SAFER: Data-Efficient and Safe Reinforcement Learning via Skill Acquisition
This work addresses the problem of enabling safe and data-efficient policy learning for safety-critical robotic applications, representing an incremental improvement over existing primitive learning techniques.
The paper tackles the problem of safe reinforcement learning by addressing the limitation of existing primitive learning methods that ignore negative experiences, proposing SAFER which models safety context using contrastive training on offline datasets including both negative and positive demonstrations. The result shows that SAFER outperforms state-of-the-art methods in success and safety on complex robotic grasping tasks.
Methods that extract policy primitives from offline demonstrations using deep generative models have shown promise at accelerating reinforcement learning(RL) for new tasks. Intuitively, these methods should also help to trainsafeRLagents because they enforce useful skills. However, we identify these techniques are not well equipped for safe policy learning because they ignore negative experiences(e.g., unsafe or unsuccessful), focusing only on positive experiences, which harms their ability to generalize to new tasks safely. Rather, we model the latentsafetycontextusing principled contrastive training on an offline dataset of demonstrations from many tasks, including both negative and positive experiences. Using this late variable, our RL framework, SAFEty skill pRiors (SAFER) extracts task-specific safe primitive skills to safely and successfully generalize to new tasks. In the inference stage, policies trained with SAFER learn to compose safe skills into successful policies. We theoretically characterize why SAFER can enforce safe policy learning and demonstrate its effectiveness on several complex safety-critical robotic grasping tasks inspired by the game Operation, in which SAFERoutperforms state-of-the-art primitive learning methods in success and safety.