Safe Control using Learned Safety Filters and Adaptive Conformal Inference

2026-04-20

AI Summary

This work addresses the challenge of providing reliable safety guarantees in high-dimensional systems, where learning-based safety filters often fail due to prediction errors. To overcome this limitation, the paper proposes Adaptive Conformal Filtering (ACoFi), a novel framework that integrates Hamilton-Jacobi reachability analysis with adaptive conformal inference to dynamically adjust the policy switching threshold in response to prediction uncertainty. ACoFi offers an asymptotically valid, user-specified upper bound on the miscoverage rate of safety predictions, thereby delivering a soft yet theoretically grounded safety guarantee. Experimental evaluations on Dubins car and Safety Gymnasium benchmarks demonstrate that ACoFi significantly outperforms fixed-threshold baselines, achieving higher safety performance and fewer constraint violations—particularly in out-of-distribution scenarios.

Technology Category

• Natural Language Processing: Safety and Robustness
• Reasoning under Uncertainty: Sequential Decision Making
• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Safety filters have been shown to be effective tools to ensure the safety of control systems with unsafe nominal policies. To address scalability challenges in traditional synthesis methods, learning-based approaches have been proposed for designing safety filters for systems with high-dimensional state and control spaces. However, the inevitable errors in the decisions of these models raise concerns about their reliability and the safety guarantees they offer. This paper presents Adaptive Conformal Filtering (ACoFi), a method that combines learned Hamilton-Jacobi reachability-based safety filters with adaptive conformal inference. Under ACoFi, the filter dynamically adjusts its switching criteria based on the observed errors in its predictions of the safety of actions. The range of possible safety values of the nominal policy's output is used to quantify uncertainty in safety assessment. The filter switches from the nominal policy to the learned safe one when that range suggests it might be unsafe. We show that ACoFi guarantees that the rate of incorrectly quantifying uncertainty in the predicted safety of the nominal policy is asymptotically upper bounded by a user-defined parameter. This gives a soft safety guarantee rather than a hard safety guarantee. We evaluate ACoFi in a Dubins car simulation and a Safety Gymnasium environment, empirically demonstrating that it significantly outperforms the baseline method that uses a fixed switching threshold by achieving higher learned safety values and fewer safety violations, especially in out-of-distribution scenarios.

Problem

• safe control
• safety filters
• conformal inference
• uncertainty quantification
• learning-based safety

Innovation

• Adaptive Conformal Inference
• Safety Filters
• Hamilton-Jacobi Reachability
• Uncertainty Quantification
• Safe Control