Association Rule Mining for Unknown Video Games

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearchpeer review

Authors

External Research Organisations

  • Otto-von-Guericke University Magdeburg
View graph of relations

Details

Original languageEnglish
Title of host publicationStudies in Fuzziness and Soft Computing
PublisherSpringer Science and Business Media Deutschland GmbH
Pages257-270
Number of pages14
Publication statusPublished - 27 Oct 2020
Externally publishedYes

Publication series

NameStudies in Fuzziness and Soft Computing
Volume394
ISSN (Print)1434-9922
ISSN (electronic)1860-0808

Abstract

Computational intelligence agents can reach expert levels in many known games, such as Chess, Go, and Morris. Those systems incorporate powerful machine learning algorithms, which are able to learn from, e.g., observations, play-traces, or by reinforcement learning. While many black box systems, such as deep neural networks, are able to achieve high performance in a wide range of applications, they generally lack interpretability. Additionally, previous systems often focused on a single or a small set of games, which makes it a cumbersome task to rebuild and retrain the agent for each possible application. This paper proposes a method, which extracts an interpretable set of game rules for previously unknown games. Frequent pattern mining is used to find common observation patterns in the game environment. Finally, game rules as well as winning-/losing-conditions are extracted via association rule analysis. Our evaluation shows that a wide range of game rules can be successfully extracted from previously unknown games. We further highlight how the application of fuzzy methods can advance our efforts in generating explainable artifical intelligence (AI) agents.

ASJC Scopus subject areas

Cite this

Association Rule Mining for Unknown Video Games. / Dockhorn, Alexander; Saxton, Chris; Kruse, Rudolf.
Studies in Fuzziness and Soft Computing. Springer Science and Business Media Deutschland GmbH, 2020. p. 257-270 (Studies in Fuzziness and Soft Computing; Vol. 394).

Research output: Chapter in book/report/conference proceedingContribution to book/anthologyResearchpeer review

Dockhorn, A, Saxton, C & Kruse, R 2020, Association Rule Mining for Unknown Video Games. in Studies in Fuzziness and Soft Computing. Studies in Fuzziness and Soft Computing, vol. 394, Springer Science and Business Media Deutschland GmbH, pp. 257-270. https://doi.org/10.1007/978-3-030-54341-9_22
Dockhorn, A., Saxton, C., & Kruse, R. (2020). Association Rule Mining for Unknown Video Games. In Studies in Fuzziness and Soft Computing (pp. 257-270). (Studies in Fuzziness and Soft Computing; Vol. 394). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-54341-9_22
Dockhorn A, Saxton C, Kruse R. Association Rule Mining for Unknown Video Games. In Studies in Fuzziness and Soft Computing. Springer Science and Business Media Deutschland GmbH. 2020. p. 257-270. (Studies in Fuzziness and Soft Computing). doi: 10.1007/978-3-030-54341-9_22
Dockhorn, Alexander ; Saxton, Chris ; Kruse, Rudolf. / Association Rule Mining for Unknown Video Games. Studies in Fuzziness and Soft Computing. Springer Science and Business Media Deutschland GmbH, 2020. pp. 257-270 (Studies in Fuzziness and Soft Computing).
Download
@inbook{dee134e3540d4c41b89ef957344ea59e,
title = "Association Rule Mining for Unknown Video Games",
abstract = "Computational intelligence agents can reach expert levels in many known games, such as Chess, Go, and Morris. Those systems incorporate powerful machine learning algorithms, which are able to learn from, e.g., observations, play-traces, or by reinforcement learning. While many black box systems, such as deep neural networks, are able to achieve high performance in a wide range of applications, they generally lack interpretability. Additionally, previous systems often focused on a single or a small set of games, which makes it a cumbersome task to rebuild and retrain the agent for each possible application. This paper proposes a method, which extracts an interpretable set of game rules for previously unknown games. Frequent pattern mining is used to find common observation patterns in the game environment. Finally, game rules as well as winning-/losing-conditions are extracted via association rule analysis. Our evaluation shows that a wide range of game rules can be successfully extracted from previously unknown games. We further highlight how the application of fuzzy methods can advance our efforts in generating explainable artifical intelligence (AI) agents.",
author = "Alexander Dockhorn and Chris Saxton and Rudolf Kruse",
year = "2020",
month = oct,
day = "27",
doi = "10.1007/978-3-030-54341-9_22",
language = "English",
series = "Studies in Fuzziness and Soft Computing",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "257--270",
booktitle = "Studies in Fuzziness and Soft Computing",
address = "Germany",

}

Download

TY - CHAP

T1 - Association Rule Mining for Unknown Video Games

AU - Dockhorn, Alexander

AU - Saxton, Chris

AU - Kruse, Rudolf

PY - 2020/10/27

Y1 - 2020/10/27

N2 - Computational intelligence agents can reach expert levels in many known games, such as Chess, Go, and Morris. Those systems incorporate powerful machine learning algorithms, which are able to learn from, e.g., observations, play-traces, or by reinforcement learning. While many black box systems, such as deep neural networks, are able to achieve high performance in a wide range of applications, they generally lack interpretability. Additionally, previous systems often focused on a single or a small set of games, which makes it a cumbersome task to rebuild and retrain the agent for each possible application. This paper proposes a method, which extracts an interpretable set of game rules for previously unknown games. Frequent pattern mining is used to find common observation patterns in the game environment. Finally, game rules as well as winning-/losing-conditions are extracted via association rule analysis. Our evaluation shows that a wide range of game rules can be successfully extracted from previously unknown games. We further highlight how the application of fuzzy methods can advance our efforts in generating explainable artifical intelligence (AI) agents.

AB - Computational intelligence agents can reach expert levels in many known games, such as Chess, Go, and Morris. Those systems incorporate powerful machine learning algorithms, which are able to learn from, e.g., observations, play-traces, or by reinforcement learning. While many black box systems, such as deep neural networks, are able to achieve high performance in a wide range of applications, they generally lack interpretability. Additionally, previous systems often focused on a single or a small set of games, which makes it a cumbersome task to rebuild and retrain the agent for each possible application. This paper proposes a method, which extracts an interpretable set of game rules for previously unknown games. Frequent pattern mining is used to find common observation patterns in the game environment. Finally, game rules as well as winning-/losing-conditions are extracted via association rule analysis. Our evaluation shows that a wide range of game rules can be successfully extracted from previously unknown games. We further highlight how the application of fuzzy methods can advance our efforts in generating explainable artifical intelligence (AI) agents.

UR - http://www.scopus.com/inward/record.url?scp=85094557810&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-54341-9_22

DO - 10.1007/978-3-030-54341-9_22

M3 - Contribution to book/anthology

AN - SCOPUS:85094557810

T3 - Studies in Fuzziness and Soft Computing

SP - 257

EP - 270

BT - Studies in Fuzziness and Soft Computing

PB - Springer Science and Business Media Deutschland GmbH

ER -

By the same author(s)