Learning Unknown Constraints without Unsafe Data via Optimality and Counterfactual Regularization

This research proposes CF-KKT, a constraint learning framework that leverages learned dynamics and locally optimal demonstrations to recover unknown constraints without requiring known dynamics or risky exploration. The method combines the data efficiency and safety of CIOC with the flexibility of ICRL, achieving improved safety and data efficiency in high-dimensional robotic control tasks compared to state-of-the-art offline ICRL baselines.