Safe and Robust Experience Sharing for Deterministic Policy Gradient Algorithms
This addresses memory limitations in off-policy deep reinforcement learning for continuous action domains, though it appears to be an incremental improvement on existing methods.
The paper tackles the challenge of learning in high-dimensional continuous control tasks with limited experience replay memory by introducing an experience sharing mechanism for deterministic policies. The method achieves safe experience sharing across multiple agents and shows robust performance under strict memory constraints in OpenAI Gym continuous control tasks.
Learning in high dimensional continuous tasks is challenging, mainly when the experience replay memory is very limited. We introduce a simple yet effective experience sharing mechanism for deterministic policies in continuous action domains for the future off-policy deep reinforcement learning applications in which the allocated memory for the experience replay buffer is limited. To overcome the extrapolation error induced by learning from other agents' experiences, we facilitate our algorithm with a novel off-policy correction technique without any action probability estimates. We test the effectiveness of our method in challenging OpenAI Gym continuous control tasks and conclude that it can achieve a safe experience sharing across multiple agents and exhibits a robust performance when the replay memory is strictly limited.