Saved in:
Bibliographic Details
Main Authors: Saadat, Kimiya, Zhao, Richard
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.16653
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909476969250816
author Saadat, Kimiya
Zhao, Richard
author_facet Saadat, Kimiya
Zhao, Richard
contents Playing two-player games using reinforcement learning and self-play can be challenging due to the complexity of two-player environments and the possible instability in the training process. We propose that a reinforcement learning algorithm can train more efficiently and achieve improved performance in a two-player game if it leverages the knowledge from the single-player version of the same game. This study examines the proposed idea in ten different Atari 2600 environments using the Atari 2600 RAM as the input state. We discuss the advantages of using transfer learning from a single-player training process over training in a two-player setting from scratch, and demonstrate our results in a few measures such as training time and average total reward. We also discuss a method of calculating RAM complexity and its relationship to performance.
format Preprint
id arxiv_https___arxiv_org_abs_2410_16653
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing Two-Player Performance Through Single-Player Knowledge Transfer: An Empirical Study on Atari 2600 Games
Saadat, Kimiya
Zhao, Richard
Machine Learning
Artificial Intelligence
Playing two-player games using reinforcement learning and self-play can be challenging due to the complexity of two-player environments and the possible instability in the training process. We propose that a reinforcement learning algorithm can train more efficiently and achieve improved performance in a two-player game if it leverages the knowledge from the single-player version of the same game. This study examines the proposed idea in ten different Atari 2600 environments using the Atari 2600 RAM as the input state. We discuss the advantages of using transfer learning from a single-player training process over training in a two-player setting from scratch, and demonstrate our results in a few measures such as training time and average total reward. We also discuss a method of calculating RAM complexity and its relationship to performance.
title Enhancing Two-Player Performance Through Single-Player Knowledge Transfer: An Empirical Study on Atari 2600 Games
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2410.16653