Learning a subspace of policies for online adaptation in Reinforcement Learning

TL;DR

This paper introduces a method for online adaptation in reinforcement learning by learning a subspace of policies, enabling better generalization to unseen environment variations without extensive tuning or additional components.

Contribution

The authors propose a novel subspace policy learning approach that improves generalization and online adaptation in RL without relying on meta-RL or extra modules.

Findings

Outperforms baseline methods on various benchmarks.

Learns policies that achieve high rewards in unseen environments.

Simple to tune and does not require extra components.

Abstract

Deep Reinforcement Learning (RL) is mainly studied in a setting where the training and the testing environments are similar. But in many practical applications, these environments may differ. For instance, in control systems, the robot(s) on which a policy is learned might differ from the robot(s) on which a policy will run. It can be caused by different internal factors (e.g., calibration issues, system attrition, defective modules) or also by external changes (e.g., weather conditions). There is a need to develop RL methods that generalize well to variations of the training conditions. In this article, we consider the simplest yet hard to tackle generalization setting where the test environment is unknown at train time, forcing the agent to adapt to the system's new dynamics. This online adaptation process can be computationally expensive (e.g., fine-tuning) and cannot rely on meta-RL techniques since there is just a single train environment. To do so, we propose an approach where we learn a subspace of policies within the parameter space. This subspace contains an infinite number of policies that are trained to solve the training environment while having different parameter values. As a consequence, two policies in that subspace process information differently and exhibit different behaviors when facing variations of the train environment. Our experiments carried out over a large variety of benchmarks compare our approach with baselines, including diversity-based methods. In comparison, our approach is simple to tune, does not need any extra component (e.g., discriminator) and learns policies able to gather a high reward on unseen environments.