Investigating MCTS modifications in general video game playing

Frederik Frydenberg, Kasper R. Andersen, Sebastian Risi, Julian Togelius

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Abstract

    While Monte Carlo tree search (MCTS) methods have shown promise in a variety of different board games, more complex video games still present significant challenges. Recently, several modifications to the core MCTS algorithm have been proposed with the hope to increase its effectiveness on arcade-style video games. This paper investigates of how well these modifications perform in general video game playing using the general video game AI (GVG-AI) framework and introduces a new MCTS modification called UCT reverse penalty that penalizes the MCTS controller for exploring recently visited children. The results of our experiments show that a combination of two MCTS modifications can improve the performance of the vanilla MCTS controller, but the effectiveness of the modifications highly depends on the particular game being played.

    Original languageEnglish (US)
    Title of host publication2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages107-113
    Number of pages7
    ISBN (Print)9781479986217
    DOIs
    StatePublished - Nov 4 2015
    Event2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Tainan, Taiwan, Province of China
    Duration: Aug 31 2015Sep 2 2015

    Other

    Other2015 IEEE Conference on Computational Intelligence and Games, CIG 2015
    CountryTaiwan, Province of China
    CityTainan
    Period8/31/159/2/15

    Fingerprint

    Controllers
    Experiments

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Computer Science Applications
    • Information Systems

    Cite this

    Frydenberg, F., Andersen, K. R., Risi, S., & Togelius, J. (2015). Investigating MCTS modifications in general video game playing. In 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings (pp. 107-113). [7317937] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CIG.2015.7317937

    Investigating MCTS modifications in general video game playing. / Frydenberg, Frederik; Andersen, Kasper R.; Risi, Sebastian; Togelius, Julian.

    2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015. p. 107-113 7317937.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Frydenberg, F, Andersen, KR, Risi, S & Togelius, J 2015, Investigating MCTS modifications in general video game playing. in 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings., 7317937, Institute of Electrical and Electronics Engineers Inc., pp. 107-113, 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015, Tainan, Taiwan, Province of China, 8/31/15. https://doi.org/10.1109/CIG.2015.7317937
    Frydenberg F, Andersen KR, Risi S, Togelius J. Investigating MCTS modifications in general video game playing. In 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2015. p. 107-113. 7317937 https://doi.org/10.1109/CIG.2015.7317937
    Frydenberg, Frederik ; Andersen, Kasper R. ; Risi, Sebastian ; Togelius, Julian. / Investigating MCTS modifications in general video game playing. 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2015. pp. 107-113
    @inproceedings{50015b70d57c485385124ebd2e4de0b1,
    title = "Investigating MCTS modifications in general video game playing",
    abstract = "While Monte Carlo tree search (MCTS) methods have shown promise in a variety of different board games, more complex video games still present significant challenges. Recently, several modifications to the core MCTS algorithm have been proposed with the hope to increase its effectiveness on arcade-style video games. This paper investigates of how well these modifications perform in general video game playing using the general video game AI (GVG-AI) framework and introduces a new MCTS modification called UCT reverse penalty that penalizes the MCTS controller for exploring recently visited children. The results of our experiments show that a combination of two MCTS modifications can improve the performance of the vanilla MCTS controller, but the effectiveness of the modifications highly depends on the particular game being played.",
    author = "Frederik Frydenberg and Andersen, {Kasper R.} and Sebastian Risi and Julian Togelius",
    year = "2015",
    month = "11",
    day = "4",
    doi = "10.1109/CIG.2015.7317937",
    language = "English (US)",
    isbn = "9781479986217",
    pages = "107--113",
    booktitle = "2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",

    }

    TY - GEN

    T1 - Investigating MCTS modifications in general video game playing

    AU - Frydenberg, Frederik

    AU - Andersen, Kasper R.

    AU - Risi, Sebastian

    AU - Togelius, Julian

    PY - 2015/11/4

    Y1 - 2015/11/4

    N2 - While Monte Carlo tree search (MCTS) methods have shown promise in a variety of different board games, more complex video games still present significant challenges. Recently, several modifications to the core MCTS algorithm have been proposed with the hope to increase its effectiveness on arcade-style video games. This paper investigates of how well these modifications perform in general video game playing using the general video game AI (GVG-AI) framework and introduces a new MCTS modification called UCT reverse penalty that penalizes the MCTS controller for exploring recently visited children. The results of our experiments show that a combination of two MCTS modifications can improve the performance of the vanilla MCTS controller, but the effectiveness of the modifications highly depends on the particular game being played.

    AB - While Monte Carlo tree search (MCTS) methods have shown promise in a variety of different board games, more complex video games still present significant challenges. Recently, several modifications to the core MCTS algorithm have been proposed with the hope to increase its effectiveness on arcade-style video games. This paper investigates of how well these modifications perform in general video game playing using the general video game AI (GVG-AI) framework and introduces a new MCTS modification called UCT reverse penalty that penalizes the MCTS controller for exploring recently visited children. The results of our experiments show that a combination of two MCTS modifications can improve the performance of the vanilla MCTS controller, but the effectiveness of the modifications highly depends on the particular game being played.

    UR - http://www.scopus.com/inward/record.url?scp=84964507597&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84964507597&partnerID=8YFLogxK

    U2 - 10.1109/CIG.2015.7317937

    DO - 10.1109/CIG.2015.7317937

    M3 - Conference contribution

    SN - 9781479986217

    SP - 107

    EP - 113

    BT - 2015 IEEE Conference on Computational Intelligence and Games, CIG 2015 - Proceedings

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -