Biasing MCTS with Features for General Games

Soemers, Dennis;Piette, Eric;Browne, Cameron
(2019) 2019 IEEE Congress on Evolutionary Computation (CEC) — Location: Wellington, New Zealand (10.June.2019)

Files

cec19.pdf
  • Open Access
  • Adobe PDF
  • 543.43 KB

Details

Authors
  • Soemers, Dennis
    Author
  • Piette, EricUCLouvain
    Author
  • Browne, Cameron
    Author
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.
Affiliations

Citations

Soemers, D., Piette, E., & Browne, C. (2019). Biasing MCTS with Features for General Games. 2019 IEEE Congress on Evolutionary Computation. Published. 2019 IEEE Congress on Evolutionary Computation (CEC), Wellington, New Zealand. https://doi.org/10.1109/cec.2019.8790141