Skill-based Safe Reinforcement Learning with Risk Planning

TL;DR

This paper introduces a novel Safe Skill Planning approach that leverages offline demonstration data and risk prediction to improve safety and efficiency in reinforcement learning for robotics.

Contribution

The paper proposes a two-stage Safe Skill Planning method combining PU learning and risk-aware planning to enhance safe RL with offline data and online adaptation.

Findings

Outperforms previous safe RL methods in benchmark environments

Effectively integrates offline demonstration data for risk prediction

Achieves safer and more efficient policy learning

Abstract

Safe Reinforcement Learning (Safe RL) aims to ensure safety when an RL agent conducts learning by interacting with real-world environments where improper actions can induce high costs or lead to severe consequences. In this paper, we propose a novel Safe Skill Planning (SSkP) approach to enhance effective safe RL by exploiting auxiliary offline demonstration data. SSkP involves a two-stage process. First, we employ PU learning to learn a skill risk predictor from the offline demonstration data. Then, based on the learned skill risk predictor, we develop a novel risk planning process to enhance online safe RL and learn a risk-averse safe policy efficiently through interactions with the online RL environment, while simultaneously adapting the skill risk predictor to the environment. We conduct experiments in several benchmark robotic simulation environments. The experimental results demonstrate that the proposed approach consistently outperforms previous state-of-the-art safe RL methods.