Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation
Shangding Gu, Bilgehan Sel, Yuhao Ding, Lu Wang, Qingwei Lin, Ming Jin, Alois Knoll
Abstract
Ensuring the safety of Reinforcement Learning (RL) is crucial for its deployment in real-world applications. Nevertheless, managing the trade-off between reward and safety during exploration presents a significant challenge. Improving reward performance through policy adjustments may adversely affect safety performance. In this study, we aim to address this conflicting relation by leveraging the theory of gradient manipulation. Initially, we analyze the conflict between reward and safety gradients. Subsequently, we tackle the balance between reward and safety optimization by proposing a soft switching policy optimization method, for which we provide convergence analysis. Based on our theoretical examination, we provide a safe RL framework to overcome the aforementioned challenge, and we develop a Safety-MuJoCo Benchmark to assess the performance of safe RL algorithms. Finally, we evaluate the effectiveness of our method on the Safety-MuJoCo Benchmark and a popular safe benchmark, Omnisafe. Experimental results demonstrate that our algorithms outperform several state-of-the-art baselines in terms of balancing reward and safety optimization.
BibTeX
@article{Gu_Sel_Ding_Wang_Lin_Jin_Knoll_2024, title={Balance Reward and Safety Optimization for Safe Reinforcement Learning: A Perspective of Gradient Manipulation}, volume={38}, url={https://ojs.aaai.org/index.php/AAAI/article/view/30102}, DOI={10.1609/aaai.v38i19.30102}, abstractNote={Ensuring the safety of Reinforcement Learning (RL) is crucial for its deployment in real-world applications. Nevertheless, managing the trade-off between reward and safety during exploration presents a significant challenge. Improving reward performance through policy adjustments may adversely affect safety performance. In this study, we aim to address this conflicting relation by leveraging the theory of gradient manipulation. Initially, we analyze the conflict between reward and safety gradients. Subsequently, we tackle the balance between reward and safety optimization by proposing a soft switching policy optimization method, for which we provide convergence analysis. Based on our theoretical examination, we provide a safe RL framework to overcome the aforementioned challenge, and we develop a Safety-MuJoCo Benchmark to assess the performance of safe RL algorithms. Finally, we evaluate the effectiveness of our method on the Safety-MuJoCo Benchmark and a popular safe benchmark, Omnisafe. Experimental results demonstrate that our algorithms outperform several state-of-the-art baselines in terms of balancing reward and safety optimization.}, number={19}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Gu, Shangding and Sel, Bilgehan and Ding, Yuhao and Wang, Lu and Lin, Qingwei and Jin, Ming and Knoll, Alois}, year={2024}, month={Mar.}, pages={21099-21106} }