A Bayesian Tactician

Abstract : We describe a generative Bayesian model of tactical attacks in strategy games, which can be used both to predict attacks and to take tactical decisions. This model is designed to easily integrate and merge information from other (probabilistic) estimations and heuristics. In particular, it handles uncertainty in enemy units' positions as well as their probable tech tree. We claim that learning, being it supervised or through reinforcement, adapts to skewed data sources. We evaluated our approach on StarCraft1: the parameters are learned on a new (freely available) dataset of game states, deterministically re-created from replays, and the whole model is evaluated for prediction in realistic conditions. It is also the tactical decision-making component of a competitive StarCraft AI.
Document type :
Conference papers
Computer Games Workshop at ECAI 2012, Aug 2012, Montpellier, France. pp. 114-125, 2012


https://hal.archives-ouvertes.fr/hal-00752868
Contributor : Gabriel Synnaeve <>
Submitted on : Friday, November 16, 2012 - 3:31:50 PM
Last modification on : Tuesday, October 28, 2014 - 6:35:33 PM

File

BayesianTactician.pdf
fileSource_public_author

Identifiers

  • HAL Id : hal-00752868, version 1

Collections

Citation

Gabriel Synnaeve, Pierre Bessiere. A Bayesian Tactician. Computer Games Workshop at ECAI 2012, Aug 2012, Montpellier, France. pp. 114-125, 2012. <hal-00752868>

Export

Share

Metrics

Consultation de
la notice

469

Téléchargement du document

200