Learning latent state representation for speeding up exploration

TL;DR

This paper proposes a method that leverages prior experience to learn latent state representations, improving exploration efficiency in high-dimensional reinforcement learning tasks with sparse rewards.

Contribution

It introduces a representation learning approach that uses prior related task data to enhance exploration in high-dimensional spaces.

Findings

Learned representations predict rewards effectively.

Using representations improves exploration in high-dimensional environments.

Method benefits meta-exploration in simulated tasks.

Abstract

Exploration is an extremely challenging problem in reinforcement learning, especially in high dimensional state and action spaces and when only sparse rewards are available. Effective representations can indicate which components of the state are task relevant and thus reduce the dimensionality of the space to explore. In this work, we take a representation learning viewpoint on exploration, utilizing prior experience to learn effective latent representations, which can subsequently indicate which regions to explore. Prior experience on separate but related tasks help learn representations of the state which are effective at predicting instantaneous rewards. These learned representations can then be used with an entropy-based exploration method to effectively perform exploration in high dimensional spaces by effectively lowering the dimensionality of the search space. We show the benefits of this representation for meta-exploration in a simulated object pushing environment.