Deep Synoptic Monte Carlo Planning in Reconnaissance Blind Chess

TL;DR

This paper presents DSMCP, a novel planning algorithm for large imperfect information games, demonstrated by its success in reconnaissance blind chess, and introduces new inference and bandit techniques.

Contribution

It introduces DSMCP, a new Monte Carlo planning method using synopses for uncertainty, and develops Penumbra, a program that won the 2020 reconnaissance blind chess competition.

Findings

Penumbra outperformed 33 competitors in the 2020 competition.

DSMCP effectively handles large imperfect information games.

Algorithm variants with caution, paranoia, and new bandit methods were evaluated.

Abstract

This paper introduces deep synoptic Monte Carlo planning (DSMCP) for large imperfect information games. The algorithm constructs a belief state with an unweighted particle filter and plans via playouts that start at samples drawn from the belief state. The algorithm accounts for uncertainty by performing inference on "synopses," a novel stochastic abstraction of information states. DSMCP is the basis of the program Penumbra, which won the official 2020 reconnaissance blind chess competition versus 33 other programs. This paper also evaluates algorithm variants that incorporate caution, paranoia, and a novel bandit algorithm. Furthermore, it audits the synopsis features used in Penumbra with per-bit saliency statistics.