DEP-RL: Embodied Exploration for Reinforcement Learning in Overactuated and Musculoskeletal Systems

TL;DR

This paper introduces DEP-RL, a reinforcement learning method that uses differential extrinsic plasticity to enable efficient exploration and rapid learning in complex, overactuated musculoskeletal systems, outperforming existing methods.

Contribution

The paper presents DEP-RL, a novel integration of differential extrinsic plasticity into RL, improving exploration and learning speed in large, overactuated musculoskeletal models.

Findings

DEP-RL achieves faster learning than existing methods.

DEP-RL demonstrates higher robustness in tasks.

Sample efficiency is significantly improved.

Abstract

Muscle-actuated organisms are capable of learning an unparalleled diversity of dexterous movements despite their vast amount of muscles. Reinforcement learning (RL) on large musculoskeletal models, however, has not been able to show similar performance. We conjecture that ineffective exploration in large overactuated action spaces is a key problem. This is supported by the finding that common exploration noise strategies are inadequate in synthetic examples of overactuated systems. We identify differential extrinsic plasticity (DEP), a method from the domain of self-organization, as being able to induce state-space covering exploration within seconds of interaction. By integrating DEP into RL, we achieve fast learning of reaching and locomotion in musculoskeletal systems, outperforming current approaches in all considered tasks in sample efficiency and robustness.