Work

Safe Reinforcement Learning Under High Relative Degree Constraints for Discrete‐Time Fully Cooperative Games