Publication Date

Spring 2023

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Chris Pollett

Second Advisor

Mark Stamp

Third Advisor

Kevin Smith


Deep Q-Learning, Video games, Atari, Reinforcement learning


Games for the Atari 2600 console provide great environments for testing reinforcement learning algorithms. In reinforcement learning algorithms, an agent typically learns about its environment via the delivery of periodic rewards. Deep Q-Learning, a variant of Q-Learning, utilizes neural networks which train a Q-function to predict the highest future reward given an input state and action. Deep Q-learning has shown great results in training agents to play Atari 2600 games like Space Invaders and Breakout. However, Deep Q-Learning has historically struggled with learning how to play games with greater emphasis on exploration and delayed rewards, like Ms. PacMan. In this project, we train a neural network that learns control policies for Ms. PacMan. We experiment with various methods to boost Q-agent performance. Through our novel training method, Frame-Diff, we are able to not only optimize, but beat, known benchmarks for reinforcement learning agents for Ms. PacMan by up to 50%.