Beyond Optimism: Exploration With Partially Observable Rewards

TL;DR

This paper introduces a new exploration strategy for reinforcement learning that guarantees convergence to optimal policies even with partially observable rewards, addressing limitations of optimism-based methods.

Contribution

A novel exploration approach that handles unobservable rewards and a set of benchmark environments for evaluating exploration strategies in RL.

Findings

Our method outperforms existing exploration algorithms in benchmark environments.

The proposed strategy guarantees convergence to optimal policies under partial observability.

Benchmark environments facilitate evaluation of exploration in RL with unobservable rewards.

Abstract

Exploration in reinforcement learning (RL) remains an open challenge. RL algorithms rely on observing rewards to train the agent, and if informative rewards are sparse the agent learns slowly or may not learn at all. To improve exploration and reward discovery, popular algorithms rely on optimism. But what if sometimes rewards are unobservable, e.g., situations of partial monitoring in bandits and the recent formalism of monitored Markov decision process? In this case, optimism can lead to suboptimal behavior that does not explore further to collapse uncertainty. With this paper, we present a novel exploration strategy that overcomes the limitations of existing methods and guarantees convergence to an optimal policy even when rewards are not always observable. We further propose a collection of tabular environments for benchmarking exploration in RL (with and without unobservable rewards) and show that our method outperforms existing ones.