Memory-Augmented Monte Carlo Tree Search

Chenjun Xiao, Jincheng Mei, Martin Müller

AAAI 2018 pp. 1455-1462

doi:10.1609/AAAI.V32I1.11531 /aaai/2018/xiao2018aaai-memory/

Abstract

This paper proposes and evaluates Memory-Augmented Monte Carlo Tree Search (M-MCTS), which provides a new approach to exploit generalization in online real-time search. The key idea of M-MCTS is to incorporate MCTS with a memory structure, where each entry contains information of a particular state. This memory is used to generate an approximate value estimation by combining the estimations of similar states. We show that the memory based value approximation is better than the vanilla Monte Carlo estimation with high probability under mild conditions. We evaluate M-MCTS in the game of Go. Experimental results show that M-MCTS outperforms the original MCTS with the same number of simulations.

PDF AAAI Semantic Scholar

Cite

Text

Xiao et al. "Memory-Augmented Monte Carlo Tree Search." AAAI Conference on Artificial Intelligence, 2018. doi:10.1609/AAAI.V32I1.11531

Markdown

[Xiao et al. "Memory-Augmented Monte Carlo Tree Search." AAAI Conference on Artificial Intelligence, 2018.](https://mlanthology.org/aaai/2018/xiao2018aaai-memory/) doi:10.1609/AAAI.V32I1.11531

BibTeX

@inproceedings{xiao2018aaai-memory,
  title     = {{Memory-Augmented Monte Carlo Tree Search}},
  author    = {Xiao, Chenjun and Mei, Jincheng and Müller, Martin},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2018},
  pages     = {1455-1462},
  doi       = {10.1609/AAAI.V32I1.11531},
  url       = {https://mlanthology.org/aaai/2018/xiao2018aaai-memory/}
}