Biasing MCTS with Features for General Games

Dennis Soemers*, Eric Piette, Cameron Browne

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingAcademicpeer-review

137 Downloads (Pure)

Abstract

This paper proposes using a linear function approximator, rather than a deep neural network (DNN), to bias a Monte Carlo tree search (MCTS) player for general games. This is unlikely to match the potential raw playing strength of DNNs, but has advantages in terms of generality, interpretability and resources (time and hardware) required for training. Features describing local patterns are used as inputs. The features are formulated in such a way that they are easily interpretable and applicable to a wide range of general games, and might encode simple local strategies. We gradually create new features during the same self-play training process used to learn feature weights. We evaluate the playing strength of an MCTS player biased by learnt features against a standard upper confidence bounds for trees (UCT) player in multiple different board games, and demonstrate significantly improved playing strength in the majority of them after a small number of self-play training games.
Original languageEnglish
Title of host publicationIEEE Congress on Evolutionary Computation
Subtitle of host publication(CEC'19)
Pages450-457
Number of pages8
DOIs
Publication statusPublished - 11 Jun 2019

Keywords

  • GO
  • features
  • games
  • learning
  • search

Cite this