New math framework auto-tunes AI robustness training

Original: A Unified and Constrained View of Regularization-Based Robust Reinforcement Learning

Source: arxiv.org ↗