Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards
This addresses the challenge of efficient exploration in high-dimensional robotics control for researchers and practitioners, though it is incremental as it builds on existing DDPG methods.
The paper tackles the problem of deep reinforcement learning in robotics with sparse rewards by incorporating human demonstrations into DDPG, eliminating the need for engineered rewards and improving performance on simulated insertion tasks and a real-world clip insertion task.
We propose a general and model-free approach for Reinforcement Learning (RL) on real robotics with sparse rewards. We build upon the Deep Deterministic Policy Gradient (DDPG) algorithm to use demonstrations. Both demonstrations and actual interactions are used to fill a replay buffer and the sampling ratio between demonstrations and transitions is automatically tuned via a prioritized replay mechanism. Typically, carefully engineered shaping rewards are required to enable the agents to efficiently explore on high dimensional control problems such as robotics. They are also required for model-based acceleration methods relying on local solvers such as iLQG (e.g. Guided Policy Search and Normalized Advantage Function). The demonstrations replace the need for carefully engineered rewards, and reduce the exploration problem encountered by classical RL approaches in these domains. Demonstrations are collected by a robot kinesthetically force-controlled by a human demonstrator. Results on four simulated insertion tasks show that DDPG from demonstrations out-performs DDPG, and does not require engineered rewards. Finally, we demonstrate the method on a real robotics task consisting of inserting a clip (flexible object) into a rigid object.