frozen_lake.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Lets start with importing the libraries required"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import gym\n",
    "import numpy as np\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading the Frozen Lake Environment from OpenAI Gym"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "env = gym.make('FrozenLake-v0')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define the Placeholder for the Input Data and Variables for the Weights"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The activation function we are using is the argmax, which returns the maximum value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "import gym\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "env = gym.make('FrozenLake-v0')\n",
    "#The grid with 4x4 gives 16 possible states, hence we have an array of 16 states.\n",
    "inputs = tf.placeholder(shape=[1,16],dtype=tf.float32)\n",
    "#Each state has 4 possible outcomes, hence we have 16x4 matrix with weights uniformly distributed\n",
    "weights = tf.Variable(tf.random_uniform([16,4],0,0.1))\n",
    "#Find the dot product of inputs and the weights\n",
    "Q1 = tf.matmul(inputs,weights)\n",
    "#The next state will be the opted based on the argmax function.\n",
    "output = tf.argmax(Q1,1)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

Thanks Bhaiya its awesome