Program-Based Strategy Induction for Reinforcement Learning

TL;DR

This paper introduces a Bayesian program induction approach to discover interpretable, discrete strategies in reinforcement learning tasks, revealing complex behaviors that traditional models overlook.

Contribution

It presents a novel method for strategy discovery that balances simplicity and effectiveness, capturing strategies that are hard to identify with classical models.

Findings

Identified strategies like asymmetric learning and adaptive exploration.

Revealed discrete state switching in bandit tasks.

Demonstrated interpretability of discovered strategies.

Abstract

Typical models of learning assume incremental estimation of continuously-varying decision variables like expected rewards. However, this class of models fails to capture more idiosyncratic, discrete heuristics and strategies that people and animals appear to exhibit. Despite recent advances in strategy discovery using tools like recurrent networks that generalize the classic models, the resulting strategies are often onerous to interpret, making connections to cognition difficult to establish. We use Bayesian program induction to discover strategies implemented by programs, letting the simplicity of strategies trade off against their effectiveness. Focusing on bandit tasks, we find strategies that are difficult or unexpected with classical incremental learning, like asymmetric learning from rewarded and unrewarded trials, adaptive horizon-dependent random exploration, and discrete state switching.