{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import sys\n",
"sys.path.append('../..')\n",
"from hippocampus.environments import SimpleMDP, HexWaterMaze, TwoStepTask\n",
"from hippocampus.experiments.reliability_in_twostep import CombinedAgent\n",
"from definitions import FIGURE_FOLDER\n",
"# TODO: do this on linear track "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np\n",
"import seaborn as sns\n",
"import matplotlib.pyplot as plt\n",
"from tqdm.notebook import tqdm\n",
"import pandas as pd\n",
"import os\n",
"import matplotlib"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"sns.palplot(sns.color_palette())"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"ag = CombinedAgent(env=SimpleMDP(5, reward_probability=.85),inv_temp=10)\n",
"\n",
"init_p_sr = .5\n",
"ag.p_sr = init_p_sr\n",
"\n",
"ag.HPC.learning_rate =.01"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"\n",
"df = pd.DataFrame({})\n",
"for ep in tqdm(range(1,400)):\n",
" results = ag.one_episode(deterministic_policy=False)\n",
" results['trial'] = ep \n",
" df = df.append(results, ignore_index=True)\n",
" ag.HPC.learning_rate *=.95"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"plt.plot(np.array([df['omega'].iloc[i] for i in range(len(df))]))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"dls_reliab = pd.concat([pd.Series([0.]), df['DLS reliability']])\n",
"hpc_reliab = pd.concat([pd.Series([0.]), df['HPC reliability']])\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"font = {'size': 22}\n",
"\n",
"matplotlib.rc('font', **font)\n",
"\n",
"\n",
"fig, ax = plt.subplots()\n",
"#df.plot(ax=ax, x='trial', y=['DLS reliability', 'HPC reliability'])\n",
"ax.plot(dls_reliab, color=sns.color_palette()[1], linewidth=2)\n",
"ax.plot(hpc_reliab, color=sns.color_palette()[2], linewidth=2)\n",
"\n",
"# Move left and bottom spines outward by 10 points\n",
"ax.spines['left'].set_position(('outward', 10))\n",
"ax.spines['bottom'].set_position(('outward', 10))\n",
"# Hide the right and top spines\n",
"ax.spines['right'].set_visible(False)\n",
"ax.spines['top'].set_visible(False)\n",
"# Only show ticks on the left and bottom spines\n",
"ax.yaxis.set_ticks_position('left')\n",
"ax.xaxis.set_ticks_position('bottom')\n",
"\n",
"plt.ylabel('Reliability')\n",
"plt.xlabel('Trial')\n",
"plt.legend(['DLS reliability', 'HPC reliability'])\n",
"plt.ylim([-.1,1])\n",
"plt.tight_layout()\n",
"plt.savefig(os.path.join(FIGURE_FOLDER, 'reliability.pdf'))\n",
"#plt.xlim([-10,150])\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"#df.plot(x='trial', y='P(SR)')\n",
"fig, ax = plt.subplots()\n",
"\n",
"font = {'size': 22}\n",
"\n",
"\n",
"ax.plot(pd.concat([pd.Series([init_p_sr]), df['P(SR)']]), color=sns.color_palette()[0],linewidth=2)\n",
"\n",
"# Move left and bottom spines outward by 10 points\n",
"ax.spines['left'].set_position(('outward', 10))\n",
"ax.spines['bottom'].set_position(('outward', 10))\n",
"# Hide the right and top spines\n",
"ax.spines['right'].set_visible(False)\n",
"ax.spines['top'].set_visible(False)\n",
"# Only show ticks on the left and bottom spines\n",
"ax.yaxis.set_ticks_position('left')\n",
"ax.xaxis.set_ticks_position('bottom')\n",
"\n",
"plt.ylabel('Pr(HPC)')\n",
"plt.xlabel('Trial')\n",
"\n",
"plt.ylim([0,1])\n",
"plt.tight_layout()\n",
"plt.savefig(os.path.join(FIGURE_FOLDER, 'psr.pdf'))\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"n_agents = 10\n",
"n_trials = 300\n",
"dls_reliab_M = np.zeros((n_agents, n_trials))\n",
"hpc_reliab_M = np.zeros((n_agents, n_trials))\n",
"psr_M = np.zeros((n_agents, n_trials))\n",
"\n",
"for ia in tqdm(range(n_agents)):\n",
"\n",
" ag = CombinedAgent(env=SimpleMDP(5, reward_probability=.8))\n",
"\n",
" init_p_sr = .5\n",
" ag.p_sr = init_p_sr\n",
" ag.HPC.learning_rate=.01\n",
" df = pd.DataFrame({})\n",
" for ep in tqdm(range(1, n_trials),leave=False):\n",
" results = ag.one_episode()\n",
" results['trial'] = ep \n",
" df = df.append(results, ignore_index=True)\n",
"\n",
" dls_reliab_M[ia,:] = pd.concat([pd.Series([0.]), df['DLS reliability']])\n",
" hpc_reliab_M[ia,:] = pd.concat([pd.Series([0.]), df['HPC reliability']])\n",
" psr_M[ia,:] = pd.concat([pd.Series([init_p_sr]), df['P(SR)']])\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"fig, ax = plt.subplots()\n",
"#df.plot(ax=ax, x='trial', y=['DLS reliability', 'HPC reliability'])\n",
"ax.plot(dls_reliab_M.mean(axis=0))\n",
"ax.plot(hpc_reliab_M.mean(axis=0))\n",
"plt.ylim([-.1,1])\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"fig, ax = plt.subplots()\n",
"#df.plot(ax=ax, x='trial', y=['DLS reliability', 'HPC reliability'])\n",
"ax.plot(dls_reliab_M.mean(axis=0))\n",
"ax.plot(hpc_reliab_M.mean(axis=0))\n",
"plt.ylim([-.1,1])\n",
"plt.xlim([-10,150])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"omegas = df['omega']"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"alloms = np.concatenate(np.array(df['omega'])).reshape(ag.env.nr_states, -1)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"plt.plot(alloms[4])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"df.plot(x='trial', y='omega_dls')\n",
"plt.ylim([0,1])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"omg0 = [df['omega'][i][0] for i in range(272)]\n",
"omg1 = [df['omega'][i][1] for i in range(272)]\n",
"omg2 = [df['omega'][i][2] for i in range(272)]\n",
"omg3 = [df['omega'][i][3] for i in range(272)]\n",
"omg4 = [df['omega'][i][4] for i in range(272)]\n",
"omg5 = [df['omega'][i][5] for i in range(272)]\n",
"omg6 = [df['omega'][i][6] for i in range(272)]\n",
"omg7 = [df['omega'][i][7] for i in range(272)]\n",
"omg8 = [df['omega'][i][8] for i in range(272)]\n",
"\n",
"\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"plt.plot(omg0)\n",
"plt.plot(omg1)\n",
"plt.plot(omg2)\n",
"plt.plot(omg3)\n",
"plt.plot(omg4)\n",
"plt.plot(omg5)\n",
"plt.plot(omg6)\n",
"plt.plot(omg7)\n",
"plt.plot(omg8)\n",
"\n",
"\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"alloms = np.concatenate([omg0, omg1, omg2, omg3, omg4, omg5, omg6, omg7, omg8]).reshape( -1, len(omg0))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"plt.plot(alloms.mean(axis=0))\n",
"plt.plot(df['omega_dls'])\n",
"plt.ylim([0,1])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"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": "3.6.8"
}
},
"nbformat": 4,
"nbformat_minor": 2
}