Multi-objective adaptation of a parameterized GVGAI agent towards several games

Ahmed Khalifa, Mike Preuss, Julian Togelius

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

    Abstract

    This paper proposes a benchmark for multi-objective optimization based on video game playing. The challenge is to optimize an agent to perform well on several different games, where each objective score corresponds to the performance on a different game. The benchmark is inspired from the quest for general intelligence in the form of general game playing, and builds on the General Video Game AI (GVGAI) framework. As it is based on game-playing, this benchmark incorporates salient aspects of game-playing problems such as discontinuous feedback and a non-trivial amount of stochasticity. We argue that the proposed benchmark thus provides a different challenge from many other benchmarks for multi-objective optimization algorithms currently available. We also provide initial results on categorizing the space offered by this benchmark and applying a standard multi-objective optimization algorithm to it.

    Original languageEnglish (US)
    Title of host publicationEvolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings
    PublisherSpringer Verlag
    Pages359-374
    Number of pages16
    Volume10173 LNCS
    ISBN (Print)9783319541563
    DOIs
    StatePublished - 2017
    Event9th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2017 - Munster, Germany
    Duration: Mar 19 2017Mar 22 2017

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume10173 LNCS
    ISSN (Print)03029743
    ISSN (Electronic)16113349

    Other

    Other9th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2017
    CountryGermany
    CityMunster
    Period3/19/173/22/17

    Fingerprint

    Video Games
    Multiobjective optimization
    Game
    Benchmark
    Multi-objective Optimization
    Optimization Algorithm
    Feedback
    Stochasticity
    Optimise

    Keywords

    • GVGAI
    • MCTS
    • Multi-objective optimization

    ASJC Scopus subject areas

    • Theoretical Computer Science
    • Computer Science(all)

    Cite this

    Khalifa, A., Preuss, M., & Togelius, J. (2017). Multi-objective adaptation of a parameterized GVGAI agent towards several games. In Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings (Vol. 10173 LNCS, pp. 359-374). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10173 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-54157-0_25

    Multi-objective adaptation of a parameterized GVGAI agent towards several games. / Khalifa, Ahmed; Preuss, Mike; Togelius, Julian.

    Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings. Vol. 10173 LNCS Springer Verlag, 2017. p. 359-374 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10173 LNCS).

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

    Khalifa, A, Preuss, M & Togelius, J 2017, Multi-objective adaptation of a parameterized GVGAI agent towards several games. in Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings. vol. 10173 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10173 LNCS, Springer Verlag, pp. 359-374, 9th International Conference on Evolutionary Multi-Criterion Optimization, EMO 2017, Munster, Germany, 3/19/17. https://doi.org/10.1007/978-3-319-54157-0_25
    Khalifa A, Preuss M, Togelius J. Multi-objective adaptation of a parameterized GVGAI agent towards several games. In Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings. Vol. 10173 LNCS. Springer Verlag. 2017. p. 359-374. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-54157-0_25
    Khalifa, Ahmed ; Preuss, Mike ; Togelius, Julian. / Multi-objective adaptation of a parameterized GVGAI agent towards several games. Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings. Vol. 10173 LNCS Springer Verlag, 2017. pp. 359-374 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
    @inproceedings{c0b10f1c93a34878b045705915c3c522,
    title = "Multi-objective adaptation of a parameterized GVGAI agent towards several games",
    abstract = "This paper proposes a benchmark for multi-objective optimization based on video game playing. The challenge is to optimize an agent to perform well on several different games, where each objective score corresponds to the performance on a different game. The benchmark is inspired from the quest for general intelligence in the form of general game playing, and builds on the General Video Game AI (GVGAI) framework. As it is based on game-playing, this benchmark incorporates salient aspects of game-playing problems such as discontinuous feedback and a non-trivial amount of stochasticity. We argue that the proposed benchmark thus provides a different challenge from many other benchmarks for multi-objective optimization algorithms currently available. We also provide initial results on categorizing the space offered by this benchmark and applying a standard multi-objective optimization algorithm to it.",
    keywords = "GVGAI, MCTS, Multi-objective optimization",
    author = "Ahmed Khalifa and Mike Preuss and Julian Togelius",
    year = "2017",
    doi = "10.1007/978-3-319-54157-0_25",
    language = "English (US)",
    isbn = "9783319541563",
    volume = "10173 LNCS",
    series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
    publisher = "Springer Verlag",
    pages = "359--374",
    booktitle = "Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings",
    address = "Germany",

    }

    TY - GEN

    T1 - Multi-objective adaptation of a parameterized GVGAI agent towards several games

    AU - Khalifa, Ahmed

    AU - Preuss, Mike

    AU - Togelius, Julian

    PY - 2017

    Y1 - 2017

    N2 - This paper proposes a benchmark for multi-objective optimization based on video game playing. The challenge is to optimize an agent to perform well on several different games, where each objective score corresponds to the performance on a different game. The benchmark is inspired from the quest for general intelligence in the form of general game playing, and builds on the General Video Game AI (GVGAI) framework. As it is based on game-playing, this benchmark incorporates salient aspects of game-playing problems such as discontinuous feedback and a non-trivial amount of stochasticity. We argue that the proposed benchmark thus provides a different challenge from many other benchmarks for multi-objective optimization algorithms currently available. We also provide initial results on categorizing the space offered by this benchmark and applying a standard multi-objective optimization algorithm to it.

    AB - This paper proposes a benchmark for multi-objective optimization based on video game playing. The challenge is to optimize an agent to perform well on several different games, where each objective score corresponds to the performance on a different game. The benchmark is inspired from the quest for general intelligence in the form of general game playing, and builds on the General Video Game AI (GVGAI) framework. As it is based on game-playing, this benchmark incorporates salient aspects of game-playing problems such as discontinuous feedback and a non-trivial amount of stochasticity. We argue that the proposed benchmark thus provides a different challenge from many other benchmarks for multi-objective optimization algorithms currently available. We also provide initial results on categorizing the space offered by this benchmark and applying a standard multi-objective optimization algorithm to it.

    KW - GVGAI

    KW - MCTS

    KW - Multi-objective optimization

    UR - http://www.scopus.com/inward/record.url?scp=85014225919&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85014225919&partnerID=8YFLogxK

    U2 - 10.1007/978-3-319-54157-0_25

    DO - 10.1007/978-3-319-54157-0_25

    M3 - Conference contribution

    SN - 9783319541563

    VL - 10173 LNCS

    T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

    SP - 359

    EP - 374

    BT - Evolutionary Multi-Criterion Optimization - 9th International Conference, EMO 2017, Proceedings

    PB - Springer Verlag

    ER -