Predicting resource locations in game maps using deep convolutional neural networks

Scott Lee, Aaron Isaksen, Christoffer Holmgard, Julian Togelius

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

    Abstract

    We describe an application of neural networks to predict the placements of resources in StarCraft II maps. Networks are trained on existing maps taken from databases of maps actively used in online competitions and tested on unseen maps with resources (minerals and vespene gas) removed. This method is potentially useful for Al-Assisted game design tools, allowing the suggestion of resource and base placements consonant with implicit StarCraft II design principles for fully or partially sketched heightmaps. By varying the thresholds for the placement of resources, more or fewer resources can be created consistently with the pattern of a single map. We further propose that these networks can be used to help understand the design principles of StarCraft II maps, and by extension other, similar types of game content.

    Original languageEnglish (US)
    Title of host publicationWS-16-21
    Subtitle of host publicationArtificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics
    PublisherAI Access Foundation
    Pages46-52
    Number of pages7
    VolumeWS-16-21 - WS-16-23
    ISBN (Electronic)9781577357735
    StatePublished - 2016
    Event12th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2016 - Burlingame, United States
    Duration: Oct 8 2016Oct 9 2016

    Other

    Other12th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2016
    CountryUnited States
    CityBurlingame
    Period10/8/1610/9/16

    Fingerprint

    Neural networks
    Mineral resources
    Gases

    ASJC Scopus subject areas

    • Engineering(all)

    Cite this

    Lee, S., Isaksen, A., Holmgard, C., & Togelius, J. (2016). Predicting resource locations in game maps using deep convolutional neural networks. In WS-16-21: Artificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics (Vol. WS-16-21 - WS-16-23, pp. 46-52). AI Access Foundation.

    Predicting resource locations in game maps using deep convolutional neural networks. / Lee, Scott; Isaksen, Aaron; Holmgard, Christoffer; Togelius, Julian.

    WS-16-21: Artificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics. Vol. WS-16-21 - WS-16-23 AI Access Foundation, 2016. p. 46-52.

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

    Lee, S, Isaksen, A, Holmgard, C & Togelius, J 2016, Predicting resource locations in game maps using deep convolutional neural networks. in WS-16-21: Artificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics. vol. WS-16-21 - WS-16-23, AI Access Foundation, pp. 46-52, 12th AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, AIIDE 2016, Burlingame, United States, 10/8/16.
    Lee S, Isaksen A, Holmgard C, Togelius J. Predicting resource locations in game maps using deep convolutional neural networks. In WS-16-21: Artificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics. Vol. WS-16-21 - WS-16-23. AI Access Foundation. 2016. p. 46-52
    Lee, Scott ; Isaksen, Aaron ; Holmgard, Christoffer ; Togelius, Julian. / Predicting resource locations in game maps using deep convolutional neural networks. WS-16-21: Artificial Intelligence in Adversarial Games; WS-16-22: Experimental AI in Games; WS-16-23: Player Analytics. Vol. WS-16-21 - WS-16-23 AI Access Foundation, 2016. pp. 46-52
    @inproceedings{b3fe1c8f4829439da630feb05a6ec965,
    title = "Predicting resource locations in game maps using deep convolutional neural networks",
    abstract = "We describe an application of neural networks to predict the placements of resources in StarCraft II maps. Networks are trained on existing maps taken from databases of maps actively used in online competitions and tested on unseen maps with resources (minerals and vespene gas) removed. This method is potentially useful for Al-Assisted game design tools, allowing the suggestion of resource and base placements consonant with implicit StarCraft II design principles for fully or partially sketched heightmaps. By varying the thresholds for the placement of resources, more or fewer resources can be created consistently with the pattern of a single map. We further propose that these networks can be used to help understand the design principles of StarCraft II maps, and by extension other, similar types of game content.",
    author = "Scott Lee and Aaron Isaksen and Christoffer Holmgard and Julian Togelius",
    year = "2016",
    language = "English (US)",
    volume = "WS-16-21 - WS-16-23",
    pages = "46--52",
    booktitle = "WS-16-21",
    publisher = "AI Access Foundation",

    }

    TY - GEN

    T1 - Predicting resource locations in game maps using deep convolutional neural networks

    AU - Lee, Scott

    AU - Isaksen, Aaron

    AU - Holmgard, Christoffer

    AU - Togelius, Julian

    PY - 2016

    Y1 - 2016

    N2 - We describe an application of neural networks to predict the placements of resources in StarCraft II maps. Networks are trained on existing maps taken from databases of maps actively used in online competitions and tested on unseen maps with resources (minerals and vespene gas) removed. This method is potentially useful for Al-Assisted game design tools, allowing the suggestion of resource and base placements consonant with implicit StarCraft II design principles for fully or partially sketched heightmaps. By varying the thresholds for the placement of resources, more or fewer resources can be created consistently with the pattern of a single map. We further propose that these networks can be used to help understand the design principles of StarCraft II maps, and by extension other, similar types of game content.

    AB - We describe an application of neural networks to predict the placements of resources in StarCraft II maps. Networks are trained on existing maps taken from databases of maps actively used in online competitions and tested on unseen maps with resources (minerals and vespene gas) removed. This method is potentially useful for Al-Assisted game design tools, allowing the suggestion of resource and base placements consonant with implicit StarCraft II design principles for fully or partially sketched heightmaps. By varying the thresholds for the placement of resources, more or fewer resources can be created consistently with the pattern of a single map. We further propose that these networks can be used to help understand the design principles of StarCraft II maps, and by extension other, similar types of game content.

    UR - http://www.scopus.com/inward/record.url?scp=85021974049&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85021974049&partnerID=8YFLogxK

    M3 - Conference contribution

    AN - SCOPUS:85021974049

    VL - WS-16-21 - WS-16-23

    SP - 46

    EP - 52

    BT - WS-16-21

    PB - AI Access Foundation

    ER -