Competitive reinforcement learning in Atari games

Title
Competitive reinforcement learning in Atari games
Publication Date
2017
Author(s)
McKenzie, Mark
Loxley, Peter
( author )
OrcID: https://orcid.org/0000-0003-3659-734X
Email: ploxley@une.edu.au
UNE Id une-id:ploxley
Billingsley, William
( author )
OrcID: https://orcid.org/0000-0002-1720-9076
Email: wbilling@une.edu.au
UNE Id une-id:wbilling
Wong, Sebastien
Editor
Editor(s): Peng W, Alahakoon D, Li X
Type of document
Conference Publication
Language
en
Entity Type
Publication
Publisher
Springer
Place of publication
Germany
Series
Lecture Notes in Computer Science
DOI
10.1007/978-3-319-63004-5_2
UNE publication id
une:22306
Abstract
This research describes a study into the ability of a state of the art reinforcement learning algorithm to learn to perform multiple tasks. We demonstrate that the limitation of learning to performing two tasks can be mitigated with a competitive training method. We show that this approach results in improved generalization of the system when performing unforeseen tasks. The learning agent assessed is an altered version of the DeepMind deep Q–learner network (DQN), which has been demonstrated to outperform human players for a number of Atari 2600 games. The key findings of this paper is that there were significant degradations in performance when learning more than one game, and how this varies depends on both similarity and the comparative complexity of the two games.
Link
Citation
AI 2017: Advances in Artificial Intelligence, 10400(LNAI), p. 14-26
ISBN
9783319630045
9783319630038
Start page
14
End page
26

Files:

NameSizeformatDescriptionLink