Contextual Policy Transfer in Reinforcement Learning Domains via Deep Mixtures-of-Experts

TL;DR

This paper introduces a deep mixture-of-experts approach for context-aware transfer in reinforcement learning, enabling effective knowledge transfer from model-based sources to model-free agents in diverse environments.

Contribution

It presents a novel deep mixture-of-experts model for state-dependent transfer of source policies based on estimated dynamics, improving robustness and interpretability.

Findings

Effective transfer in OpenAI-Gym benchmarks

Robustness to dynamics estimation errors

Compatibility with standard RL algorithms

Abstract

In reinforcement learning, agents that consider the context, or current state, when selecting source policies for transfer have been shown to outperform context-free approaches. However, none of the existing approaches transfer knowledge contextually from model-based learners to a model-free learner. This could be useful, for instance, when source policies are intentionally learned on diverse simulations with plentiful data but transferred to a real-world setting with limited data. In this paper, we assume knowledge of estimated source task dynamics and policies, and common sub-goals but different dynamics. We introduce a novel deep mixture-of-experts formulation for learning state-dependent beliefs over source task dynamics that match the target dynamics using state trajectories collected from the target task. The mixture model is easy to interpret, demonstrates robustness to estimation errors in dynamics, and is compatible with most learning algorithms. We then show how this model can be incorporated into standard policy reuse frameworks, and demonstrate its effectiveness on benchmarks from OpenAI-Gym.