Scribd
Upload
66 views3 pages

Course Objectives: Teaching A Reinforcement Learning Agent To Play Atari Using Deep Q-Learning

This document outlines a course on teaching a reinforcement learning agent to play Atari using deep Q-learning. The course objectives are to construct and train a neural network using Q-learning techniques in Tensorflow, utilize OpenAI Gym for model training, and improve Q-learning with techniques like dueling Q and prioritized experience replay. The course is divided into an overview, a hands-on project to develop a deep Q-learning agent in Atari, and a graded quiz. The project involves tasks to explore Q-learning theory and implement a Tensorflow model, memory, policy and action functions, and a training loop to play Atari games.

Uploaded by

Nihar Ranjan Sahoo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
66 views3 pages

Course Objectives: Teaching A Reinforcement Learning Agent To Play Atari Using Deep Q-Learning

This document outlines a course on teaching a reinforcement learning agent to play Atari using deep Q-learning. The course objectives are to construct and train a neural network using Q-learning techniques in Tensorflow, utilize OpenAI Gym for model training, and improve Q-learning with techniques like dueling Q and prioritized experience replay. The course is divided into an overview, a hands-on project to develop a deep Q-learning agent in Atari, and a graded quiz. The project involves tasks to explore Q-learning theory and implement a Tensorflow model, memory, policy and action functions, and a training loop to play Atari games.

Uploaded by

Nihar Ranjan Sahoo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Teaching a Reinforcement Learning Agent

to play Atari using Deep Q-Learning

Course Objectives
In this course, we are going to focus on three main learning objectives:

1. Construct and train a Neural Network in Tensorflow using Q-Learning techniques


2. Utilize OpenAI Gym for model training
3. Improve Q-Learning techniques with enhancements such as Dueling Q and Prioritized
Experience Replay (PER)

By the end of this course, you will be able to code a basic multilayered Deep Learning model in
Python using Tensorflow, allowing you to train a reinforcement learning agent to play Atari using the
Q-Learning technique, and preparing you for further progression in your path to learning to
implement AI and machine learning techniques in Python.

Course Structure
This course is divided into 3 parts:

1. Course Overview: This introductory reading material.


2. Teaching a Reinforcement Learning Agent to play Atari using Deep Q-
Learning: This is the hands on project that we will work on in Rhyme.
3. Graded Quiz: This is the final assignment that you need to pass in order to finish the
course successfully.

Project Structure
The hands-on project on Teaching a Reinforcement Learning Agent to play Atari using Deep Q-
Learning is divided into following tasks:
Task 1: Overview Q-Learning Agents, Tensorflow, and
OpenAI Gym

Task 2: Explore Deep-Q Learning in Theory

Task 3: Building a Tensorflow Model

Task 4: Explore Activation Functions and Model Input

Task 5: Explore Keras Initializers and Optimizers

Task 6: Implement Memory, Policy, and Action Functions

Task 7: Defining the Training Function with Q-Learning

Task 8: Explore Double Q, Dueling Q and Prioritized


Experience Replay

Teaching an RL Agent to play Atari using Deep Q-Learning


Now we will write a script that will serve as a generic model of a multilayered artificial neural network
(ANN) that can be trained, using Q-learning RL techniques, to autonomously play Atari. We will
accomplish this by completing each task in the project:

 Develop a Tensorflow model using both convolutional and dense layers


 Learn how to connect our agent to OpenAI's Gym interface
 Review types of activation functions, initializers, and optimizers
 Implement a memory class to allow the agent to use past experience during training in
experience replay
 Implement a loss function
 Implement a training loop
 Discuss the benefits and advantages of enhanced Q-learning techniques such as
Double Q, Dueling Q, and Prioritized Experience Replay

You might also like