Behavioral evaluation of hanabi rainbow DQN agents and rule-based agents

Rodrigo Canaan, Xianbo Gao, Youjin Chung, Julian Togelius, Andy Nealen, Stefan Menzel

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

    Abstract

    Hanabi is a multiplayer cooperative card game, where only your partners know your cards. All players succeed or fail together. This makes the game an excellent testbed for studying collaboration. Recently, it has been shown that deep neural networks can be trained through self-play to play the game very well. However, such agents generally do not play well with others. In this paper, we investigate the consequences of training Rainbow DQN agents with human-inspired rule-based agents. We analyze with which agents Rainbow agents learn to play well, and how well playing skill transfers to agents they were not trained with. We also analyze patterns of communication between agents to elucidate how collaboration happens. A key finding is that while most agents only learn to play well with partners seen during training, one particular agent leads the Rainbow algorithm towards a much more general policy. The metrics and hypotheses advanced in this paper can be used for further study of collaborative agents.

    Original languageEnglish (US)
    Title of host publicationProceedings of the 16th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2020
    EditorsLevi Lelis, David Thue
    PublisherThe AAAI Press
    Pages31-37
    Number of pages7
    ISBN (Electronic)9781577358497
    StatePublished - 2020
    Event16th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2020 - Virtual, Online
    Duration: Oct 19 2020Oct 23 2020

    Publication series

    NameProceedings of the 16th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2020

    Conference

    Conference16th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2020
    CityVirtual, Online
    Period10/19/2010/23/20

    ASJC Scopus subject areas

    • Visual Arts and Performing Arts
    • Artificial Intelligence

    Fingerprint Dive into the research topics of 'Behavioral evaluation of hanabi rainbow DQN agents and rule-based agents'. Together they form a unique fingerprint.

    Cite this