcZwDp0VMyxfzPAf5JZjZvgOSzLM5sQCNZcPfX19ZfEHAQDe3t6Gf8qlZlty9kvAzLD8OXuGAfnlmBk2Zk6GnfuNSyIiIpIcNidEREQkKWxOiIiISFLYnBAREZGksDkhIiIiSWFzQkRERJLCrxI7iejoaEeXYDYhhKNLIAmyVoaZL3IUSzPszFnllRMiIiKSFDYnREREJClsToiIiEhSJNmcXLlyBTNmzEC7du2gVCoREBCAoUOH4ujRowDu/PTtjh07TNabM2cOBgwYYHgcHx8PhUKBZ5991mTZ5557DgqFAvHx8TY6CnJ2zDHJHTNMjiLJ5mTMmDE4deoU1q1bh7Nnz2Lnzp0YMGAArl27ZvG2NBoNtmzZgvLycsO8W7duYfPmzWjXrp01yyYywhyT3DHD5CiS+7ZOWVkZjhw5gszMTPTv3x8AEBQUhF69etVre1FRUTh//jwyMjIwYcIEAEBGRgY0Gg0eeOABq9VN9GfMMckdM0yOJLkrJ97e3vD29saOHTug1+utss0pU6YgLS3N8Hjt2rVISEiwyraJasMck9wxw+RIkmtOXF1dkZ6ejnXr1qFZs2bo168f5s+fj++++67e23zmmWdw5MgRXLhwARcvXsRXX32FiRMn3nc9vV4PnU5nNBGZQyo5ZoapvqSSYYA5dkaSa06AO+9zXr58GTt37sTQoUORmZmJqKgopKen12t7arUaI0aMwLp165CWloYRI0ZArVbfd73U1FSoVCrDpNFo6rV/ck5SyDEzTA0hhQwDzLEzkmRzAgAeHh4YMmQIFixYgKysLMTHx2PhwoUAAB8fH2i1WpN1ysrKoFKpat1eQkKC4f8CzL2MmJKSAq1Wa5gKCwvrf0DklBydY2aYGsrRGQaYY2ck2ebkbp07d8bNmzcBAGFhYTh+/LjRuBAC2dnZ6NSpU63rDxs2DBUVFaioqMDQoUPN2qdSqYSvr6/RRNQQ9s4xM0zWxtdisgfJfVvn6tWriIuLQ0JCArp27QofHx98++23WLZsGWJjYwEAf//73zF58mSEhYUhJiYG5eXlWL16Nc6dO4fExMRat+vi4oK8vDzDvxPZEnNMcscMkyNJrjnx9vZG79698cYbb+DcuXO4ffs2NBoNpk2bhvnz5wMAnnzySQghsGLFCrzyyivw8PBAjx49cPjwYQQFBdW5bXbbZC/MMckdM0yOpBDOfNtDC+l0OqhUKmi1Wtn8cZ04cUJWdyQGbHMnTjmeO1uQ4/Ng7QzL9SVPjufOVuT2XNQ3w3LNal0sOW+y+cwJEREROQc2J0RERCQpbE6IiIhIUticOIns7GwIIWQxEdXGWhkmchRLM+zM2JwQERGRpLA5ISIiIklhc0JERESSwuaEiIiIJIXNCREREUkKmxMiIiKSFMndW0cOTp48CW9vb8NjPz8/BAcH49atWzhz5ozJ8lFRUQCAH3/80XA3zxrt27dH8+bN8euvv5rcBtzHxwcdOnRAVVUVTp06ZbLdyMhIuLm54dy5cya3LW/Tpg38/f2h0+kAwHCjLQDw9PREeHg4ACAnJ8fkK2vh4eHw9PTExYsXcfXqVaMxf39/tGnTBtevX0d+fr7RmJubGyIjIwEAubm5uH37ttF4hw4d4OPjg0uXLqGkpMRorEWLFggKCkJ5eblRrQCgUCjQo0cPw3GUl5cbjQcHB8PPzw8lJSW4dOmS0ZhKpUJISIhJLc6OGWaGGwO55LjG3blwthy3bNkSZhNkNq1WKwCYTBMmTBBCCJGfn1/reI0+ffqYjK1fv14IIcS///1vk7GYmJh77vfKlStCCCFGjRplMrZy5UohhBCvvfaayViPHj0MNbm7u5uMnz59WgghxNSpU03G5s2bJ4QQ4uDBgyZjbdq0MWy3TZs2JuMHDx4UQggxb948k7GpU6cKIYQ4ffq0yZi7u7thuz169DAZ37ZtmxBCiJUrV5qMjRo1SgghxLlz5wQAodVqrREF2WKGmeHGQG45zs7OrnU9Z8txzfNnToZ54z8L1Ny06NChQ7Lo1v39/ZGZmYmBAwdiw4YNhg7d2br1kJAQXL16FWq1WjY3CrMVZpgZbgzkluNLly4hOjraKMOA8+W4ZcuWZt/4j82JBeR2J0zgj7thZmdnG/4wnZEcz50tyPF5YIbvkOO5sxW5PRfM8B28KzERERHJFpsTIiIikhQ2J0RERCQpbE6IiIhIUticEBERkaSwOSEiIiJJ4S/EOonoaEdXYD388rtzklOGmVGqzd0ZZk7qxisnREREJClsToiIiEhSJNucKBQK7Nixw9FlENUbM0xyxwyTozi0OYmPj8fjjz9e61hRURGGDx9ueKxQKAyTj48PevbsiYyMDMP4okWLjJapmfbv329YpqysDImJiQgMDISHhwfCw8Oxe/dumx0fNX7MMMkdM0xSJNkPxAYEBJjMS0tLw7Bhw1BWVobly5cjLi4OR44cQd++fQEAERERRn8EANC8eXMAQEVFBYYMGYJWrVph+/btaNu2LQoLC+Hj42P7gyGnxAyT3DHD5CiSbU4UCgU++eQTo46+WbNmCAgIQEBAAN577z1s2bIFO3fuNPxRuLq61vrHBABr167FtWvXkJWVBTc3NwBAUFCQzY+DnBczTHLHDJOjSPYzJ/fj5uYGV1dXk1tB16XmjycxMRH+/v7o0qULli5diqqqKhtXSlQ7ZpjkjhkmW5HslZN70ev1WL58OXQ6HR599FHD/NzcXHh7exsed+7cGceOHQMAnD9/Hl9++SUmTJiA3bt3Iz8/H4mJiaisrMSCBQvq3I9erzc81ul0NjoicjbMMMmdvTJcsy/m2LnIqjkZP348XFxcUF5eDpVKhRUrVhh9WKtTp07YuXOn4bFSqTT8e3V1NVq1aoXVq1fDxcUF0dHRuHz5MpYvX17nH0VqaioWL15suwMip8MMk9zZO8MAc+yMZNWcvPHGGxg8eDB8fX3RqlUrk3F3d3eEhobWum5gYCDc3Nzg4uJimBceHo7i4mJUVFTA3d3dZJ2UlBS8+OKLhsc6nQ4ajcYKR0LOihkmubN3hgHm2BnJqjkJCAioM/T3069fP2zatAnV1dVo0uTOR23Onj2LwMDAOv8glEqlUddP1FDMMMmdvTMMMMfOyOHNiVarxcmTJ43m1XztzJpmzpyJf/3rX0hKSsKsWbOQn5+PpUuXYvbs2VbfFzkXZpjkjhkmqXF4c5KZmYkePXoYzZs8ebLV96PRaLBv3z688MIL6Nq1K9q0aYOkpCTMnTvX6vsi58IMk9wxwyQ1CiF4X0Rz6XQ6qFQqaLVa+Pr6Orocs5w4cQLR0dEAsgFEObocq6hPYuV47mxBjs+DHDNsi1dVOZ47W5Hbc1FXhp3tv76WnDfZ/s4JERERNU5sToiIiEhS2JwQERGRpLA5cRLZ2Xfe32wMEzknOWWYqDZ3Z5jqxuaEiIiIJMVqzUllZSUKCgqstTkiIiJyUlZrTr7//nsEBwdba3NERETkpPi2DhEREUmK2b8QGxV17x8/Ki8vb3AxRERERGY3J2fOnMG4cePqfOumqKgIZ8+etVphRERE5JzMbk66dOmC3r17Y+bMmbWOnzx5EmvWrLFaYUREROSczP7MycMPP4wff/yxznEfHx888sgjVimKiIiInJfZV05WrVp1z/GQkBAcPHiwofUQERGRkzP7ysmCBQtQVVVV53hBQQGGDBlilaKIiIjIeZndnKSnp6Nnz57Izc01GVu9ejW6dOkCV1ezL8QQERER1crs5uT06dOIjIzEgw8+iNTUVFRXV6OgoACDBw9GcnIyXn/9dXz++ee2rJWIiIicgNmXOnx9ffHhhx9izJgxmDFjBrZu3Yqff/4Zffv2RW5uLjQajS3rJCIiIidh8S/E9u7dG5GRkfjuu+9QXV2N5ORkNiZERERkNRY1J5s3b0ZERASqq6uRl5eHmTNnYvjw4UhKSuIvxBIREZFVmN2cjB07FtOnT8eiRYtw4MABdOrUCcuWLUNmZib27NmDbt264ejRo7aslYiIiJyA2Z85KSoqQk5ODkJDQ43m9+3bF6dOncLcuXPRv39/VFRUWL3IxkShcHQFRKaYSyLbi442nSeE/euQA7Obk8OHD6NJk9ovtHh4eODNN9/EmDFjrFYYEREROSez39apqzH5M/58PRERETWUxd/WISIiIrKlRtGcxMfHQ6FQmEzDhg0zLJOTk4O4uDj4+/vDw8MDHTt2xLRp03D27FkHVk70B+aY5I4ZJmtpFM0JAAwbNgxFRUVG0+bNmwEAu3btQp8+faDX67Fx40bk5eVh/fr1UKlUePXVVx1cOdEfmGOSO2aYrKHR3AxHqVQiICDAZP7vv/+OKVOm4LHHHsMnn3ximB8cHIzevXujrKzMjlUS3RtzTHLHDJM1NJorJ3XZu3cvSktLkZycXOt4s2bN7FsQUT0wxyR3zDBZotFcOdm1axe8vb2N5s2dOxdKpRIAEBYWZvE29Xo99Hq94bFOp2tYkUT3Ye0cM8Nkb3wtJmtoNM3JwIED8e677xrNa968OdasWVPvbaampmLx4sUNLY3IbNbOMTNM9sbXYrKGRvO2jpeXF0JDQ42m5s2bo2PHjgCAH374weJtpqSkQKvVGqbCwkJrl01kxNo5ZobJ3vhaTNbQaJqTusTExECtVmPZsmW1jt/rQ1hKpRK+vr5GE5Ej1DfHzDBJBV+LyRKN5m0dvV6P4uJio3murq5Qq9X44IMPEBcXh9GjR2P27NkIDQ1FaWkptm3bhoKCAmzZssVBVRMZY45J7phhsoZGc+Vkz549CAwMNJoefvhhAEBsbCyysrLg5uaGp59+GmFhYRg/fjy0Wi2WLFni4MqJ/sAck9wxw2QNCiF4T0Rz6XQ6qFQqaLXael9WtP/dX08AiEZ2djaioqLsvXPJsMa5awzqeh6kfVdiZhhghv9Mbs/FiRMnEB0dDSAbgHGGnem/wJact0Zz5YSIiIgaBzYnREREJClsToiIiEhS2JzYmRD2nbKzHX3EJAf2ziUzTM4oO9s031Q7NidEREQkKWxOiIiISFLYnBAREZGksDkhIiIiSWFzQkRERJLC5oSIiIgkhc0JERERSQqbEyIiIpIUNidEREQkKWxOiIiISFLYnBAREZGksDkhIiIiSWFzQkRERJLC5oSIiIgkhc0JERERSQqbEyIiIpIUV0cXQPYRHW3d7Qlh3e0R3U99M8ysklTUZJiZvD9eOSEiIiJJYXNCREREksLmhIiIiCRFUs1JfHw8FAoFFAoF3Nzc4O/vjyFDhmDt2rWorq42LNe+fXsoFAps2bLFZBsRERFQKBRIT083mp+Tk4O4uDj4+/vDw8MDHTt2xLRp03D27FlbHxY5GeaY5I4ZJkeTVHMCAMOGDUNRUREuXLiAzz//HAMHDkRSUhJGjhyJyspKw3IajQZpaWlG63799dcoLi6Gl5eX0fxdu3ahT58+0Ov12LhxI/Ly8rB+/XqoVCq8+uqrdjkuci7MMckdM0wOJSRk8uTJIjY21mT+gQMHBACxZs0aIYQQQUFBYt68eUKpVIqCggLDctOmTROzZs0SKpVKpKWlCSGEuHnzplCr1eLxxx+vdZ+//fab2fVptVoBQGi1WrPXcbTs7GwBQADZ4s5nxK0zyY09z52Uc+yMGW4smOE/yC3Hd2fYWVly3iR35aQ2gwYNQrdu3ZCRkWGY5+/vj6FDh2LdunUAgN9//x1bt25FQkKC0bp79+5FaWkpkpOTa912s2bNbFY30Z8xxyR3zDDZiyyaEwAICwvDhQsXjOYlJCQgPT0dQghs374dISEh6N69u9Ey+fn5hvUtpdfrodPpjCaihrB3jplhsja+FpM9yKY5EUJAoVAYzRsxYgRu3LiB//znP1i7dq1Jp16zXn2lpqZCpVIZJo1GU+9tEQH2zzEzTNbG12KyB9k0J3l5eQgODjaa5+rqimeeeQYLFy7EN998gwkTJpis17FjRwDADz/8YPE+U1JSoNVqDVNhYWH9iif6f/bOMTNM1sbXYrIHWTQnX375JXJzczFmzBiTsYSEBBw6dAixsbHw8/MzGY+JiYFarcayZctq3XZZWVmd+1UqlfD19TWaiOrLETlmhsma+FpM9iK5e+vo9XoUFxejqqoKJSUl2LNnD1JTUzFy5EhMmjTJZPnw8HCUlpaiadOmtW7Py8sLH3zwAeLi4jB69GjMnj0boaGhKC0txbZt21BQUFDrd/SJGoI5JrljhsmRJNec7NmzB4GBgXB1dYWfnx+6deuGt956C5MnT0aTJrVf6GnRosU9txkbG4usrCykpqbi6aefhk6ng0ajwaBBg7BkyRJbHAY5OeaY5I4ZJkdSiIZ8SsnJ6HQ6qFQqaLVa2VxWPHHiBKKjowFkA4iy2nbllho5njtbkOPz0NAMyy2rdZHjubMVuT0Xd2e4sWTSUpacN1l85oSIiIicB5sTIiIikhQ2J0RERCQpbE6cRHa2Ne+s4+ijIWdU3wwTSUVNhun+2JwQERGRpLA5ISIiIklhc0JERESSwuaEiIiIJIXNCREREUkKmxMiIiKSFMndW0fKan7pX6fTObgS8924ccPwTznVbW01x+7sd2tghuWLGf6D3HLMDN9hSYZ5bx0L/PLLL9BoNI4ugxqgsLAQbdu2dXQZDsMMy5+zZxhgjuXOnAyzObFAdXU1Ll++DB8fHygUCsP8mjtrFhYWSu4mVKztDiEErl+/jtatW9d5R1VnwAxbFzPsGLXlmDmpP3vVZ0mG+baOBZo0aXLPbs/X11eSwQNYGwCoVCqb70PqmGHbYIbt6145Zk7qzx71mZth526/iYiISHLYnBAREZGksDmxAqVSiYULF0KpVDq6FBOsjcwh5XPB2sgcUj4XUq4NkGZ9/EAsERERSQqvnBAREZGksDkhIiIiSWFzQkRERJLC5oSIiIgkhc2JlbVv3x4KhcJomjdvnsPqeeeddxAcHAwPDw9ER0fj8OHDDqulxqJFi0yeo4CAAEeXRf+PGb4/Zlj6pJRjZthy/IVYG/jHP/6BadOmGR57e3s7pI6tW7dizpw5eOedd9CvXz+8//77GD58OM6cOYN27do5pKYaERER2L9/v+Gxi4uLA6uhuzHD98cMS58UcswM1w+bExvw8fGRRAf6+uuvY+rUqfjb3/4GAFi1ahX27t2Ld999F6mpqQ6tzdXVVRLPEdWOGb4/Zlj6pJBjZrh++LaODfzzn/9EixYt0L17d/z3f/83Kioq7F5DRUUFsrOzERMTYzQ/JiYGWVlZdq/nbvn5+WjdujWCg4Mxbtw4nD9/3tEl0Z8ww/fHDEufo3PMDNcfr5xYWVJSEqKiouDn54djx44hJSUFP//8Mz744AO71lFaWoqqqir4+/sbzff390dxcbFda7lb79698eGHH6Jjx44oKSnBkiVL8NBDD+H7779HixYtHFobMcPmYIalTwo5ZoYbQNB9LVy4UAC453T8+PFa192+fbsAIEpLS+1a86VLlwQAkZWVZTR/yZIlolOnTnat5X5u3Lgh/P39xcqVKx1dSqPFDNsWM2wfcssxM1x/vHJihueffx7jxo275zLt27evdX6fPn0AAD/99JNdu1G1Wg0XFxeT7vzKlSsmXbyjeXl5ITIyEvn5+Y4updFihm2LGbYPueWYGa4/NidmUKvVUKvV9Vo3JycHABAYGGjNku7L3d0d0dHR+OKLL/DXv/7VMP+LL75AbGysXWu5H71ej7y8PPzlL39xdCmNFjNsW8ywfcgtx8xwAzj60k1jkpWVJV5//XWRk5Mjzp8/L7Zu3Spat24tRo8e7ZB6tmzZItzc3MT//M//iDNnzog5c+YILy8vceHCBYfUU+Oll14SmZmZ4vz58+Lrr78WI0eOFD4+Pg6vi5hhczHD0ialHDPD9cPmxIqys7NF7969hUqlEh4eHqJTp05i4cKF4ubNmw6r6e233xZBQUHC3d1dREVFiUOHDjmslhpPPfWUCAwMFG5ubqJ169biiSeeEN9//72jyyLBDJuLGZY2qeWYGbacQgghHH31hoiIiKgGf+eEiIiIJIXNCREREUkKmxMiIiKSFDYnREREJClsToiIiEhS2JwQERGRpLA5ISIiIklhc0JERESSwuaEiIiIJIXNCREREUkKmxMiIiKSFDYnREREJCn/BxrenDkmSLo/AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 600x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to visualise the GP on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Do you want to visualise the learnt preference model on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Type 0 or 1 which you think larger. Type 2 if unsure.0\n",
      "You chose X0\n",
      "Probability of correct selection: 3.78e-01 ± 2.33e-01\n",
      "Estimated total correct selection rate amongst sure samples: 7.24e-01 ± 2.12e-01\n",
      "Is your selection correct? Yes if 1: 1.0\n",
      "Is your selection sure? Yes if 1: 1\n",
      "9) Best value: 1.80121e+01\n",
      "4) parameters: beta 1.897e+00 gamma 1.600e-01\n",
      "training models...\n",
      "generating candidates...\n",
      "X0 (preference): tensor([1.4490, 0.3178, 0.9943])\n",
      "X1 (normal UCB): tensor([0.9479, 0.5200, 0.2496])\n",
      "Do you need more explanation on the two candidates?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x400 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to know the attribution to each feature?\n",
      "Type y if yes, otherwise type n as noy\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████| 7/7 [00:00<00:00, 3466.78it/s]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 600x400 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Do you want to visualise the GP on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Do you want to visualise the learnt preference model on the plane where two candidates are placed?\n",
      "Type y if yes, otherwise type n as non\n",
      "Type 0 or 1 which you think larger. Type 2 if unsure.0\n",
      "You chose X0\n",
      "Probability of correct selection: 3.47e-01 ± 2.72e-01\n",
      "Estimated total correct selection rate amongst sure samples: 7.26e-01 ± 2.08e-01\n",
      "Is your selection correct? Yes if 1: 1.0\n",
      "Is your selection sure? Yes if 1: 1\n",
      "10) Best value: 1.80121e+01\n"
     ]
    }
   ],
   "source": [
    "results = []\n",
    "for t in range(n_iterations):\n",
    "    beta, gamma = state(t)\n",
    "    result, dataset_obj, dataset_duel = coexbo(\n",
    "        dataset_obj, dataset_duel, beta, gamma,\n",
    "    )\n",
    "    print(f\"{len(dataset_obj[0])}) Best value: {dataset_obj[2].max().item():.5e}\")\n",
    "    results.append(result)\n",
    "\n",
    "results = torch.tensor(results)\n",
    "\n",
    "all_results = (dataset_obj, dataset_duel, results)\n",
    "with open(file_save_name, 'wb') as handle:\n",
    "    pickle.dump(all_results, handle)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b615b72",
   "metadata": {},
   "source": [
    "# Results\n",
    "[overhead(s), best observation, Euclidean distance, correct answer, sure answer]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d260c915",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[55.6231, 16.8844,  0.5581,  1.0000,  1.0000],\n",
       "        [21.5042, 16.8844,  0.8105,  1.0000,  1.0000],\n",
       "        [18.6132, 16.8844,  1.0282,  1.0000,  1.0000],\n",
       "        [16.4089, 18.0121,  0.7665,  1.0000,  1.0000],\n",
       "        [20.2186, 18.0121,  1.1312,  1.0000,  1.0000]])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0b3c76eb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}