Reinforcement Learning with Prototypical Representations
This addresses the problem of sample inefficiency in reinforcement learning for researchers and practitioners, though it appears incremental as it builds on existing representation learning methods.
The paper tackles the challenge of sample-efficient reinforcement learning in image-based environments by proposing Proto-RL, a self-supervised framework that uses prototypical representations to tie representation learning with exploration, achieving state-of-the-art downstream policy learning on difficult continuous control tasks.
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.