Twice Sequential Monte Carlo for Tree Search

TL;DR

This paper introduces TSMCTS, a novel variant of Sequential Monte Carlo Tree Search, which improves scalability, reduces variance, and mitigates path degeneracy in model-based reinforcement learning.

Contribution

The paper proposes TSMCTS, enhancing SMC for tree search by addressing variance and path degeneracy, and demonstrating superior performance and scalability.

Findings

TSMCTS outperforms SMC and modern MCTS in various environments.

It scales better with increased search depth and compute.

It reduces estimator variance and mitigates path degeneracy.

Abstract

Model-based reinforcement learning (RL) methods that leverage search are responsible for many milestone breakthroughs in RL. Sequential Monte Carlo (SMC) recently emerged as an alternative to the Monte Carlo Tree Search (MCTS) algorithm which drove these breakthroughs. SMC is easier to parallelize and more suitable to GPU acceleration. However, it also suffers from large variance and path degeneracy which prevent it from scaling well with increased search depth, i.e., increased sequential compute. To address these problems, we introduce Twice Sequential Monte Carlo Tree Search (TSMCTS). Across discrete and continuous environments TSMCTS outperforms the SMC baseline as well as a popular modern version of MCTS as a policy improvement operator, scales favorably with sequential compute, reduces estimator variance and mitigates the effects of path degeneracy while retaining the properties that make SMC natural to parallelize.