ACDER: Augmented Curiosity-Driven Experience Replay
This addresses the challenge of sample efficiency in robotic manipulation for RL researchers, though it is incremental as it builds on Hindsight Experience Replay.
The paper tackles the problem of inefficient exploration in reinforcement learning for robotic manipulation tasks with sparse rewards by proposing ACDER, which combines goal-oriented curiosity-driven exploration and dynamic initial states selection. The method significantly outperforms existing methods on three basic tasks and achieves satisfactory performance on a multi-step task.
Exploration in environments with sparse feedback remains a challenging research problem in reinforcement learning (RL). When the RL agent explores the environment randomly, it results in low exploration efficiency, especially in robotic manipulation tasks with high dimensional continuous state and action space. In this paper, we propose a novel method, called Augmented Curiosity-Driven Experience Replay (ACDER), which leverages (i) a new goal-oriented curiosity-driven exploration to encourage the agent to pursue novel and task-relevant states more purposefully and (ii) the dynamic initial states selection as an automatic exploratory curriculum to further improve the sample-efficiency. Our approach complements Hindsight Experience Replay (HER) by introducing a new way to pursue valuable states. Experiments conducted on four challenging robotic manipulation tasks with binary rewards, including Reach, Push, Pick&Place and Multi-step Push. The empirical results show that our proposed method significantly outperforms existing methods in the first three basic tasks and also achieves satisfactory performance in multi-step robotic task learning.