Jointly-Learned State-Action Embedding for Efficient Reinforcement Learning

TL;DR

This paper introduces a novel joint embedding method for states and actions in reinforcement learning, improving generalization and performance in large, complex domains by combining model-free and model-based approaches.

Contribution

It presents the first theoretical foundation for joint state-action embeddings and a practical architecture that leverages environment models for better reinforcement learning performance.

Findings

Outperforms state-of-the-art models in gaming, robotic control, and recommender systems.

Effective in both discrete and continuous large state/action spaces.

Enhances generalization by capturing similarities in embedding spaces.

Abstract

While reinforcement learning has achieved considerable successes in recent years, state-of-the-art models are often still limited by the size of state and action spaces. Model-free reinforcement learning approaches use some form of state representations and the latest work has explored embedding techniques for actions, both with the aim of achieving better generalization and applicability. However, these approaches consider only states or actions, ignoring the interaction between them when generating embedded representations. In this work, we establish the theoretical foundations for the validity of training a reinforcement learning agent using embedded states and actions. We then propose a new approach for jointly learning embeddings for states and actions that combines aspects of model-free and model-based reinforcement learning, which can be applied in both discrete and continuous domains. Specifically, we use a model of the environment to obtain embeddings for states and actions and present a generic architecture that leverages these to learn a policy. In this way, the embedded representations obtained via our approach enable better generalization over both states and actions by capturing similarities in the embedding spaces. Evaluations of our approach on several gaming, robotic control, and recommender systems show it significantly outperforms state-of-the-art models in both discrete/continuous domains with large state/action spaces, thus confirming its efficacy.