Reinforcement Learning with Prototypical Representations

TL;DR

Proto-RL introduces a self-supervised approach that learns prototypical representations to improve exploration and generalization in reinforcement learning, achieving state-of-the-art results on complex continuous control tasks.

Contribution

The paper presents Proto-RL, a novel framework that jointly learns representations and exploration strategies using prototypes, decoupling representation learning from task-specific data.

Findings

Achieves state-of-the-art performance on continuous control benchmarks.

Pre-trained representations enable efficient downstream policy learning.

Prototypes effectively summarize exploratory experience.

Abstract

Learning effective representations in image-based environments is crucial for sample efficient Reinforcement Learning (RL). Unfortunately, in RL, representation learning is confounded with the exploratory experience of the agent -- learning a useful representation requires diverse data, while effective exploration is only possible with coherent representations. Furthermore, we would like to learn representations that not only generalize across tasks but also accelerate downstream exploration for efficient task-specific training. To address these challenges we propose Proto-RL, a self-supervised framework that ties representation learning with exploration through prototypical representations. These prototypes simultaneously serve as a summarization of the exploratory experience of an agent as well as a basis for representing observations. We pre-train these task-agnostic representations and prototypes on environments without downstream task information. This enables state-of-the-art downstream policy learning on a set of difficult continuous control tasks.