Combining Reinforcement Learning and Tensor Networks, with an Application to Dynamical Large Deviations

TL;DR

This paper introduces ACTeN, a novel framework combining tensor networks with reinforcement learning to efficiently solve complex dynamical optimization problems, demonstrated on challenging stochastic models.

Contribution

The paper presents a new actor-critic method using tensor networks for policy and value function approximation, suitable for large, factorisable state and action spaces.

Findings

Successfully sampled rare trajectories in the East model and ASEP.

Demonstrated effectiveness of tensor networks in complex RL tasks.

Potential for broad applications in physics and multi-agent RL.

Abstract

We present a framework to integrate tensor network (TN) methods with reinforcement learning (RL) for solving dynamical optimisation tasks. We consider the RL actor-critic method, a model-free approach for solving RL problems, and introduce TNs as the approximators for its policy and value functions. Our "actor-critic with tensor networks" (ACTeN) method is especially well suited to problems with large and factorisable state and action spaces. As an illustration of the applicability of ACTeN we solve the exponentially hard task of sampling rare trajectories in two paradigmatic stochastic models, the East model of glasses and the asymmetric simple exclusion process (ASEP), the latter being particularly challenging to other methods due to the absence of detailed balance. With substantial potential for further integration with the vast array of existing RL methods, the approach introduced here is promising both for applications in physics and to multi-agent RL problems more generally.