Improvements to Increase the Efficiency of the AlphaZero Algorithm: A Case Study in the Game ’Connect 4’

Colin Clausen, Simon Reichhuber, Ingo Thomsen, Sven Tomforde

Abstract

AlphaZero is a recent approach to self-teaching gameplay without the need for human expertise. It suffers from the massive computation and hardware requirements, which are responsible for the reduced applicability of the approach. This paper focuses on possible improvements with the goal to reduce the required computation resources. We propose and investigate three modifications: We model the self-learning phase as an evolutionary process, study the game process as a tree and use network-internal features as auxiliary targets. Then behaviour and performance of these modifications are evaluated in the game Connect 4 as a test scenario.

Download


Paper Citation


in Harvard Style

Clausen C., Reichhuber S., Thomsen I. and Tomforde S. (2021). Improvements to Increase the Efficiency of the AlphaZero Algorithm: A Case Study in the Game ’Connect 4’.In Proceedings of the 13th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART, ISBN 978-989-758-484-8, pages 803-811. DOI: 10.5220/0010245908030811


in Bibtex Style

@conference{icaart21,
author={Colin Clausen and Simon Reichhuber and Ingo Thomsen and Sven Tomforde},
title={Improvements to Increase the Efficiency of the AlphaZero Algorithm: A Case Study in the Game ’Connect 4’},
booktitle={Proceedings of the 13th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,},
year={2021},
pages={803-811},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0010245908030811},
isbn={978-989-758-484-8},
}


in EndNote Style

TY - CONF

JO - Proceedings of the 13th International Conference on Agents and Artificial Intelligence - Volume 2: ICAART,
TI - Improvements to Increase the Efficiency of the AlphaZero Algorithm: A Case Study in the Game ’Connect 4’
SN - 978-989-758-484-8
AU - Clausen C.
AU - Reichhuber S.
AU - Thomsen I.
AU - Tomforde S.
PY - 2021
SP - 803
EP - 811
DO - 10.5220/0010245908030811