human_rl_v2.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/lynn0503/NMA2022DL_Mohinga_Project/blob/main/human_rl_v2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "SllQYWzCHA-L"
      },
      "source": [
        "# Using RL to Model Cognitive Tasks\n",
        "\n",
        "**By Neurmatch Academy**\n",
        "\n",
        "__Content creators:__ Morteza Ansarinia, Yamil Vidal\n",
        "\n",
        "__Production editor:__ Spiros Chavlis\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "84B_IMDlHA-T"
      },
      "source": [
        "<p align='center'><img src='https://github.com/NeuromatchAcademy/widgets/blob/master/sponsors.png?raw=True'/></p>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "DUcgD2CGHA-U"
      },
      "source": [
        "---\n",
        "# Objective\n",
        "\n",
        "- This project aims to use behavioral data to train an agent and then use the agent to investigate data produced by human subjects. Having a computational agent that mimics humans in such tests, we will be able to compare its mechanics with human data.\n",
        "\n",
        "- In another conception, we could fit an agent that learns many cognitive tasks that require abstract-level constructs such as executive functions. This is a multi-task control problem.\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "dqlJguOgHA-W"
      },
      "source": [
        "---\n",
        "# Setup"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "execution": {},
        "id": "ge0IRlo_HA-Y",
        "outputId": "1c47a2ea-6769-4447-8cf5-726b75e41414",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[?25l\r\u001b[K     |█                               | 10 kB 28.2 MB/s eta 0:00:01\r\u001b[K     |██                              | 20 kB 21.9 MB/s eta 0:00:01\r\u001b[K     |███▏                            | 30 kB 16.5 MB/s eta 0:00:01\r\u001b[K     |████▏                           | 40 kB 14.6 MB/s eta 0:00:01\r\u001b[K     |█████▏                          | 51 kB 7.1 MB/s eta 0:00:01\r\u001b[K     |██████▎                         | 61 kB 8.4 MB/s eta 0:00:01\r\u001b[K     |███████▎                        | 71 kB 9.0 MB/s eta 0:00:01\r\u001b[K     |████████▍                       | 81 kB 8.5 MB/s eta 0:00:01\r\u001b[K     |█████████▍                      | 92 kB 9.4 MB/s eta 0:00:01\r\u001b[K     |██████████▍                     | 102 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████▌                    | 112 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████▌                   | 122 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |█████████████▋                  | 133 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |██████████████▋                 | 143 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████████▋                | 153 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████████▊               | 163 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |█████████████████▊              | 174 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |██████████████████▉             | 184 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████████████▉            | 194 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████████████▉           | 204 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |██████████████████████          | 215 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████████████████         | 225 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 235 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████       | 245 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████      | 256 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▏    | 266 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▏   | 276 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▏  | 286 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▎ | 296 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▎| 307 kB 7.8 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 314 kB 7.8 MB/s \n",
            "\u001b[33mWARNING: dm-acme 0.4.0 does not provide the extra 'tensorflow'\u001b[0m\n",
            "\u001b[K     |████████████████████████████████| 3.8 MB 41.6 MB/s \n",
            "\u001b[K     |████████████████████████████████| 72 kB 644 kB/s \n",
            "\u001b[K     |████████████████████████████████| 342 kB 48.3 MB/s \n",
            "\u001b[K     |████████████████████████████████| 197 kB 52.8 MB/s \n",
            "\u001b[K     |████████████████████████████████| 145 kB 56.1 MB/s \n",
            "\u001b[K     |████████████████████████████████| 117 kB 54.0 MB/s \n",
            "\u001b[K     |████████████████████████████████| 6.4 MB 23.8 MB/s \n",
            "\u001b[K     |████████████████████████████████| 4.0 MB 42.7 MB/s \n",
            "\u001b[K     |████████████████████████████████| 668.3 MB 18 kB/s \n",
            "\u001b[K     |████████████████████████████████| 5.7 MB 28.9 MB/s \n",
            "\u001b[K     |████████████████████████████████| 462 kB 51.5 MB/s \n",
            "\u001b[K     |████████████████████████████████| 596 kB 53.3 MB/s \n",
            "\u001b[K     |████████████████████████████████| 217 kB 41.6 MB/s \n",
            "\u001b[K     |████████████████████████████████| 51 kB 5.5 MB/s \n",
            "\u001b[K     |████████████████████████████████| 272 kB 39.8 MB/s \n",
            "\u001b[?25h  Building wheel for dm-acme (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[K     |████████████████████████████████| 18.8 MB 3.4 MB/s \n",
            "\u001b[K     |████████████████████████████████| 88 kB 8.6 MB/s \n",
            "\u001b[K     |████████████████████████████████| 4.9 MB 38.7 MB/s \n",
            "\u001b[K     |████████████████████████████████| 206 kB 47.0 MB/s \n",
            "\u001b[K     |████████████████████████████████| 99 kB 8.6 MB/s \n",
            "\u001b[?25h  Building wheel for bsuite (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "\u001b[K     |████████████████████████████████| 254 kB 8.2 MB/s \n",
            "\u001b[K     |████████████████████████████████| 104 kB 8.7 MB/s \n",
            "\u001b[?25h"
          ]
        }
      ],
      "source": [
        "# @title Install dependencies\n",
        "# !pip install --upgrade pip setuptools wheel --quiet\n",
        "!pip install dm-acme[jax,tensorflow] --quiet\n",
        "!pip install dm-acme[envs] --quiet\n",
        "!pip install dm-sonnet --quiet\n",
        "!pip install trfl --quiet"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "execution": {},
        "id": "lYr0Rp3oHA-e"
      },
      "outputs": [],
      "source": [
        "# Imports\n",
        "\n",
        "import time\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import sonnet as snt\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import dm_env\n",
        "\n",
        "import acme\n",
        "from acme import specs\n",
        "from acme import wrappers\n",
        "from acme import EnvironmentLoop\n",
        "from acme.agents.tf import dqn\n",
        "from acme.utils import loggers\n",
        "from acme.utils.observers import action_norm \n",
        "from tqdm import tqdm\n",
        "from datetime import datetime "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "cellView": "form",
        "execution": {},
        "id": "h6qra9eCHA-e"
      },
      "outputs": [],
      "source": [
        "# @title Figure settings\n",
        "from IPython.display import clear_output, display, HTML\n",
        "%matplotlib inline\n",
        "sns.set()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "amLHDeiXHA-f"
      },
      "source": [
        "---\n",
        "# Background\n",
        "\n",
        "- Cognitive scientists use standard lab tests to tap into specific processes in the brain and behavior. Some examples of those tests are Stroop, N-back, Digit Span, TMT (Trail making tests), and WCST (Wisconsin Card Sorting Tests).\n",
        "\n",
        "- Despite an extensive body of research that explains human performance using descriptive what-models, we still need a more sophisticated approach to gain a better understanding of the underlying processes (i.e., a how-model).\n",
        "\n",
        "- Interestingly, many of such tests can be thought of as a continuous stream of stimuli and corresponding actions, that is in consonant with the RL formulation. In fact, RL itself is in part motivated by how the brain enables goal-directed behaviors using reward systems, making it a good choice to explain human performance.\n",
        "\n",
        "- One behavioral test example would be the N-back task.\n",
        "\n",
        "  - In the N-back, participants view a sequence of stimuli, one by one, and are asked to categorize each stimulus as being either match or non-match. Stimuli are usually numbers, and feedback is given at both timestep and trajectory levels.\n",
        "\n",
        "  - The agent is rewarded when its response matches the stimulus that was shown N steps back in the episode. A simpler version of the N-back uses two-choice action schema, that is match vs non-match. Once the present stimulus matches the one presented N step back, then the agent is expected to respond to it as being a `match`.\n",
        "\n",
        "\n",
        "- Given a trained RL agent, we then find correlates of its fitted parameters with the brain mechanisms. The most straightforward composition could be the correlation of model parameters with the brain activities."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "jSi9vuPjHA-g"
      },
      "source": [
        "## Datasets\n",
        "\n",
        "- HCP WM task ([NMA-CN HCP notebooks](https://github.com/NeuromatchAcademy/course-content/tree/master/projects/fMRI))\n",
        "\n",
        "Any dataset that used cognitive tests would work.\n",
        "Question: limit to behavioral data vs fMRI?\n",
        "Question: Which stimuli and actions to use?\n",
        "classic tests can be modeled using 1) bounded symbolic stimuli/actions (e.g., A, B, C), but more sophisticated one would require texts or images (e.g., face vs neutral images in social stroop dataset)\n",
        "The HCP dataset from NMA-CN contains behavioral and imaging data for 7 cognitive tests including various versions of N-back."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "IKeEqNEZHA-h"
      },
      "source": [
        "## N-back task\n",
        "\n",
        "In the N-back task, participants view a sequence of stimuli, one per time, and are asked to categorize each stimulus as being either match or non-match. Stimuli are usually numbers, and feedbacks are given at both timestep and trajectory levels.\n",
        "\n",
        "In a typical neuro setup, both accuracy and response time are measured, but here, for the sake of brevity, we focus only on accuracy of responses."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "CANdLa1CHA-i"
      },
      "source": [
        "---\n",
        "# Cognitive Tests Environment\n",
        "\n",
        "First we develop an environment in that agents perform a cognitive test, here the N-back."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "ZJW-_yYdHA-j"
      },
      "source": [
        "## Human dataset\n",
        "\n",
        "We need a dataset of human perfoming a N-back test, with the following features:\n",
        "\n",
        "- `participant_id`: following the BIDS format, it contains a unique identifier for each participant.\n",
        "- `trial_index`: same as `time_step`.\n",
        "- `stimulus`: same as `observation`.\n",
        "- `response`: same as `action`, recorded response by the human subject.\n",
        "- `expected_response`: correct response.\n",
        "- `is_correct`: same as `reward`, whether the human subject responded correctly.\n",
        "- `response_time`: won't be used here.\n",
        "\n",
        "Here we generate a mock dataset with those features, but remember to **replace this with real human data.**"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### generate mock data"
      ],
      "metadata": {
        "id": "3w1cnw25HBxT"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 149,
      "metadata": {
        "execution": {},
        "id": "-LAaobwDHA-j"
      },
      "outputs": [],
      "source": [
        "def generate_mock_nback_dataset(N=2,\n",
        "                                n_participants=10,\n",
        "                                n_trials=32,\n",
        "                                stimulus_choices=list('ABCDEF'),\n",
        "                                response_choices=['match', 'non-match']):\n",
        "  \"\"\"Generate a mock dataset for the N-back task.\"\"\"\n",
        "\n",
        "  n_rows = n_participants * n_trials\n",
        "\n",
        "  participant_ids = sorted([f'sub-{pid}' for pid in range(1, n_participants + 1)] * n_trials)\n",
        "  trial_indices = list(range(1, n_trials + 1)) * n_participants\n",
        "  stimulus_sequence = np.random.choice(stimulus_choices, n_rows)\n",
        "\n",
        "  # change stimuluse_sequence to control match and non-match number\n",
        "  # stimulus_sequence=np.ones(n_rows)\n",
        "  # match prob will be slightly higher than non-match prob\n",
        "  # for i in range(n_rows):\n",
        "  #   if i<N:\n",
        "  #       stimulus_sequence[i]=np.random.choice(len(stimulus_choices))\n",
        "  #   else:\n",
        "  #       rand=np.random.random()\n",
        "  #       if rand < 0.4:\n",
        "  #           stimulus_sequence[i]=stimulus_sequence[i-N]\n",
        "  #       else:\n",
        "  #           stimulus_sequence[i]=np.random.choice(len(stimulus_choices))\n",
        "\n",
        "\n",
        "  responses = np.random.choice(response_choices, n_rows)\n",
        "  response_times = np.random.exponential(size=n_rows)\n",
        "\n",
        "  df = pd.DataFrame({\n",
        "      'participant_id': participant_ids,\n",
        "      'trial_index': trial_indices,\n",
        "      'stimulus': stimulus_sequence,\n",
        "      'response': responses,\n",
        "      'response_time': response_times\n",
        "  })\n",
        "\n",
        "  # mark matchig stimuli\n",
        "  _nback_stim = df['stimulus'].shift(N)\n",
        "  df['expected_response'] = (df['stimulus'] == _nback_stim).map({True: 'match', False: 'non-match'})\n",
        "\n",
        "  df['is_correct'] = (df['response'] == df['expected_response'])\n",
        "\n",
        "  # we don't care about burn-in trials (trial < N)\n",
        "  df.loc[df['trial_index'] <= N, 'is_correct'] = True\n",
        "  df.loc[df['trial_index'] <= N, ['response', 'response_time', 'expected_response']] = None\n",
        "\n",
        "  return df\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Analyze mock data"
      ],
      "metadata": {
        "id": "9OUzQxO1G7L7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "# ========\n",
        "# now generate the actual data with the provided function and plot some of its features\n",
        "N=2\n",
        "episode_steps=500\n",
        "mock_nback_data = generate_mock_nback_dataset()\n",
        "mock_nback_data['is_correct'] = mock_nback_data['is_correct'].astype(int)\n",
        "\n",
        "\n",
        "# sns.displot(data=mock_nback_data, x='response_time')\n",
        "# plt.suptitle('response time distribution of the mock N-back dataset', y=1.01)\n",
        "# plt.show()\n",
        "\n",
        "# sns.displot(data=mock_nback_data, x='is_correct')\n",
        "# plt.suptitle('Accuracy distribution of the mock N-back dataset', y=1.06)\n",
        "# plt.show()\n",
        "\n",
        "sns.barplot(data=mock_nback_data, y='is_correct', x='participant_id')\n",
        "plt.suptitle('Accuracy distribution of the mock N-back dataset', y=1.06)\n",
        "plt.show()\n",
        "\n",
        "# mock_nback_data1.head()\n",
        "\n",
        "# sns.displot(data=mock_nback_data, x='expected_response')\n",
        "# plt.suptitle('correct-action distribution of the mock N-back dataset', y=1.06)\n",
        "# plt.show()\n"
      ],
      "metadata": {
        "id": "LxkfEVNHF2Jv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "outputId": "a669ada8-326f-4bab-ef44-ea8b5bc86a1c"
      },
      "execution_count": 162,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "dG-N0OPuHA-k"
      },
      "source": [
        "## Implementation scheme\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "G-po6h3IHA-k"
      },
      "source": [
        "### Environment\n",
        "\n",
        "The following cell implments N-back envinronment, that we later use to train a RL agent on human data. It is capable of performing two kinds of simulation:\n",
        "- rewards the agent once the action was correct (i.e., a normative model of the environment).\n",
        "- receives human data (or mock data if you prefer), and returns what participants performed as the observation. This is more useful for preference-based RL."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 182,
      "metadata": {
        "execution": {},
        "id": "0bW8nFsQHA-k"
      },
      "outputs": [],
      "source": [
        "class NBack(dm_env.Environment):\n",
        "\n",
        "  ACTIONS = ['match', 'non-match']\n",
        "\n",
        "  def __init__(self,\n",
        "               N=2,\n",
        "               episode_steps=32,\n",
        "               n_episodes=100,\n",
        "               stimuli_choices=list('ABCDEFGHIJKLMN'),\n",
        "               human_data=None,\n",
        "               seed=1,\n",
        "               ):\n",
        "    \"\"\"\n",
        "    Args:\n",
        "      N: Number of steps to look back for the matched stimuli. Defaults to 2 (as in 2-back).\n",
        "      episode_steps\n",
        "      stimuli_choices\n",
        "      human_data\n",
        "      seed\n",
        "\n",
        "    \"\"\"\n",
        "    self.N = N\n",
        "    self.episode_steps = episode_steps\n",
        "    self.stimuli_choices = stimuli_choices\n",
        "    self.stimuli = np.empty(shape=episode_steps)  # will be filled in the `reset()`\n",
        "    self.current_episode = 0\n",
        "    self._reset_next_step = True\n",
        "    self.n_episodes=n_episodes\n",
        "    # whether mimic humans or reward the agent once it responds optimally.\n",
        "    if human_data is None:\n",
        "      self._imitate_human = False\n",
        "      self.human_data = None\n",
        "      self.human_subject_data = None\n",
        "    else:\n",
        "      self._imitate_human = True\n",
        "      self.human_data = human_data\n",
        "      self.human_subject_data = None\n",
        "\n",
        "    self._action_history = []\n",
        "    self._reward_history = []\n",
        "\n",
        "\n",
        "  def reset(self):\n",
        "\n",
        "    self._reset_next_step = False\n",
        "    self._current_step = 0\n",
        "    self._action_history.clear()\n",
        "    self._reward_history.clear()\n",
        "\n",
        "    # generate a random sequence instead of relying on human data\n",
        "    if self.human_data is None:\n",
        "      # self.stimuli = np.random.choice(self.stimuli_choices, self.episode_steps)\n",
        "      # FIXME This is a fix for acme & reverb issue with string observation. Agent should be able to handle strings\n",
        "      self.stimuli = np.random.choice(len(self.stimuli_choices), self.episode_steps).astype(np.float32)\n",
        "      # self.stimuli = np.empty(episode_steps)\n",
        "      # # match prob will be slightly higher than non-match prob\n",
        "      # for i in range(episode_steps):\n",
        "      #   if i<N:\n",
        "      #       self.stimuli[i]=np.random.choice(len(self.stimuli_choices))\n",
        "      #   else:\n",
        "      #       rand=np.random.random()\n",
        "      #       if rand < 0.4:\n",
        "      #           self.stimuli[i]=self.stimuli[i-N]\n",
        "      #       else:\n",
        "      #           self.stimuli[i]=np.random.choice(len(self.stimuli_choices))\n",
        "      # print(self.stimuli)\n",
        "    else:\n",
        "      # randomly choose a subject from the human data and follow her trials and responses.\n",
        "      # FIXME should we always use one specific human subject or randomly select one in each episode?\n",
        "      self.human_subject_data = self.human_data.query('participant_id == participant_id.sample().iloc[0]',\n",
        "                                                engine='python').sort_values('trial_index')\n",
        "      self.stimuli = self.human_subject_data['stimulus'].to_list()\n",
        "      self.stimuli = np.array([ord(s) - ord('A') + 1 for s in self.stimuli]).astype(np.float32)\n",
        "\n",
        "    return dm_env.restart(self._observation())\n",
        "\n",
        "\n",
        "  def _episode_return(self):\n",
        "    if self._imitate_human:\n",
        "      return np.mean(self.human_subject_data['response'] == self._action_history)\n",
        "    else:\n",
        "      return 0.0\n",
        "    \n",
        "\n",
        "  def step(self, action: int):\n",
        "    if self._reset_next_step:\n",
        "      return self.reset()\n",
        "\n",
        "    agent_action = NBack.ACTIONS[action]\n",
        "\n",
        "    if self._imitate_human:\n",
        "      # if it was the same action as the human subject, then reward the agent\n",
        "      human_action = self.human_subject_data['response'].iloc[self._current_step]\n",
        "      step_reward = 0. if (agent_action == human_action) else -1.\n",
        "    else:\n",
        "      # assume the agent is rationale and doesn't want to reproduce human, reward once the response it correct\n",
        "      expected_action = 'match' if (self.stimuli[self._current_step] == self.stimuli[self._current_step - self.N]) else 'non-match'\n",
        "      step_reward = 0. if (agent_action == expected_action) else -1.\n",
        "      \n",
        "    self._action_history.append(agent_action)\n",
        "    self._reward_history.append(step_reward)\n",
        "\n",
        "    self._current_step += 1\n",
        "    # Check for termination.\n",
        "    if self._current_step == self.stimuli.shape[0]:\n",
        "      self._reset_next_step = True\n",
        "      self.current_episode += 1\n",
        "      # we are using the mean of total time step rewards as the episode return\n",
        "      # saving data\n",
        "      df=pd.DataFrame()\n",
        "      # print(len(self.stimuli),len(self._action_history))\n",
        "      df['episode']=np.ones(self.episode_steps)*self.current_episode\n",
        "      df['stimuli']=self.stimuli\n",
        "      df['action']=self._action_history\n",
        "      df['reward']=self._reward_history\n",
        "      if self.current_episode == 1:\n",
        "        mode='w'\n",
        "        header=True\n",
        "      else:\n",
        "        mode='a'\n",
        "        header=False\n",
        "      # df.to_csv('data_lr.csv',header=header,index_label='trial',mode=mode)\n",
        "      \n",
        "      if self.current_episode==self.n_episodes:\n",
        "        self.current_episode=0\n",
        "\n",
        "      return dm_env.termination(reward=self._episode_return(),\n",
        "                                observation=self._observation())\n",
        "    else:\n",
        "      return dm_env.transition(reward=step_reward,\n",
        "                               observation=self._observation())\n",
        "\n",
        "  def observation_spec(self):\n",
        "    return dm_env.specs.BoundedArray(\n",
        "        shape=self.stimuli.shape,\n",
        "        dtype=self.stimuli.dtype,\n",
        "        name='nback_stimuli', minimum=0, maximum=len(self.stimuli_choices) + 1)\n",
        "\n",
        "  def action_spec(self):\n",
        "    return dm_env.specs.DiscreteArray(\n",
        "        num_values=len(NBack.ACTIONS),\n",
        "        dtype=np.int32,\n",
        "        name='action')\n",
        "\n",
        "  def _observation(self):\n",
        "\n",
        "    # agent observes only the current trial\n",
        "    # obs = self.stimuli[self._current_step - 1]\n",
        "\n",
        "    # agents observe stimuli up to the current trial\n",
        "    obs = self.stimuli[:self._current_step+1].copy()\n",
        "    obs = np.pad(obs,(0, len(self.stimuli) - len(obs)))\n",
        "\n",
        "    return obs\n",
        "\n",
        "  def plot_state(self):\n",
        "    \"\"\"Display current state of the environment.\n",
        "\n",
        "     Note: `M` mean `match`, and `.` is a `non-match`.\n",
        "    \"\"\"\n",
        "    stimuli = self.stimuli[:self._current_step - 1]\n",
        "    actions = ['M' if a=='match' else '.' for a in self._action_history[:self._current_step - 1]]\n",
        "    actions_binary = [1 if a=='M' else 0 for a in actions] \n",
        "    expected_actions = stimuli[:-self.N]==stimuli[self.N:]\n",
        "    rewards = [1 if a==e else 0   for a in actions for e in expected_actions]\n",
        "    return HTML(\n",
        "        f'<b>Environment ({self.N}-back):</b><br />'\n",
        "        f'<pre><b>Stimuli:</b> {\"\".join(map(str,map(int,stimuli)))}</pre>'\n",
        "        f'<pre><b>Actions:</b> {\"\".join(actions)}</pre>'\n",
        "        \n",
        "    )\n",
        "    \n",
        "    # f'<pre><b>Rewards:</b> {\"\".join(map(str,reward)}</pre>'\n",
        "    # print(self.N,'back')\n",
        "    # print('stimuli: ',stimuli)\n",
        "    # print('actions: ',actions)\n",
        "    # print('answers: ',expected_actions)\n",
        "    # print('rewards: ',rewards)\n",
        "\n",
        "\n",
        "  @staticmethod\n",
        "  def create_environment(N,episode_steps,n_episodes,human_data):\n",
        "    \"\"\"Utility function to create a N-back environment and its spec.\"\"\"\n",
        "\n",
        "    # Make sure the environment outputs single-precision floats.\n",
        "    environment = wrappers.SinglePrecisionWrapper(NBack(N,episode_steps,n_episodes,human_data=None))\n",
        "\n",
        "    # Grab the spec of the environment.\n",
        "    environment_spec = specs.make_environment_spec(environment)\n",
        "\n",
        "    return environment, environment_spec"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Init ENV"
      ],
      "metadata": {
        "id": "IHwq9hwVvskc"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 120,
      "metadata": {
        "execution": {},
        "id": "OLnRt42yHA-r",
        "outputId": "8aa1ec7c-2557-42af-89e7-eb1ed4552123",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "actions:\n",
            " DiscreteArray(shape=(), dtype=int32, name=action, minimum=0, maximum=1, num_values=2)\n",
            "observations:\n",
            " BoundedArray(shape=(32,), dtype=dtype('float32'), name='nback_stimuli', minimum=0.0, maximum=15.0)\n",
            "rewards:\n",
            " Array(shape=(), dtype=dtype('float32'), name='reward')\n"
          ]
        }
      ],
      "source": [
        "# fitting parameters\n",
        "N=2\n",
        "n_episodes = 200\n",
        "episode_steps= 32\n",
        "# n_total_steps = 0\n",
        "# log_loss = False\n",
        "# n_steps = n_episodes * episode_steps\n",
        "\n",
        "env, env_spec = NBack.create_environment(N=2,episode_steps=episode_steps,n_episodes=n_episodes)\n",
        "# agent = RandomAgent(env_spec)\n",
        "\n",
        "print('actions:\\n', env_spec.actions)\n",
        "print('observations:\\n', env_spec.observations)\n",
        "print('rewards:\\n', env_spec.rewards)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "pFm0KrNPHA-l"
      },
      "source": [
        "### Define a random agent\n",
        "\n",
        "For more information you can refer to NMA-DL W3D2 Basic Reinforcement learning."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 131,
      "metadata": {
        "execution": {},
        "id": "FmQl3jeeHA-l"
      },
      "outputs": [],
      "source": [
        "class RandomAgent(acme.Actor):\n",
        "\n",
        "  def __init__(self, environment_spec):\n",
        "    \"\"\"Gets the number of available actions from the environment spec.\"\"\"\n",
        "    self._num_actions = environment_spec.actions.num_values\n",
        "\n",
        "  def select_action(self, observation):\n",
        "    \"\"\"Selects an action uniformly at random.\"\"\"\n",
        "    action = np.random.randint(self._num_actions)\n",
        "    return action\n",
        "\n",
        "  def observe_first(self, timestep):\n",
        "    \"\"\"Does not record as the RandomAgent has no use for data.\"\"\"\n",
        "    pass\n",
        "\n",
        "  def observe(self, action, next_timestep):\n",
        "    \"\"\"Does not record as the RandomAgent has no use for data.\"\"\"\n",
        "    pass\n",
        "\n",
        "  def update(self):\n",
        "    \"\"\"Does not update as the RandomAgent does not learn from data.\"\"\"\n",
        "    pass"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Define DQN"
      ],
      "metadata": {
        "id": "u5GW44FzN1LB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def dqn_make_network(action_spec: specs.DiscreteArray) -> snt.Module:\n",
        "  return snt.Sequential([\n",
        "      snt.Flatten(),\n",
        "      snt.nets.MLP([50, 50, 50, 50, action_spec.num_values]),\n",
        "  ])\n"
      ],
      "metadata": {
        "id": "CksSLmV8uUZc"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Define DRQN"
      ],
      "metadata": {
        "id": "kd2YnP4yO3vZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# add recurrent layer to dqn\n",
        "# r2d2 receive sequence data\n",
        "from acme.agents.tf import r2d2\n",
        "def r2d2_make_network(action_spec: specs.DiscreteArray) -> snt.Module:\n",
        "  return snt.DeepRNN([\n",
        "            snt.Flatten(),\n",
        "            snt.LSTM(2),\n",
        "            snt.nets.MLP([50, 50, action_spec.num_values])\n",
        "            ])\n"
      ],
      "metadata": {
        "id": "GucpcNWauYrx"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Plot function"
      ],
      "metadata": {
        "id": "wBSMenJ5hpVF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# plot moving average\n",
        "def moving_average(interval, window_size):\n",
        "    window = np.ones(int(window_size)) / float(window_size)\n",
        "    return np.convolve(interval, window, 'same')\n",
        "\n",
        "# plot reward by episode\n",
        "def plot_reward_by_episode(logs_dict):\n",
        "    for key, val in logs_dict.items():\n",
        "        logs=logs_dict[key]\n",
        "        reward_ma=moving_average(logs['episode_return'],10)\n",
        "        plt.plot(reward_ma)\n",
        "\n",
        "    plt.ylabel(\"episode reward\")\n",
        "    plt.xlabel(\"episode\")\n",
        "    plt.legend(['RANDOM','DQN','DRQN'])\n",
        "    plt.title(\"total reward within episode\")\n"
      ],
      "metadata": {
        "id": "Zjy8fcedvZSd"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# test plot function \n",
        "# N = 2\n",
        "# n_episodes = 1000\n",
        "# episode_steps= 32\n",
        "plot_reward_by_episode(logs_dict)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "ari38aBF26C0",
        "outputId": "2b5056c0-5412-47fb-8a60-a094718c1984"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "xz3x4BLhHA-n"
      },
      "source": [
        "## Test agents"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### test Random Agent"
      ],
      "metadata": {
        "id": "doart4swZ5yX"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "execution": {},
        "id": "qI6T1LwqHA-v",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 395
        },
        "outputId": "350f0e96-f938-4364-d2b8-523d54b15289"
      },
      "outputs": [
        {
          "output_type": "error",
          "ename": "ValueError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-126-a8b84f9d43f7>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     11\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     12\u001b[0m   \u001b[0;31m# Make the first observation.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 13\u001b[0;31m   \u001b[0magent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobserve_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtimestep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     14\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     15\u001b[0m   \u001b[0;31m# Run an episode\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/acme/agents/agent.py\u001b[0m in \u001b[0;36mobserve_first\u001b[0;34m(self, timestep)\u001b[0m\n\u001b[1;32m     71\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     72\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mobserve_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtimestep\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mdm_env\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTimeStep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 73\u001b[0;31m     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_actor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobserve_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtimestep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     74\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     75\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mobserve\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mtypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mNestedArray\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnext_timestep\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mdm_env\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTimeStep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/acme/agents/tf/actors.py\u001b[0m in \u001b[0;36mobserve_first\u001b[0;34m(self, timestep)\u001b[0m\n\u001b[1;32m    164\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mobserve_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtimestep\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mdm_env\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTimeStep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    165\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_adder\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 166\u001b[0;31m       \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_adder\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtimestep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    167\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    168\u001b[0m     \u001b[0;31m# Set the state to None so that we re-initialize at the next policy call.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/acme/adders/reverb/base.py\u001b[0m in \u001b[0;36madd_first\u001b[0;34m(self, timestep)\u001b[0m\n\u001b[1;32m    174\u001b[0m     self._writer.append(dict(observation=timestep.observation,\n\u001b[1;32m    175\u001b[0m                              start_of_episode=timestep.first()),\n\u001b[0;32m--> 176\u001b[0;31m                         partial_step=True)\n\u001b[0m\u001b[1;32m    177\u001b[0m     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_add_first_called\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    178\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/reverb/trajectory_writer.py\u001b[0m in \u001b[0;36mappend\u001b[0;34m(self, data, partial_step)\u001b[0m\n\u001b[1;32m    291\u001b[0m     \u001b[0;31m# Flatten the data and pass it to the C++ writer for column wise append.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    292\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mpartial_step\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 293\u001b[0;31m       \u001b[0mflat_column_data_references\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_writer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAppendPartial\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mflat_column_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    294\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    295\u001b[0m       \u001b[0mflat_column_data_references\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_writer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAppend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mflat_column_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mValueError\u001b[0m: Tensor of incompatible shape provided for column 0. Got [0] which is incompatible with [100]."
          ]
        }
      ],
      "source": [
        "# all_returns = []\n",
        "# # main loop\n",
        "# for episode in range(n_episodes):\n",
        "#   episode_steps = 0\n",
        "#   episode_return = 0\n",
        "#   episode_loss = 0\n",
        "\n",
        "#   start_time = time.time()\n",
        "\n",
        "#   timestep = env.reset()\n",
        "\n",
        "#   # Make the first observation.\n",
        "#   agent.observe_first(timestep)\n",
        "\n",
        "#   # Run an episode\n",
        "#   while not timestep.last():\n",
        "\n",
        "#     # DEBUG\n",
        "#     # print(timestep)\n",
        "\n",
        "#     # Generate an action from the agent's policy and step the environment.\n",
        "#     action = agent.select_action(timestep.observation)\n",
        "#     timestep = env.step(action)\n",
        "\n",
        "#     # Have the agent observe the timestep and let the agent update itself.\n",
        "#     agent.observe(action, next_timestep=timestep)\n",
        "#     agent.update()\n",
        "\n",
        "#     # Book-keeping.\n",
        "#     episode_steps += 1\n",
        "#     n_total_steps += 1\n",
        "#     episode_return += timestep.reward\n",
        "\n",
        "#     if log_loss:\n",
        "#       episode_loss += agent.last_loss\n",
        "\n",
        "#     if n_steps is not None and n_total_steps >= n_steps:\n",
        "#       break\n",
        "\n",
        "#   # Collect the results and combine with counts.\n",
        "#   steps_per_second = episode_steps / (time.time() - start_time)\n",
        "#   result = {\n",
        "#       'episode': episode,\n",
        "#       'episode_length': episode_steps,\n",
        "#       'episode_return': episode_return,\n",
        "#   }\n",
        "#   if log_loss:\n",
        "#     result['loss_avg'] = episode_loss/episode_steps\n",
        "\n",
        "#   all_returns.append(episode_return)\n",
        "\n",
        "#   display(env.plot_state())\n",
        "#   # Log the given results.\n",
        "#   print(result)\n",
        "\n",
        "#   if n_steps is not None and n_total_steps >= n_steps:\n",
        "#     break\n",
        "\n",
        "# clear_output()\n",
        "\n",
        "# # Histogram of all returns\n",
        "# plt.figure()\n",
        "# sns.histplot(all_returns, stat=\"density\", kde=True, bins=12)\n",
        "# plt.xlabel('Return [a.u.]')\n",
        "# plt.ylabel('Density')\n",
        "# plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "ZwxHabM5HA-x"
      },
      "source": [
        "**Note:** You can simplify the environment loop using [DeepMind Acme](https://github.com/deepmind/acme)."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### test DQN"
      ],
      "metadata": {
        "id": "-OF2iOiUOL35"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "execution": {},
        "id": "y_S1FJvRHA-y"
      },
      "outputs": [],
      "source": [
        "# # test dqn\n",
        "# # construct a DQN agent\n",
        "agent = dqn.DQN(\n",
        "    environment_spec=env_spec,\n",
        "    network=dqn_make_network(env_spec.actions),\n",
        "    epsilon=[0.1],\n",
        "    discount=0.01,\n",
        "    logger=loggers.InMemoryLogger(),\n",
        "    checkpoint=False,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "execution": {},
        "id": "9TTVcl5fHA-z"
      },
      "source": [
        "Now, we run the environment loop with the DQN agent and print the training log."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "execution": {},
        "id": "6a5rvXerHA-z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "outputId": "3993f31f-efcf-4370-943b-cfb183e25544"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'DQN reward by episode')"
            ]
          },
          "metadata": {},
          "execution_count": 25
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "# # training loop\n",
        "\n",
        "\n",
        "for discount in [0.1,0.5,0.9]:\n",
        "  agent = dqn.DQN(\n",
        "    environment_spec=env_spec,\n",
        "    network=dqn_make_network(env_spec.actions),\n",
        "    epsilon=[0.1],\n",
        "    discount=discount,\n",
        "    logger=loggers.InMemoryLogger(),\n",
        "    checkpoint=False)\n",
        "  loop = EnvironmentLoop(env, agent, logger=loggers.InMemoryLogger())\n",
        "  loop.run(n_episodes)\n",
        "  logs_dqn = pd.DataFrame(loop._logger._data)\n",
        "  episode_return=logs_dqn['episode_return']\n",
        "  plt.plot(moving_average(episode_return,5),label=\"discount=\"+str(discount))\n",
        "\n",
        "plt.legend()\n",
        "plt.ylabel(\"episode reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.title(\"DQN reward by episode\")\n",
        "# print logs\n",
        "# logs_dqn = pd.DataFrame(loop._logger._data)\n",
        "# logs_dqn.tail()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### test DRQN"
      ],
      "metadata": {
        "id": "lwP03QS4INDO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# # test DRQN\n",
        "# # construct a DRQN agent\n",
        "\n",
        "n_episodes = 500\n",
        "# epsilon default 0.05\n",
        "for learning_rate in [1e-3,1e-4]:\n",
        "  agent_r2d2 = r2d2.R2D2(\n",
        "      environment_spec=env_spec,\n",
        "      network=r2d2_make_network(env_spec.actions),\n",
        "      burn_in_length=10,\n",
        "      trace_length=2,\n",
        "      replay_period=10,\n",
        "      epsilon=0.05,\n",
        "      learning_rate=learning_rate,\n",
        "      min_replay_size=32,\n",
        "      logger=loggers.InMemoryLogger(),\n",
        "      checkpoint=False)\n",
        "\n",
        "  loop = EnvironmentLoop(env, agent_r2d2, logger=loggers.InMemoryLogger())\n",
        "  loop.run(n_episodes)\n",
        "\n",
        "  logs_r2d2 = pd.DataFrame(loop._logger._data)\n",
        "  episode_return=logs_r2d2['episode_return']\n",
        "  plt.plot(moving_average(episode_return,10),label=\"lr=\"+str(learning_rate))\n",
        "\n",
        "agent_random = RandomAgent(env_spec)\n",
        "loop = EnvironmentLoop(env, agent_random, logger=loggers.InMemoryLogger())\n",
        "loop.run(n_episodes)\n",
        "logs_random = pd.DataFrame(loop._logger._data)\n",
        "episode_return=logs_random['episode_return']\n",
        "plt.plot(moving_average(episode_return,10),label=\"baseline\")\n",
        "\n",
        "plt.legend()\n",
        "plt.ylabel(\"episode reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.title(\"DRQN reward by episode\")\n",
        "\n",
        "# print logs\n",
        "# logs_r2d2 = pd.DataFrame(loop._logger._data)\n",
        "# logs_r2d2.tail()"
      ],
      "metadata": {
        "id": "1YJ2JM_PHZvx",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 391
        },
        "outputId": "419db348-babd-490b-92a6-d67c3a65a6c4"
      },
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /root/acme/27a4cebc-09a6-11ed-8643-0242ac1c0002/snapshots/network/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /root/acme/27a4cebc-09a6-11ed-8643-0242ac1c0002/snapshots/network/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /root/acme/27a4cebc-09a6-11ed-8643-0242ac1c0002/snapshots/network/assets\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "INFO:tensorflow:Assets written to: /root/acme/27a4cebc-09a6-11ed-8643-0242ac1c0002/snapshots/network/assets\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'DRQN reward by episode')"
            ]
          },
          "metadata": {},
          "execution_count": 121
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# # test DRQN\n",
        "# # construct a DRQN agent\n",
        "\n",
        "n_episodes = 200\n",
        "# epsilon default 0.05\n",
        "for discount in [0.1,0.5,0.9]:\n",
        "  agent_r2d2 = r2d2.R2D2(\n",
        "      environment_spec=env_spec,\n",
        "      network=r2d2_make_network(env_spec.actions),\n",
        "      burn_in_length=10,\n",
        "      trace_length=2,\n",
        "      replay_period=10,\n",
        "      epsilon=0.05,\n",
        "      discount=discount,\n",
        "      learning_rate=1e-4,\n",
        "      min_replay_size=32,\n",
        "      logger=loggers.InMemoryLogger(),\n",
        "      checkpoint=False)\n",
        "\n",
        "  loop = EnvironmentLoop(env, agent_r2d2, logger=loggers.InMemoryLogger())\n",
        "  loop.run(n_episodes)\n",
        "\n",
        "  logs_r2d2 = pd.DataFrame(loop._logger._data)\n",
        "  episode_return=logs_r2d2['episode_return']\n",
        "  plt.plot(moving_average(episode_return,5),label=\"discount=\"+str(discount))\n",
        "\n",
        "agent_random = RandomAgent(env_spec)\n",
        "loop = EnvironmentLoop(env, agent_random, logger=loggers.InMemoryLogger())\n",
        "loop.run(n_episodes)\n",
        "logs_random = pd.DataFrame(loop._logger._data)\n",
        "episode_return=logs_random['episode_return']\n",
        "plt.plot(moving_average(episode_return,10),label=\"baseline\")\n",
        "\n",
        "plt.legend()\n",
        "plt.ylabel(\"episode reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.title(\"DRQN reward by episode\")\n",
        "\n",
        "# print logs\n",
        "# logs_r2d2 = pd.DataFrame(loop._logger._data)\n",
        "# logs_r2d2.tail()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 373
        },
        "id": "tnjpQCY6ssNG",
        "outputId": "a9b1c77b-fed8-4852-83ba-4f9408129b65"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "INFO:tensorflow:Assets written to: /root/acme/519f3846-0975-11ed-83b6-0242ac1c0002/snapshots/network/assets\n",
            "INFO:tensorflow:Assets written to: /root/acme/519f3846-0975-11ed-83b6-0242ac1c0002/snapshots/network/assets\n",
            "INFO:tensorflow:Assets written to: /root/acme/519f3846-0975-11ed-83b6-0242ac1c0002/snapshots/network/assets\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'DRQN reward by episode')"
            ]
          },
          "metadata": {},
          "execution_count": 14
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Playground"
      ],
      "metadata": {
        "id": "l1hhXaIIfvzQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## test different parameters"
      ],
      "metadata": {
        "id": "UAVvoTrU-jRU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# test 5 back\n",
        "\n",
        "N = 2\n",
        "n_episodes = 100\n",
        "episode_steps= 32\n",
        "n_total_steps = 0\n",
        "log_loss = False\n",
        "n_steps = n_episodes * episode_steps\n",
        "# mock data\n",
        "# mock_nback_data = generate_mock_nback_dataset(N=N,n_trials=episode_steps)\n",
        "# environment\n",
        "env, env_spec = NBack.create_environment(N=N,episode_steps=episode_steps)\n",
        "obs = action_norm.ActionNormObserver()\n",
        "\n",
        "# agent\n",
        "agent_random = RandomAgent(env_spec)\n",
        "\n",
        "agent_dqn = dqn.DQN(\n",
        "        environment_spec=env_spec,\n",
        "        network=dqn_make_network(env_spec.actions),\n",
        "        epsilon=[0.1],\n",
        "        logger=loggers.InMemoryLogger(),\n",
        "        checkpoint=False,\n",
        ")\n",
        "\n",
        "agent_drqn = r2d2.R2D2(\n",
        "    environment_spec=env_spec,\n",
        "    network=r2d2_make_network(env_spec.actions),\n",
        "    burn_in_length=10,\n",
        "    trace_length=2,\n",
        "    replay_period=10,\n",
        "    epsilon=[0.1],\n",
        "    logger=loggers.InMemoryLogger(),\n",
        "    checkpoint=False,\n",
        ")\n",
        "\n",
        "# training loop\n",
        "agent_dict= {'RANDOM':agent_random,\n",
        "    'DQN':agent_dqn,\n",
        "    'DRQN':agent_drqn}\n",
        "logs_dict = {}\n",
        "\n",
        "for key, agent in tqdm(agent_dict.items()):\n",
        "    loop = EnvironmentLoop(env, agent, \n",
        "                logger=loggers.InMemoryLogger())\n",
        "    loop.run(n_episodes)\n",
        "    logs_dict[key] = pd.DataFrame(loop._logger._data)\n",
        "    # print(logs_dict[key].tail(5))\n",
        "\n"
      ],
      "metadata": {
        "id": "_9rMoNVqUPOQ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 413
        },
        "outputId": "3327e846-0aa0-49d4-a895-5f7d42d93b77"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "  0%|          | 0/3 [00:00<?, ?it/s]\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "TypeError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-44-e1a7ba68ebe0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     45\u001b[0m                 \u001b[0mlogger\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mloggers\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mInMemoryLogger\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     46\u001b[0m                 observers=obs)\n\u001b[0;32m---> 47\u001b[0;31m     \u001b[0mloop\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn_episodes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     48\u001b[0m     \u001b[0mlogs_dict\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDataFrame\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloop\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_logger\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m     \u001b[0;31m# print(logs_dict[key].tail(5))\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/acme/environment_loop.py\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, num_episodes, num_steps)\u001b[0m\n\u001b[1;32m    174\u001b[0m     \u001b[0;32mwith\u001b[0m \u001b[0msignals\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mruntime_terminator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    175\u001b[0m       \u001b[0;32mwhile\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mshould_terminate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mepisode_count\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstep_count\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 176\u001b[0;31m         \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_episode\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    177\u001b[0m         \u001b[0mepisode_count\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    178\u001b[0m         \u001b[0mstep_count\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'episode_length'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/acme/environment_loop.py\u001b[0m in \u001b[0;36mrun_episode\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     95\u001b[0m     \u001b[0;31m# Make the first observation.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_actor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobserve_first\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtimestep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 97\u001b[0;31m     \u001b[0;32mfor\u001b[0m \u001b[0mobserver\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_observers\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     98\u001b[0m       \u001b[0;31m# Initialize the observer with the current state of the env after reset\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     99\u001b[0m       \u001b[0;31m# and the initial timestep.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mTypeError\u001b[0m: 'ActionNormObserver' object is not iterable"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_reward_by_episode(logs_dict)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "zAS4HBUtHY_Z",
        "outputId": "a2b36fa7-5260-4896-c02c-6af59af1f76b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "env.plot_state()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "id": "mPbAK3JmrEwU",
        "outputId": "aab7c12d-6d7b-4867-8fc0-34b49698bfd7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ],
            "text/html": [
              "<b>Environment (2-back):</b><br /><pre><b>Stimuli:</b> 5430303033224244505035353333333</pre><pre><b>Actions:</b> MM.MMMMMMMMMMMMMMMMMMMMMMMMMMMM</pre>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# N = 2\n",
        "# n_episodes = 1000\n",
        "# episode_steps= 100\n",
        "for key, val in logs_dict.items():\n",
        "    logs=logs_dict[key]\n",
        "    reward_ma=moving_average(logs['episode_return'],10)\n",
        "    plt.plot(reward_ma)\n",
        "\n",
        "plt.ylabel(\"cumulative reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.legend(['RANDOM','DQN','DRQN'])\n",
        "# because episode_steps too large, \n",
        "# DQN learn the best policy within the first episode \n",
        "# so reward is not increasing by episode"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 303
        },
        "id": "PoexUzo1ybkM",
        "outputId": "6916d3a9-d36e-44c6-ada8-72bbd83bb85d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fab611273d0>"
            ]
          },
          "metadata": {},
          "execution_count": 24
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# N = 2\n",
        "# n_episodes = 1000\n",
        "# episode_steps= 32\n",
        "for key, val in logs_dict.items():\n",
        "    logs=logs_dict[key]\n",
        "    reward_ma=moving_average(logs['episode_return'],10)\n",
        "    plt.plot(reward_ma)\n",
        "\n",
        "plt.ylabel(\"cumulative reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.legend(['RANDOM','DQN','DRQN'])\n",
        "# episode_steps is small and dqn is learning by episode"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 303
        },
        "id": "7lg4y9-loCS7",
        "outputId": "58af0a92-0002-4943-c476-07d1fa5d0162"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fab60eebb10>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# N = 5\n",
        "# n_episodes = 1000\n",
        "# episode_steps= 32\n",
        "\n",
        "for key, val in logs_dict.items():\n",
        "    logs=logs_dict[key]\n",
        "    reward_ma=moving_average(logs['episode_return'],10)\n",
        "    plt.plot(reward_ma)\n",
        "\n",
        "plt.ylabel(\"cumulative reward\")\n",
        "plt.xlabel(\"episode\")\n",
        "plt.legend(['RANDOM','DQN','DRQN'])\n",
        "plt.title(\"5-back\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 306
        },
        "id": "Rv5Hz3emoC4C",
        "outputId": "9afb1f43-d842-43f4-d364-0a32efe38639"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f66949bc190>"
            ]
          },
          "metadata": {},
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## error pattern"
      ],
      "metadata": {
        "id": "kQodUSba-nRE"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### episode analysis"
      ],
      "metadata": {
        "id": "Y3SxB2-efW-L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "logs_drqn=logs_dict['DRQN']\n",
        "# dataframe\n",
        "print(logs_dqn.tail(10))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "99A5IpgraSm8",
        "outputId": "f2719539-128e-4d37-ed3c-f4eb299af749"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "     episode_length episode_return  steps_per_second  episodes  steps\n",
            "990              32           -4.0        346.219121       991  31712\n",
            "991              32           -8.0        336.968010       992  31744\n",
            "992              32           -5.0        352.724656       993  31776\n",
            "993              32           -7.0        350.254117       994  31808\n",
            "994              32           -6.0        321.221841       995  31840\n",
            "995              32          -10.0        347.744952       996  31872\n",
            "996              32           -4.0        347.761170       997  31904\n",
            "997              32           -8.0        346.695790       998  31936\n",
            "998              32           -5.0        328.387926       999  31968\n",
            "999              32           -9.0        321.985510      1000  32000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### trial analysis"
      ],
      "metadata": {
        "id": "UXgiy9LQywZd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "df=pd.read_csv('DRQN_2back_500episodes_32steps.csv')\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "UA11kQOuuLmD",
        "outputId": "3dfabc7f-87ab-4333-fd34-f80ba1936a00"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   trial  episode  stimuli     action  reward\n",
              "0      0      1.0      1.0  non-match    -1.0\n",
              "1      1      1.0      5.0      match    -1.0\n",
              "2      2      1.0      0.0  non-match     0.0\n",
              "3      3      1.0      2.0      match    -1.0\n",
              "4      4      1.0      0.0      match     0.0"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-c4d8ff2f-37b7-42a1-8b9c-56f3b37bfb68\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>trial</th>\n",
              "      <th>episode</th>\n",
              "      <th>stimuli</th>\n",
              "      <th>action</th>\n",
              "      <th>reward</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>non-match</td>\n",
              "      <td>-1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>match</td>\n",
              "      <td>-1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>non-match</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>match</td>\n",
              "      <td>-1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>match</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c4d8ff2f-37b7-42a1-8b9c-56f3b37bfb68')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-c4d8ff2f-37b7-42a1-8b9c-56f3b37bfb68 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-c4d8ff2f-37b7-42a1-8b9c-56f3b37bfb68');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(np.cumsum(df['reward']))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "lINnv_REM7jc",
        "outputId": "c3a3f5e2-05c2-4fb8-e09f-5233ec1b3c6e"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f1eba57b650>]"
            ]
          },
          "metadata": {},
          "execution_count": 34
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_episode = df.groupby([\"episode\"]).sum()"
      ],
      "metadata": {
        "id": "fB9q-LdLfYnq"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(df_episode['reward'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        },
        "id": "FbwZatxeesFe",
        "outputId": "ce67bd58-4821-4664-fdbf-e689b83ba601"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f1ef65cb310>]"
            ]
          },
          "metadata": {},
          "execution_count": 38
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot error idx\n",
        "fig, (ax1, ax2) = plt.subplots(1,2,figsize=(20,5))\n",
        "\n",
        "error_idx=df['trial'][df['reward']==-1][df['episode']<310]\n",
        "ax1.hist(error_idx,bins=episode_steps,density=True)\n",
        "ax1.set_ylim(0,0.05)\n",
        "ax1.set_ylabel('density')\n",
        "ax1.set_xlabel('step index')\n",
        "\n",
        "\n",
        "error_idx=df['trial'][df['reward']==-1][df['episode']>320]\n",
        "ax2.hist(error_idx,bins=episode_steps,density=True)\n",
        "ax2.set_ylim(0,0.05)\n",
        "ax2.set_xlabel('step index')\n",
        "\n",
        "\n",
        "fig.suptitle('error step index distribution before and after jump')"
      ],
      "metadata": {
        "id": "kyV0SSV-uPcf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        },
        "outputId": "325a3786-dfbf-4a06-ed43-95506aac8f44"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0.98, 'error step index distribution before and after jump')"
            ]
          },
          "metadata": {},
          "execution_count": 42
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_best= df[df['episode']>400]"
      ],
      "metadata": {
        "id": "yq0k5g3Ig4oW"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "error_sti=df_best['stimuli'][df['reward']==-1]\n",
        "plt.hist(error_sti)\n",
        "# ax2.set_ylim(0,0.05)\n",
        "plt.xlabel('stimulus')\n",
        "plt.title('error distribution of best policy')"
      ],
      "metadata": {
        "id": "HxEB7e8GmYJ6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "outputId": "9f2ff7b3-8157-4b74-911a-7f55a95c03bd"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'error distribution of best policy')"
            ]
          },
          "metadata": {},
          "execution_count": 50
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df['action binary']=[0 if x=='match' else -1 for x in df['action']]\n",
        "df['expected binary']= df['action binary']==df['reward']\n",
        "df['expected action']=['match' if x==True else 'non-match' for x in df['expected binary']]\n",
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        },
        "id": "_kI0X8q5hXbR",
        "outputId": "18ff2630-b1bc-4792-fd62-182663686173"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "   trial  episode  stimuli     action  reward  action binary  expected bianry  \\\n",
              "0      0      1.0      1.0  non-match    -1.0             -1             True   \n",
              "1      1      1.0      5.0      match    -1.0              0            False   \n",
              "2      2      1.0      0.0  non-match     0.0             -1            False   \n",
              "3      3      1.0      2.0      match    -1.0              0            False   \n",
              "4      4      1.0      0.0      match     0.0              0             True   \n",
              "\n",
              "   expected binary expected action  \n",
              "0             True           match  \n",
              "1            False       non-match  \n",
              "2            False       non-match  \n",
              "3            False       non-match  \n",
              "4             True           match  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-5bf77f97-397a-46ca-86df-9017ffa3130b\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>trial</th>\n",
              "      <th>episode</th>\n",
              "      <th>stimuli</th>\n",
              "      <th>action</th>\n",
              "      <th>reward</th>\n",
              "      <th>action binary</th>\n",
              "      <th>expected bianry</th>\n",
              "      <th>expected binary</th>\n",
              "      <th>expected action</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>non-match</td>\n",
              "      <td>-1.0</td>\n",
              "      <td>-1</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>match</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>1</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>match</td>\n",
              "      <td>-1.0</td>\n",
              "      <td>0</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>non-match</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>2</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>non-match</td>\n",
              "      <td>0.0</td>\n",
              "      <td>-1</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>non-match</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>3</td>\n",
              "      <td>1.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>match</td>\n",
              "      <td>-1.0</td>\n",
              "      <td>0</td>\n",
              "      <td>False</td>\n",
              "      <td>False</td>\n",
              "      <td>non-match</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>4</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>match</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>True</td>\n",
              "      <td>True</td>\n",
              "      <td>match</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-5bf77f97-397a-46ca-86df-9017ffa3130b')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-5bf77f97-397a-46ca-86df-9017ffa3130b button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-5bf77f97-397a-46ca-86df-9017ffa3130b');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fig, (ax1, ax2) = plt.subplots(1,2,figsize=(12,6))\n",
        "\n",
        "error_action=df['expected action'][df['reward']==-1][df['episode']>400]\n",
        "ax1.hist(error_action,density=True)\n",
        "ax1.set_xlabel('expected action')\n",
        "ax1.set_title('wrong action distribution over expected action of best policy')\n",
        "\n",
        "exp_action=df['expected action'][df['episode']>400]\n",
        "ax2.hist(exp_action,density=True)\n",
        "ax2.set_xlabel('expected action')\n",
        "ax2.set_title('expected action distribution of best policy')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 428
        },
        "id": "qmgerjfYli8v",
        "outputId": "70f267a3-0f4d-4440-cacc-f632a85a2aaa"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'expected action distribution of best policy')"
            ]
          },
          "metadata": {},
          "execution_count": 73
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x432 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs4AAAGJCAYAAACaWbVlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxN1+L//3eGRkm0hgYprttqjS0yKzHPRExV9KpS6pNbar54VE1pKapVU6tapaNWB1rK1fZSOlKtoZQiNSaRRNAmIdPJ+v3h53wTMuxEdhL6ev7lOOfstfbOXmu/zzprr+NijDECAAAAkCfXkq4AAAAAcCMgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcE5i+joaPn6+srhcNhe1iOPPKIPP/xQkvTZZ5/pscceK7Jtd+vWTTt27JAkLV68WBMmTCiybS9btkxTpkwpsu2h6E2ePFkLFiywvZxp06Zp6dKltpdztffee0/NmjWTr6+vzp8/n+2506dPq27dusrIyCj2ehXWjh071LJlS+fjrO0Xfw9Zrwd2GjZsmNauXWt7OZ988okGDBjgfOzr66tTp04VybazXoOKur0XZwbI6uzZs/rXv/4lX19fzZkz55rni6tPL0p2ZpyS9rcOzm3bttX333/vfHznnXdq9+7dcnNzK9Z6hIWF6Y033sj3dVYbz+eff67g4ODrrtfVF3RJCg8P16xZs65728hd3bp1deLEiZKuRjZXXwglKSIiQiNGjCjWeqSnp2vOnDl64403tHv3blWsWLFYyi2uYCMVXftF8SitH9ZyGjR5/fXX1atXr2Kvy+7du1WzZs08X5PT9SYnRXkNKi0Z4IMPPlDFihX1yy+/aPLkycVSptXjXRSsZpwbhS3BubR1IH8XHPf8FfdIAopWQkKCUlNTdc8995R0VQAUs5v1GhcdHa3atWvLxcWlpKsCC/IMzh9//LHCw8Odjzt27KhRo0Y5H7dq1UoHDx6UdHmU7N1331XHjh3VsWNHSdKaNWvUoUMHBQUFKTw8XLGxsc731q1bV6tXr1bHjh0VEBCgmTNn6sqPGDocDs2ZM0fBwcFq27at3nnnnTw/0S9fvlzt27eXr6+vunbtqi+//DLb82vWrFGXLl2czx84cED/+c9/FB0drfDwcPn6+uq11167ZuQgNjZW4eHhCgoKUocOHbRmzRrnNhcvXqzRo0dr4sSJ8vX1Vbdu3fTrr7/meiy/++47de7cWf7+/oqIiFDWH2zMOppnjNHs2bP1wAMPyM/PT927d9fhw4f1wQcfaP369VqxYoV8fX2df5e2bdtq+fLl6t69u5o0aaKMjIxrPkWnpaVpzJgx8vX1Va9evXTo0KFsf4eso5tXRrUvXryoxx9/XHFxcfL19ZWvr69iY2OvGcX43//+p27duikgIECPPPKIIiMjnc+1bdtWK1asUPfu3eXv768xY8YoNTU1x+OTmZmpl19+WW3atNEDDzygiRMnKjExUdLlrxffeeedbK8PCwvTF198IUmKjIzUkCFDFBQUpE6dOmnjxo3Z9mf69Ol6/PHH1aRJkxy/Ak9MTNRTTz2lkJAQtWjRQgsWLJDD4VBaWpp69Oiht99+W9Ll87J///5asmSJpMvnwKhRo3I9trGxsXryySfVtGlTtW3bVm+99ZbzOYfDoWXLljnP2969eysmJkb/+te/JEk9evSQr6+vc1+2bt2qHj16KCAgQP37989Wzm+//aZevXrJ19c3z2MsSSdPntSgQYMUHBys4OBgjR8/Xn/99Zfz+ZiYGI0cOVJNmzZVcHCwIiIiFBkZqenTp2vPnj3y9fVVQECA89hm/QaksO39amlpaZo1a5ZCQkIUEhKiWbNmKS0tTceOHVPnzp0lSYGBgRo0aFCu+/nxxx87379ixQrn/2dmZjr7i+DgYI0ePVoXLlyQJKWmpmrChAkKDg5WQECA+vTpo7Nnz2rBggXatWuXIiIi5Ovrq4iIiGvKu9J3fPDBBzmWm9s+5SRr+83tPJk5c+Y1X+mGh4dr1apVuR6Tm0lubevChQtq2bKltmzZIklKTk5Whw4dtG7dOkmXz9lp06ZpyJAh8vX11cCBAxUVFeXcbl59SUpKiubMmaM2bdrI399fAwYMUEpKigYOHCjp8jnp6+ur3bt3S5I++ugjdenSRYGBgRo6dGi2cvK6Hlxt37596tevnwICAhQSEqKIiIhs586RI0ecdW7WrJmWLVum7du369VXX9WmTZvk6+ursLAwSdm/Ocmrz71yPq9du1atW7dWcHCwXnnllVzreP78eYWHh8vPz08PPvigTp48me35rNeZbdu2qWvXrvL19VWLFi20YsWKPK83o0aN0oQJE+Tn56e1a9fmOJKeW3u/uo/KOspa3Bngl19+UZ8+feTv768+ffrol19+cdZx3bp1zmt71mv31ce4MOdtQY731fJrL7nt09Wu/sYyp3M2Pj5ejRs3zjb97sCBA2ratKnS09NzPa4lwuTh5MmTxt/f3zgcDnPmzBnTunVr06JFC+dzAQEBxuFwGGOMqVOnjhk8eLA5f/68uXTpkvn+++9NUFCQ2b9/v0lNTTURERHm4Ycfdm67Tp06Zvjw4ebPP/80UVFRJjg42Gzbts0YY8x7771nunTpYmJiYsyFCxfMo48+aurUqWPS09NzrOfGjRvNmTNnjMPhMJ9//rlp3LixiY2NdT4XEhJi9u7dazIzM83x48fN6dOnjTHGtGnTxnz33XfO7Zw6dSpbOQ8//LCZPn26SUlJMb/99psJDg4233//vTHGmEWLFpn77rvPfP311yYjI8PMnz/f9O3bN8f6JSQkmCZNmphNmzaZtLQ0s3LlSlO/fn2zZs0aY4wxH3/8senfv78xxpjt27ebXr16mT///NNkZmaao0ePOvdl0qRJ5sUXX8y27TZt2piwsDATHR1tLl26dM1+LVq0yDRo0MBZ9uuvv27atGlj0tLSnH+H48ePO7eXtYwff/zR+fe+YtGiRWb8+PHGGGP++OMP07hxY/Ptt9+atLQ0s3z5ctO+fXuTmprqrEefPn3MmTNnzPnz503nzp3Ne++9l+Mx+vDDD0379u3NyZMnTVJSkhkxYoSZMGGCMcaYtWvXmn79+jlfe+TIEePv729SU1NNcnKyadmypfnoo49Menq6OXDggAkKCjJHjhxx7o+fn5/ZtWuXcTgcJiUl5Zqyn3jiCTN16lSTnJxszp49a/r06WNWr15tjDHm999/NwEBAebo0aPm5ZdfNn379jUZGRn5HluHw2F69eplFi9ebFJTU83JkydN27Ztzfbt240xxrz22msmNDTUREZGmszMTHPw4EFz7ty5HP8mBw4cME2bNjV79uwxGRkZ5pNPPjFt2rQxqampJjU11bRu3dqsXLnSpKWlmU2bNpkGDRpcc55ccfz4cfPtt9+a1NRUk5CQYB5++GHz7LPPGmOMycjIMN27dzezZs0yycnJJiUlxfz000/GmOznaE7nyvW096u99NJLpm/fvubs2bMmISHB9OvXzyxYsMAYc20bvdqV58eOHWuSk5PNoUOHTHBwsLM9rFq1yvTt29fExMSY1NRUM3XqVDN27FhjjDGrV682//d//2cuXrxoMjIyzK+//moSExONMcYMHDjQ2V4LU25e+3R1O8vafnM7T/bu3WuaN2/u7H8TEhJMo0aNTHx8fK51vFnk17a++eYb06xZM3P27FkzZcoU8+STTzrfO2nSJNOkSROzc+dOk5qaap555hnneZ1fXzJjxgwzcOBAc+bMGZORkWF+/vlnk5qamuM5+eWXX5r27dubo0ePmvT0dLN06VJnH5bf9eBqv/76q9m9e7dJT083p06dMp07dzYrV640xhiTmJhomjdvblasWGFSUlJMYmKi2bNnjzEme199RdbzOK8+98o+TZkyxVy6dMkcPHjQNGzY0Bw9ejTHOo4ZM8aMGjXKJCcnm99//92EhIRk6y+y9mnNmzd39isXLlww+/fvN8bkfr1p0KCB+fLLL43D4TCXLl3Ktl/5tburr5l5tbWs2yvqDHD+/HkTEBBg1q5da9LT08369etNQECAs8/P6dqe1fWctwU53gUpN799ynquZb1+5HXODhs2zLz77rvO8mfNmmUiIiLyrGNJyHPEuWbNmvL09NTBgwe1a9cuhYSEqEqVKoqMjNTOnTvl7+8vV9f/t4nhw4erQoUKuvXWW7V+/Xr16dNHDRs2lIeHh8aNG6c9e/bo9OnTztc//vjjuu2223TnnXcqODjYOYq2adMmDRo0SNWqVdPtt9+u4cOH5xn+u3TpoqpVq8rV1VVdu3ZVrVq1tG/fPkmXP/UPGzZMjRo1kouLi2rVqqXq1avn+4EiJiZGv/zyiyZMmKAyZcqofv366tu3rz799FPna/z9/dWqVSu5ubmpR48e2UYBs9q+fbvuvfdede7cWbfccoseffRR3XHHHTm+1t3dXcnJyfrjjz9kjFHt2rVVpUqVPOv6yCOPyMfHR7feemuOzzds2NBZ9pAhQ5SWlqa9e/fmewzys3HjRrVq1UrNmzfXLbfcoqFDhyolJcU54nKlblWrVlWFChXUpk0b5zcUV1u/fr0GDx7sPOfGjRunjRs3KiMjQ+3bt9ehQ4ecn3TXr1+vDh06yMPDQ19//bWqV6+uPn36yN3dXQ0aNFCnTp303//+17ntdu3aOc/VMmXKZCv37Nmz2rZtm5566imVK1dOlStX1uDBg/X5559LkurUqaN///vfeuKJJ/TGG29o3rx52ea/5XZsf/31V507d04jR46Uh4eHatasqYceesg5EvDhhx9q9OjRuvvuu+Xi4qJ69erlOl/3gw8+UL9+/dS4cWO5ubmpV69euuWWW7Rnzx7t3btX6enpevTRR3XLLbeoc+fOuv/++3P9m9WqVUvNmzeXh4eHKlWqpCFDhuinn36SdHlkKy4uThMnTlS5cuVUpkwZ5+hyfq6nvee0rREjRqhy5cqqVKmSRowYoc8++8xSPa4YMWKEypUrp7p166p3797asGGDJOn999/X2LFjVa1aNXl4eGjkyJHavHmzMjIy5O7urgsXLujEiRNyc3PTfffdJy8vryIpt7D7lNt50qhRI5UvX14//PCDpMttMSgoKNd+5WaSX9sKCQlR586dNXjwYG3btk0zZ87M9v7WrVsrMDBQHh4eGjt2rPbs2aOYmJg8+5LMzEx9/PHHmjJliqpWrSo3Nzf5+fnJw8Mjxzq+//77Gj58uGrXri13d3eFh4fr4MGDioqKKtD1QJLuu+8+NWnSRO7u7qpRo4b69evnbLNff/217rjjDj322GMqU6aMvLy81LhxY0vHMa8+94qRI0fq1ltvVb169VSvXr0c26zD4dAXX3yhUaNGqVy5cqpTp06e86jd3d119OhRJSUl6fbbb1fDhg3zrGeTJk3Uvn17ubq65nqNy63dXY+izABff/21atWqpZ49e8rd3V2hoaG6++67tXXrVsv1Kcx5KxX8eBek3MLsU17nbK9evZz9osPh0Oeff64ePXoUqL7FwT2/FwQGBmrnzp06ceKEAgMDVb58ef3000/as2ePgoKCsr3Wx8fH+e+4uLhsfyBPT09VqFBBsbGxqlGjhiTJ29vb+XzZsmWVnJzsfG/WbVWrVi3POq5bt04rV650BquLFy86h/tjYmL0j3/8I7/dvEZcXJxuv/32bBfOO++8U/v373c+ztrZ3XrrrUpNTXVegK/eVtZ9cHFxybZ/WT3wwAP617/+pYiICEVFRaljx46aNGlSnhfw3LZ1RdayXV1dVbVqVcXFxeX5Hivi4uJ05513Ztu2j49Ptq98rv4b51ZuXFxctg801atXV0ZGhhISElS1alW1atVKn3/+uYYPH64NGzbo2WeflSRFRUVp37592QKew+FwfjUp5X18oqOjlZGRoZCQEOf/ZWZmZntPz549tWDBAnXs2FH//Oc/s70/r2MbFxd3Tb2uPD5z5ozl8zI6Olrr1q3LNl0lPT1dcXFxcnFxUdWqVbPNjcv6N7na2bNnNWvWLO3atUvJyckyxui2226TdLmt3Hnnndecv1ZcT3vPaVtZ9+HOO+8s8Pma9e9XvXp1HT58WNLlYzlixIhsH/hdXV2VkJCgHj166MyZMxo3bpz++usvhYWFaezYsbrllluuu9zC7lNe58mVi0zz5s312Wef5Tl15WYSFRWVZ9uSpIceekjvvPOOwsPDr/lAmrXNenp66vbbb1dcXFyefcn58+eVmpqa7w1uV0RHR2v27NmaO3eu8/+MMYqNjS3Q9UCSjh07pjlz5mj//v26dOmSHA6Hs60V9vom5d3nXpH1Gle2bFldvHjxmu2cO3dOGRkZ2fYhrz5o0aJFeuWVV/TCCy+obt26Gj9+vHx9fXN9fX7Xfyn3dnc9ijoDXH1M7rzzzhynR+SmMOetVPDjbbXcwu5TXudsu3btNH36dJ06dUrHjh2Tl5eXGjVqZLmuxSXfK2RQUJC2bNmiqKgohYeH67bbbtP69eu1e/du53zMK7JevKtUqZJtLszFixd14cIFVa1aNd9KeXt768yZM87HWf99taioKD399NNatWqVfH19nZ/8rvDx8blmvpUVVapU0Z9//qmkpCRnw4mJibFU/6tdvT/GGMXExOT6+kGDBmnQoEFKSEjQmDFj9Prrr2vMmDG53jiQ3w0FWcvOzMxUbGyscxS7bNmyunTpkvP5+Ph45z7mt90qVapk66Cu7FdhjtHV50t0dLTc3d1VuXJlSVJoaKiWLFmiwMBApaamOlcd8PHxUWBgoFauXFngMiU5Rx5//PHHXAPjzJkz1aZNG3377bfatWtXtg4qt2Pr5uamGjVqOOdh51TuyZMnVadOnXzr6OPjo/DwcP373/++5rmdO3cqNjZWxhjn3ys6OjrXC/yLL74oFxcXrV+/XhUqVNBXX33lnLPr4+OjmJiYHDt+K+dCYdt7TtuKjo7WvffeK+lyu8vvW5erxcTEqHbt2pIuH48r769WrZpmz54tf3//HN83cuRIjRw5UqdPn9bw4cN11113qW/fvtddbmH3Ka/zJCwsTKGhoTp06JAiIyPVvn17y/W8kfn4+OTZthwOh6ZNm6aePXvqvffeU+/evVWrVi3n81nbbHJysv78809VqVIlz74kMzNTZcqU0alTp1SvXr1sz+XUNq602awf4K84ceJEga4HM2bMUIMGDfTCCy/Iy8tLq1at0ubNm53lZJ3Pml+9ssqrz83rmnu1SpUqyd3dPdu5n9f+NGrUSK+88orS09P17rvvasyYMdq2bVuhr29Xysup3ZUtW1YpKSnO1509e9byfhVlBrjS/q+uc4sWLSxvozDnrVTw42213MLuU17nbJkyZdSlSxd99tln+uOPP0rlaLNkYVWNwMBA7dixQykpKapWrZoCAgL0zTff6MKFC2rQoEGu7wsNDdUnn3yigwcPKi0tTS+++KIaNWrkHH3KS5cuXfTWW28pNjZWf/31l1577bVcX3vp0iW5uLioUqVKki7fJHDkyBHn8w8++KDeeOMN7d+/X8YYnThxwtlZ3HHHHbmuLenj4yNfX1+9+OKLSk1N1aFDh/TRRx/l2BHmp1WrVjpy5Ii++OILZWRk6K233sq1Ae/bt8/59XvZsmXl4eHhHB2rXLlytq++rTpw4ICz7DfffFMeHh7Or0bq1aunDRs2yOFwaPv27c6vAK+Ud+HCBecNI1fr0qWLtm3bph9++EHp6el644035OHhUaBPs1eEhobqzTff1KlTp5ScnKwFCxaoS5cuzgDXqlUrRUdHa9GiReratavzmLRu3VrHjx/XunXrlJ6ervT0dO3bty/bTYp5qVKlipo3b645c+YoKSlJmZmZOnnypHbu3Cnp8rcZBw4c0HPPPaenn35akydPzjZSmtuxbdSokTw9PbV8+XKlpKTI4XDo8OHDzilEffv21cKFC3X8+HEZY3To0CHntyRXn5d9+/bV+++/r71798oYo4sXL+rrr79WUlKS8yvct956S+np6friiy/yvEElOTlZ5cqVU/ny5RUbG6vXX3/d+VyjRo3k7e2tF154QRcvXlRqaqp+/vlnSZfPhdjY2FxvaLue9n61bt266ZVXXtG5c+d07tw5LV26VN27dy/QNl5++WVdunRJR44c0SeffKKuXbtKkgYMGKCXXnrJ2QecO3dOX331lSTpxx9/1O+//y6HwyEvLy+5u7s7z7O8+gor5RZ2n/I6T6pVq6b7779f//nPf9SxY8dcv8a+2eTXtpYtWyYXFxfNnj1bQ4cO1aRJk7KtprNt2zbt2rVLaWlpWrhwoRo3biwfH588+xJXV1f16dNHzz33nGJjY+VwOLR7926lpaWpUqVKcnV1zXZ+9O/fX8uXL3deixITE7Vp0yZJBbseSJfbrKenpzw9PRUZGanVq1c7n2vdurXi4+O1atUqpaWlKSkpyTkNr3LlyoqKilJmZmaO282vz7XKzc1NHTp00JIlS3Tp0iUdPXo017Wi09LS9NlnnykxMVG33HKLPD09s13f8rre5CW3dle/fn1t27ZNFy5cUHx8vN58881s7yvODHD8+HGtX79eGRkZ2rhxo44eParWrVtb3kZhztuiON65lVvYfcrrnJUu3xi/du1abdmy5cYNznfddZc8PT2do2xeXl6qUaOG/Pz88lzrsFmzZho9erSefPJJhYSE6NSpU5YX8H7ooYfUvHlzhYWFqWfPnmrVqpXc3d1zLO+ee+7RY489pv79+6tZs2Y6fPiw/Pz8nM936dJF4eHhGj9+vPz8/DRixAj9+eefki7PyX7llVcUEBCQ7U7cK1588UVFRUWpRYsWGjlypJ588kk1a9bM0j5kValSJS1cuFAvvPCCgoODdeLEiWx1zCo5OVlPP/20goKC1KZNG1WoUEFDhw6VdPlDwNGjRxUQEKAnnnjCcvnt2rXTxo0bFRgYqE8//VSLFy92fv08ZcoUbd26VQEBAVq/fn22UavatWurW7duat++vQICAq75Cubuu+/W888/r2eeeUZNmzbV1q1btWzZslzn/eWlT58+CgsL08CBA9WuXTt5eHho6tSpzuc9PDzUoUMHff/99woNDXX+v5eXl1asWKGNGzeqRYsWCgkJ0fz583MNeDmZN2+e0tPT1bVrVwUGBmrUqFGKj49XdHS0nnvuOc2dO1eenp7q3r277rvvPj333HP5Hls3NzctW7ZMhw4dUrt27dS0aVM9/fTTSkpKkiQNGTJEXbp00WOPPSY/Pz9NmTLFuRrGyJEjNXnyZAUEBGjjxo26//779cwzzygiIkKBgYHq2LGjPvnkE+dxWbx4sdauXaugoCBt3LhRHTp0yHVfR44cqd9++00BAQEaPny4cwUcSc46nzhxQm3atFHLli2dF/umTZvqnnvuUUhISI5rDF9Pe7/aE088ofvuu09hYWEKCwtTw4YNC3S+S3LeBT948GA99thjzqk4gwYNUtu2bfXYY4/J19dXDz30kDNwnT17VqNGjZK/v7+6du2qoKAgZ8c9aNAgbd68WYGBgc5pQgUpt7D7lNd5Il2eRnT48OFSe4GxQ15ta//+/Vq1apXmzp0rNzc3Pf7445Iur7x0RWhoqJYuXarg4GAdOHBAzz//vKT8+5JJkyapTp06evDBBxUUFKT58+crMzNTZcuWVXh4uAYMGKCAgADt2bNHHTp00LBhwzRu3Dj5+fkpNDRU27dvl1Sw68GVcjds2CA/Pz9NnTrVGQqv1PmNN97Q1q1b1bx5c3Xq1Mm5ctCVFWiCg4NznHOcX59bENOmTdPFixfVvHlzTZ48Wb179871tZ9++qnatm0rPz8/vf/++87jn9/1Ji+5tbsePXqoXr16zjaf9dhJxZcBKlasqGXLlmnlypUKDg7W66+/rmXLljkH/Kwo7Hl7vcc7t3ILu095nbOSnPcjNWzY0NL9aCXBxZg81sEpJbZt26YZM2YUaCI9YLfFixfrxIkTmj9/fklXBSXs9OnTateunQ4cOFCoOeKF9dNPP+k///mPtm7dyhqwFkyePFlVq1bV2LFjS7oqQKlXUu1l0KBB6t69e4GmyRWnUvnLgSkpKdq2bZsyMjIUGxurpUuX/m3m7wGAFenp6Xrrrbf04IMPEpoB3BT27dun3377TV26dCnpquSqVAZnY4wWLVqkwMBA9ezZU7Vr19bo0aNLuloAUCpERkYqMDBQ8fHxGjx4cElXBwCu26RJkzRkyBA99dRTBV4KtDjdEFM1AAAAgJJWKkecAQAAgNKG4AwAAABYQHAGAAAALCi+dZMK4fz5ZGVmFmwKduXKXkpISLKpRgD+bgrap7i6uqhiRU8ba1R60WcDKGmF6VMK0m+X6uCcmWkK3AlfeR8AFBX6FGvoswGUBnb2KUzVAAAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAF7nZufOvWrVq4cKGMMTLGaOTIkerYsaOdRQIAAAC2sC04G2M0ceJEvfvuu6pTp44OHTqkAQMGqH379nJ1tW+gOy3dIW/v8rZt/2opqRlK/OtSsZUHADcT+mwANxJbR5xdXV2VmJgoSUpMTFSVKlVsDc2S5HGLm7qP/9TWMrJa/0IPJRZbaQBwc6HPBnAjsS04u7i46KWXXtITTzyhcuXKKTk5WcuXL7erOAAAAMBWtgXnjIwMvfrqq3r55Zfl7++vn3/+WWPGjNHnn38uT09PS9uoXNnLruoVqeL8mhFA8aONAwAkG4PzwYMHFRcXJ39/f0mSv7+/ypYtq8jISDVq1MjSNhISkpSZaQpUbklc4OLj+eIPuFl5e5cvUBt3dXW5YT70AwAKxrYJx9WqVdOZM2f0xx9/SJIiIyOVkJCgf/zjH3YVCQAAANjGthFnb29vzZgxQ6NHj5aLi4skafbs2apQoYJdRQIAAAC2sXVVjbCwMIWFhdlZBACgCLH+PgDkztbgDAC4cZTU+vsAcKOgJwQAOJXE+vsAcKNgxBkAIEW+zB0AAB5zSURBVIn19wEgPwRnAICk619//0ZZho91uYGbm51tnOAMAJB0/evvs/Y+gJJW0LX3pYKtv8/ENQCAJNbfB4D8MOIMAJDE+vsAkB+CMwDAifX3ASB3TNUAAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYIG7XRs+ffq0RowY4XycmJiopKQk7dy5064iAQAAANvYFpxr1KihTz/91Pl41qxZcjgcdhUHAAAA2KpYpmqkpaVp/fr16tOnT3EUBwAAABS5YgnOW7ZsUdWqVdWwYcPiKA4AAAAocrZN1cjq448/LtRoc+XKXjbUpuh5e5cv6SoAsBFtHAAgFUNwjo2N1U8//aR58+YV+L0JCUnKzDQFek9JXODi4xOLvUwAxcPbu3yB2rirq8sN86EfAFAwtgfntWvXqlWrVqpYsaLdRQEArgOrIQFA3oolOE+ZMsXuYgAA14nVkAAgb7YH582bN9tdBACgiF1ZDWnFihUlXRUAKDX45UAAwDVYDQkArlUsq2oAAG4shVkN6Ua5KZJVUoCbm51tnOAMAMimsKshsRISgJJW0JWQpIKthsRUDQBANqyGBAA5IzgDALJZu3ZtoX60CgBudkzVAABkw2pIAJAzRpwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACxwt3Pjqampmj17tn744QeVKVNGTZo00TPPPGNnkQAAAIAtbA3Ozz//vMqUKaPNmzfLxcVFZ8+etbM4AAAAwDa2Befk5GStW7dO27Ztk4uLiyTpjjvusKs4AAAAwFa2zXE+deqUKlSooCVLlqh379565JFHtGvXLruKAwAAAGxl24izw+HQqVOn1KBBA02aNEl79+5VeHi4vvzyS3l5eVnaRuXK1l5X0ry9y5d0FQDYiDYOAJBsDM4+Pj5yd3dXaGioJKlx48aqWLGijh07pvvvv9/SNhISkpSZaQpUbklc4OLjE4u9TADFw9u7fIHauKuryw3zoR8AUDC2TdWoVKmSgoOD9d1330mSjh07poSEBNWqVcuuIgEA1yk1NVXTp09Xx44d1b17d02dOrWkqwQApYatq2rMnDlTTz31lObOnSt3d3fNmzdPt912m51FAgCuA6shAUDubA3ONWvW1Ntvv21nEQCAIsJqSACQN345EAAgidWQACA/to44AwBuHNe7GtKNclMkq6QANzc72zjBGQAg6fpXQ2IlJAAlraArIUkFWw2JqRoAAEmshgQA+WHEGQDgxGpIAJA7gjMAwInVkAAgd0zVAAAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwwN3Ojbdt21YeHh4qU6aMJGnChAlq0aKFnUUCAAAAtrA1OEvSokWLVKdOHbuLAQAAAGzFVA0AAADAAttHnCdMmCBjjPz9/TVu3DjddtttdhcJAAAAFDlbg/O7774rHx8fpaWladasWYqIiND8+fMtv79yZS8ba1d0vL3Ll3QVANiINg4AkGwOzj4+PpIkDw8PPfzww/r3v/9doPcnJCQpM9MU6D0lcYGLj08s9jIBFA9v7/IFauOuri43zIf+nHBTNwDkzrbgfPHiRTkcDpUvX17GGG3cuFH169e3qzgAQBHhpm4AyJltwTkhIUFPPvmkHA6HMjMzVbt2bU2fPt2u4gAAAABb2Raca9asqXXr1tm1eQCATbipGwByZvuqGgCAG8f13NR9o8zt5mZP4OZmZxsnOAMAnK7npm5u6AZQ0gp6Q7dUsJu6+QEUAICkyzd1JyZevuBwUzcAXIsRZwCAJG7qBoD8EJwBAJK4qRsA8sNUDQAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMCCYgnOS5YsUd26dXX48OHiKA4AAAAocrYH5wMHDmjPnj2qXr263UUBAAAAtrE1OKelpSkiIkIzZsywsxgAAADAdrYG54ULFyosLEw1atSwsxgAAADAdu52bXj37t3av3+/JkyYUOhtVK7sVYQ1so+3d/mSrgIAG/3d2viSJUu0ePFirV+/XnXq1Cnp6gBAqWFbcP7pp58UGRmpdu3aSZLOnDmjoUOH6rnnnlNISIilbSQkJCkz0xSo3JK4wMXHJxZ7mQCKh7d3+QK1cVdXlxvmQ39OuC8FAHJn21SN4cOH69tvv9WWLVu0ZcsWVatWTStWrLAcmgEAxYv7UgAgb6zjDACQxH0pAJAf26ZqXG3Lli3FVRQAoIC4LwXAzcLONl5swRkAUHpxXwqAm0FB70uRCnZvCsEZAKDhw4dr+PDhzsdt27bVsmXLWFUDALJgjjMAAABgASPOAIBrcF8KAFyLEWcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAnc7N/7EE0/o9OnTcnV1Vbly5TR16lTVr1/fziIBAAAAW9ganOfOnavy5ctLkr766is99dRTWrt2rZ1FAgAAALawdarGldAsSUlJSXJxcbGzOAAAAMA2to44S9KUKVP03XffyRij119/3e7iAADXgSl2AJA724PzrFmzJEnr1q3TvHnz9Nprr1l+b+XKXnZVq0h5e5fP/0UAblh/pzbOFDsAyJ3twfmKnj17atq0aTp//rwqVqxo6T0JCUnKzDQFKqckLnDx8YnFXiaA4uHtXb5AbdzV1eWG+dCfE6bYAUDubAvOycnJ+uuvv+Tj4yNJ2rJli26//XZVqFDBriIBAEWAKXYAkDPbgvOlS5c0evRoXbp0Sa6urrr99tu1bNkyRi8AoJQr7BS7G2Wk/e809Qb4O7KzjdsWnO+44w6tWbPGrs0DAGxW0Cl2TK8DUNIKOr1OKtgUO345EAAg6fIUu5iYGOdjptgBQHbFdnMgAKB0Y4odAOSN4AwAkMQUOwDID1M1AAAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAne7Nnz+/HlNnDhRJ0+elIeHh2rVqqWIiAhVqlTJriIBAAAA29g24uzi4qJhw4Zp8+bNWr9+vWrWrKn58+fbVRwAAABgK9uCc4UKFRQcHOx83KRJE0VHR9tVHADgOp0/f16PP/64OnXqpO7du2vkyJE6d+5cSVcLAEqNYpnjnJmZqdWrV6tt27bFURwAoBD4phAA8mbbHOesnnnmGZUrV04DBw4s0PsqV/ayqUZFy9u7fElXAYCN/i5tPKdvClevXl2CNQKA0sX24Dx37lydOHFCy5Ytk6trwQa4ExKSlJlpCvSekrjAxccnFnuZAIqHt3f5ArVxV1eXG+ZDf14K803hjbLff5cPQsDflZ1t3Nbg/OKLL2r//v1avny5PDw87CwKAFCECvNNIYMdAEpaQQc7pIINeNgWnI8cOaJXX31V//znP9W/f39JUo0aNbR06VK7igTwN1D+trK6tUyxzDKTJKWlO4qtrNLier4pBICbmW1Xn3vvvVe///67XZsH8Dd1axl3dR//abGVt/6FHsVWVmnAN4UAkLviG7YBAJRqfFMIAHkjOAMAJPFNIQDkh8lrAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsYDk6AAAA2OJm+7VXgjMAAABscbP92itTNQAAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAluD89y5c9W2bVvVrVtXhw8ftrMoAAAAwFa2Bud27drp3XffVfXq1e0sBgBQBBjsAIC82RqcAwIC5OPjY2cRAIAiwmAHAOTNvaQrAAAoHQICAkq6CgBQqpXq4Fy5sldJV8ESb+/yJV0FADaijQMApFIenBMSkpSZaQr0npK4wMXHJxZ7mcDfVWlv466uLjfMh/6idqPsNx+EgJubnW28VAdnAMCNg8EOAFe7Edp4QQY8bL058Nlnn1XLli115swZDRkyRN26dbOzOAAAAMA2to44P/3003r66aftLAIAUESeffZZffHFFzp79qyGDBmiChUq6PPPPy/pagFAqcFUDQCAJAY7ACA//OQ2AAAAYAHBGQAAALCA4AwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAMAAAAWEBwBgAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMACgjMAAABgAcEZAAAAsIDgDAAAAFhAcAYAAAAsIDgDAAAAFhCcAQAAAAsIzgAAAIAFBGcAAADAAoIzAAAAYAHBGQAAALCA4AwAAABYQHAGAAAALLA1OB87dkz9+vVTp06d1K9fPx0/ftzO4gAA14l+GwByZ2twnj59uh5++GFt3rxZDz/8sKZNm2ZncQCA60S/DQC5sy04JyQk6LffflNoaKgkKTQ0VL/99pvOnTtnV5EAgOtAvw0AeXO3a8MxMTGqWrWq3NzcJElubm6qUqWKYmJiVKlSJUvbcHV1KVTZVSqWLdT7Cquw9QRQOKW5jd/I/cH19tv02QByUtrbeEFeb1twLgoVK3oW6n0rnu5YxDXJW+XKXsVaHvB3RxsvneizAeTkZmrjtk3V8PHxUWxsrBwOhyTJ4XAoLi5OPj4+dhUJALgO9NsAkDfbgnPlypVVv359bdiwQZK0YcMG1a9f3/I0DQBA8aLfBoC8uRhjjF0bj4yM1OTJk/XXX3/ptttu09y5c3X33XfbVRwA4DrRbwNA7mwNzgAAAMDNgl8OBAAAACwgOAMAAAAWEJwBAAAACwjOAAAAgAUEZwAAAMCCmy4479ixQ99++62l19atW1fJyck21wjA380nn3yiY8eO5fu6HTt2qHfv3sVQo5sLxxdAbuzOgTddcN65c6e+++67kq4GgL+xtWvX6vjx4yVdjZsWxxdAbuzOge62bbkQ6tatqzFjxuirr77ShQsX9Oyzz+r777/XN998o4yMDC1cuFC1a9dWfHy8xo0bp+TkZKWmpqpVq1aaOHGifv/9d73//vvKzMzU999/r27dumn48OHaunWrFi9erIyMDLm6umrOnDmqV6+eJOntt9/Wl19+qQsXLmjixInq1KlTCR8FALmpW7euxo4dm2Ob3b59u1588UU5HA5VqlRJERERqlWrlnbs2KHZs2ercePG2r17t1xcXLRgwQLVrl37mu3v2LFDs2bNUqNGjbR37165u7tr3rx5WrJkiY4cOSIfHx8tXrxY5cqV0w8//KCXXnpJqampcjgcCg8PV7du3fTxxx9r//79evbZZ/XSSy9p0qRJatasmV599VVt2LBBLi4uKleunN577z1Jl3/Wetq0afnWrThwfAGUpBsiB5pSpE6dOuadd94xxhizceNG06RJE7NlyxZjjDHLly8348ePN8YYk5KSYpKSkowxxqSlpZlHHnnEbNu2zRhjzKJFi8ycOXOc2/zjjz9Ms2bNzLFjx4wxxqSmpprExERneW+//bYxxphdu3aZkJAQ+3cSQKHl1mbPnj1rgoODzZEjR4wxxqxZs8Y8+OCDxhhjfvzxR9OgQQNz4MABY4wxL7/8shk3blyO27/y2t9++80YY8yMGTNMixYtTExMjDHGmGHDhpk1a9YYY4y5cOGCycjIMMYYEx8fb1q0aGEuXLhgjDFm4MCBzr7LGGM++eQT89BDDzn7nnPnzhW4bsWB4wugJN0IObDUTdXo0qWLJKlhw4aSpDZt2kiS7rvvPp08eVLS5RGEefPmKSwsTL1799aRI0d06NChHLf3/fffq2XLlvrnP/8pSfLw8JCXl5fz+a5du0qSmjRpori4OKWmptqyXwCKRk5tdu/evapXr57uueceSVKfPn108OBBJSUlSZLuuusuNWjQwPm+U6dO5br9u+66S/Xr15ckNWjQQPXr11e1atUkXe6XTpw4IUk6d+6cRo0apdDQUA0dOlR//vlnrvNut27dqgEDBjj7nooVK2Yrz2rdigPHF0BJKu05sFRN1ZCkMmXKSJJcXV3l4eHh/H9XV1dlZGRIklauXKm//vpLH374ocqUKaOpU6cWOvBeKc/NzU2SlJGR4fw/AKVPTm02P7n1JTNnztQvv/wiSVqwYME1r3Vzc8vWH7i5uTn7mhkzZqht27ZasmSJXFxc1KlTp0L1Q7nVraRwfAGUpNKeA0vdiLMViYmJ8vb2VpkyZRQbG6v//e9/zue8vLyUmJjofNy8eXNt377deSNJWlqac5QEwM2hSZMmOnTokCIjIyVdvnmsQYMG2UYVcjJ9+nR9+umn+vTTT3X33XcXqMzExERVr15dLi4u+u6775wjpZLk6emZrR9q06aNVq9e7ex7zp8/X6CyShrHF0BpUpI5sNSNOFvxyCOPaPTo0QoNDVXVqlX1wAMPOJ9r37691q1bpx49ejgnhT/zzDMaO3asHA6H3NzcNGfOHNWtW7cE9wBAUapUqZLmzZunCRMmKCMjQ5UqVdLzzz9va5njx4/XzJkztXjxYt1///3Z+pR+/fppzpw5WrFihSZNmqSePXsqNjZW/fr1k7u7u8qVK6d3333X1voVJY4vgNKkJHOgizHGFNWOAAAAADerG3KqBgAAAFDcCM4AAACABQRnAAAAwAKCMwAAAGABwRkAAACwgOAM/P8eeeQRffjhh7aXM2zYMK1du9b2cgDgZkafjZJAcMZN4fTp06pbt26p+1WwxYsXa8KECdn+7/XXX1evXr1KqEYAUPLos3GjIjgDAAAAFhCcUWixsbF68skn1bRpU7Vt21ZvvfWWJOnChQtq2bKltmzZIklKTk5Whw4dtG7dOknS5MmTNW3aNA0ZMkS+vr4aOHCgoqKinNuNjIzUkCFDFBQUpE6dOmnjxo3O51JSUjRnzhy1adNG/v7+GjBggFJSUjRw4EBJUmBgoHx9fbV7925J0kcffaQuXbooMDBQQ4cOzVbOd999p86dO8vf318RERHK67eA9u3bp379+ikgIEAhISGKiIhQWlqa8/kjR44469ysWTMtW7ZM27dv16uvvqpNmzbJ19dXYWFhkrJ/vZiZmamXX35Zbdq00QMPPKCJEyc6fyr0yojM2rVr1bp1awUHB+uVV14p5F8LwN8dfTZ9NoqAAQrB4XCYXr16mcWLF5vU1FRz8uRJ07ZtW7N9+3ZjjDHffPONadasmTl79qyZMmWKefLJJ53vnTRpkmnSpInZuXOnSU1NNc8884zp37+/McaY5ORk07JlS/PRRx+Z9PR0c+DAARMUFGSOHDlijDFmxowZZuDAgebMmTMmIyPD/PzzzyY1NdWcOnXK1KlTx6SnpzvL+fLLL0379u3N0aNHTXp6ulm6dKnp16+fMcaYhIQE06RJE7Np0yaTlpZmVq5caerXr2/WrFmT4/7++uuvZvfu3SY9Pd2cOnXKdO7c2axcudIYY0xiYqJp3ry5WbFihUlJSTGJiYlmz549xhhjFi1aZMaPH59tWwMHDnSW8+GHH5r27dubkydPmqSkJDNixAgzYcIEY4xx7tOUKVPMpUuXzMGDB03Dhg3N0aNHr+tvB+Dvhz6bPhtFgxFnFMqvv/6qc+fOaeTIkfLw8FDNmjX10EMPOUcaQkJC1LlzZw0ePFjbtm3TzJkzs72/devWCgwMlIeHh8aOHas9e/YoJiZGX3/9tapXr64+ffrI3d1dDRo0UKdOnfTf//5XmZmZ+vjjjzVlyhRVrVpVbm5u8vPzk4eHR451fP/99zV8+HDVrl1b7u7uCg8P18GDBxUVFaXt27fr3nvvVefOnXXLLbfo0Ucf1R133JHr/t53331q0qSJ3N3dVaNGDfXr108//fSTJOnrr7/WHXfcoccee0xlypSRl5eXGjdubOk4rl+/XoMHD1bNmjXl6empcePGaePGjdnm/Y0cOVK33nqr6tWrp3r16unQoUOWtg0AV9Bn02ejaLiXdAVwY4qKilJcXJwCAgKc/+dwOLI9fuihh/TOO+8oPDxcFStWzPb+atWqOf/t6emp22+/XXFxcYqKitK+ffuu2W5YWJjOnz+v1NRU1axZ01Ido6OjNXv2bM2dO9f5f8YYxcbGKi4uLlsdXFxc5OPjk+u2jh07pjlz5mj//v26dOmSHA6HGjZsKEmKiYnRP/7xD0t1ulpcXJyqV6/ufFy9enVlZGQoISHB+X9ZLw5ly5bVxYsXC1UWgL8v+mz6bBQNgjMKxcfHRzVq1NAXX3yR4/MOh0PTpk1Tz5499d5776l3796qVauW8/kzZ844/52cnKw///xTVapUkY+PjwIDA7Vy5cprtpmZmakyZcro1KlTqlevXrbnXFxccqxjeHi4c55aVidOnMhWB2OMYmJict3fGTNmqEGDBnrhhRfk5eWlVatWafPmzc5yss7py69eWVWpUiXbHL7o6Gi5u7urcuXK2eoHANeDPps+G0WDqRoolEaNGsnT01PLly9XSkqKHA6HDh8+rH379kmSli1bJhcXF82ePVtDhw7VpEmT5HA4nO/ftm2bdu3apbS0NC1cuFCNGzeWj4+PWrdurePHj2vdunVKT09Xenq69u3bp8jISLm6uqpPnz567rnnFBsbK4fDod27dystLU2VKlWSq6urTp065Syjf//+Wr58uY4cOSJJSkxM1KZNmyRJrVq10pEjR/TFF18oIyNDb731ls6ePZvr/iYnJ8vT01Oenp6KjIzU6tWrnc+1bt1a8fHxWrVqldLS0pSUlKS9e/dKkipXrqyoqChlZmbmuN3Q0FC9+eabOnXqlJKTk7VgwQJ16dJF7u58pgVQdOiz6bNRNAjOKBQ3NzctW7ZMhw4dUrt27dS0aVM9/fTTSkpK0v79+7Vq1SrNnTtXbm5uevzxxyVJy5cvd74/NDRUS5cuVXBwsA4cOKDnn39ekuTl5aUVK1Zo48aNatGihUJCQjR//nzn3dCTJk1SnTp19OCDDyooKEjz589XZmamypYtq/DwcA0YMEABAQHas2ePOnTooGHDhmncuHHy8/NTaGiotm/fLkmqVKmSFi5cqBdeeEHBwcE6ceKE/Pz8ct3fSZMmacOGDfLz89PUqVPVtWtX53NeXl564403tHXrVjVv3lydOnXSjh07JEmdO3eWJAUHB+e4DmifPn0UFhamgQMHql27dvLw8NDUqVOv508DANegz6bPRtFwMSaP9VwAG0yePFlVq1bV2LFjS7oqAIB80GcD/w8jzgAAAIAFBGcAAADAAqZqAAAAABYw4gwAAABYQHAGAAAALCA4AwAAABYQnAEAAAALCM4AAACABQRnAAAAwIL/D/ghS4Ld6HvCAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "e0-ObUpxpids"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## feed mock data to agent"
      ],
      "metadata": {
        "id": "gNqotihKSch3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "N=2\n",
        "n_episodes = 500\n",
        "episode_steps=32\n",
        "\n",
        "mock_nback_data = generate_mock_nback_dataset(\n",
        "    N=2,n_participants=n_episodes,n_trials=episode_steps)\n",
        "\n",
        "env, env_spec = NBack.create_environment(\n",
        "    N=2,episode_steps=episode_steps,n_episodes=n_episodes,\n",
        "    human_data=mock_nback_data)\n",
        "\n",
        "agent = dqn.DQN(\n",
        "    environment_spec=env_spec,\n",
        "    network=dqn_make_network(env_spec.actions),\n",
        "    epsilon=[0.05],\n",
        "    discount=0.5,\n",
        "    learning_rate=1e-5,\n",
        "    min_replay_size=32,\n",
        "    logger=loggers.InMemoryLogger(),\n",
        "    checkpoint=False)\n",
        "\n",
        "loop = EnvironmentLoop(env, agent, logger=loggers.InMemoryLogger())\n",
        "loop.run(n_episodes)\n",
        "logs_dqn = pd.DataFrame(loop._logger._data)\n",
        "logs_dqn.head()\n",
        "\n",
        "env, env_spec = NBack.create_environment(\n",
        "    N=2,episode_steps=episode_steps,n_episodes=n_episodes,\n",
        "    human_data=None)\n",
        "\n",
        "agent_optimal = dqn.DQN(\n",
        "    environment_spec=env_spec,\n",
        "    network=dqn_make_network(env_spec.actions),\n",
        "    epsilon=[0.05],\n",
        "    discount=0.5,\n",
        "    learning_rate=1e-5,\n",
        "    min_replay_size=32,\n",
        "    logger=loggers.InMemoryLogger(),\n",
        "    checkpoint=False)\n",
        "\n",
        "loop = EnvironmentLoop(env, agent_optimal, logger=loggers.InMemoryLogger())\n",
        "loop.run(n_episodes)\n",
        "logs_dqn_optimal = pd.DataFrame(loop._logger._data)\n",
        "logs_dqn_optimal.head()\n",
        "\n",
        "agent_random = RandomAgent(env_spec)\n",
        "loop = EnvironmentLoop(env, agent_random, logger=loggers.InMemoryLogger())\n",
        "loop.run(n_episodes)\n",
        "logs_random = pd.DataFrame(loop._logger._data)\n",
        "episode_return=logs_random['episode_return']\n",
        "\n"
      ],
      "metadata": {
        "id": "HluY0JtFy39g"
      },
      "execution_count": 187,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(moving_average(logs_dqn['episode_return'],5),label='mock human')\n",
        "plt.plot(moving_average(logs_dqn_optimal['episode_return'],5),label='not mock human')\n",
        "plt.plot(moving_average(episode_return,10),label=\"baseline\")\n",
        "plt.legend()\n",
        "plt.xlabel('episode')\n",
        "plt.ylabel('reward')\n",
        "plt.title('DQN')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "id": "EjsVAuareD5s",
        "outputId": "febba0a4-bece-4b71-8724-2d842749fe04"
      },
      "execution_count": 189,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'DQN')"
            ]
          },
          "metadata": {},
          "execution_count": 189
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "mock_nback_data = generate_mock_nback_dataset(\n",
        "    N=2,n_participants=n_episodes,n_trials=episode_steps)\n",
        "\n",
        "df=mock_nback_data.head(10000)\n",
        "                \n",
        "sns.barplot(data=df, y='is_correct', x='participant_id')\n",
        "plt.suptitle('Accuracy distribution of the mock N-back dataset', y=1.06)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "NRGhbg_8T2uj",
        "outputId": "233a984c-f3aa-4ef0-8773-e6e90eec7f9f"
      },
      "execution_count": 173,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "mock_nback_data.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vWCLms4EXHXw",
        "outputId": "5ca82532-23e4-4bed-a1a9-65746ce013a2"
      },
      "execution_count": 190,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 16000 entries, 0 to 15999\n",
            "Data columns (total 7 columns):\n",
            " #   Column             Non-Null Count  Dtype  \n",
            "---  ------             --------------  -----  \n",
            " 0   participant_id     16000 non-null  object \n",
            " 1   trial_index        16000 non-null  int64  \n",
            " 2   stimulus           16000 non-null  object \n",
            " 3   response           15000 non-null  object \n",
            " 4   response_time      15000 non-null  float64\n",
            " 5   expected_response  15000 non-null  object \n",
            " 6   is_correct         16000 non-null  bool   \n",
            "dtypes: bool(1), float64(1), int64(1), object(4)\n",
            "memory usage: 765.8+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "O6jWr-wzX77o"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "human_rl_v2",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernel": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}