Runtime Safety through Adaptive Shielding: From Hidden Parameter Inference to Provable Guarantees
This addresses safety-critical applications in robotics and autonomous systems by providing provable guarantees, though it is incremental as it builds on existing constrained hidden-parameter MDP frameworks.
The paper tackles safety risks from hidden parameters like robot mass in reinforcement learning by developing a runtime shielding mechanism that adapts online using function encoders and conformal prediction, proving probabilistic safety guarantees and showing significant reductions in safety violations with minimal overhead in experiments.
Variations in hidden parameters, such as a robot's mass distribution or friction, pose safety risks during execution. We develop a runtime shielding mechanism for reinforcement learning, building on the formalism of constrained hidden-parameter Markov decision processes. Function encoders enable real-time inference of hidden parameters from observations, allowing the shield and the underlying policy to adapt online. The shield constrains the action space by forecasting future safety risks (such as obstacle proximity) and accounts for uncertainty via conformal prediction. We prove that the proposed mechanism satisfies probabilistic safety guarantees and yields optimal policies among the set of safety-compliant policies. Experiments across diverse environments with varying hidden parameters show that our method significantly reduces safety violations and achieves strong out-of-distribution generalization, while incurring minimal runtime overhead.