State-Dependent Conformal Perception Bounds for Neuro-Symbolic Verification of Autonomous Systems

Thomas Waite, Yuang Geng, Trevor Turnquist, Ivan Ruchkin, Radoslav Ivanov
Proceedings of the International Conference on Neuro-symbolic Systems, PMLR 288:127-143, 2025.

Abstract

It remains a challenge to provide safety guarantees for autonomous systems with neural perception and control. A typical approach obtains symbolic bounds on perception error (e.g., using conformal prediction) and performs verification under these bounds. However, these bounds can lead to drastic conservatism in the resulting end-to-end safety guarantee. This paper proposes an approach to synthesize symbolic perception error bounds that serve as an optimal interface between perception performance and control verification. The key idea is to consider our error bounds to be heteroskedastic with respect to the system’s state — not time like in previous approaches. These bounds can be obtained with two gradient-free optimization algorithms. We demonstrate that our bounds lead to tighter safety guarantees than the state-of-the-art in a case study on a mountain car.

Cite this Paper

BibTeX
@InProceedings{pmlr-v288-waite25a, title = {State-Dependent Conformal Perception Bounds for Neuro-Symbolic Verification of Autonomous Systems}, author = {Waite, Thomas and Geng, Yuang and Turnquist, Trevor and Ruchkin, Ivan and Ivanov, Radoslav}, booktitle = {Proceedings of the International Conference on Neuro-symbolic Systems}, pages = {127--143}, year = {2025}, editor = {Pappas, George and Ravikumar, Pradeep and Seshia, Sanjit A.}, volume = {288}, series = {Proceedings of Machine Learning Research}, month = {28--30 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v288/main/assets/waite25a/waite25a.pdf}, url = {https://proceedings.mlr.press/v288/waite25a.html}, abstract = {It remains a challenge to provide safety guarantees for autonomous systems with neural perception and control. A typical approach obtains symbolic bounds on perception error (e.g., using conformal prediction) and performs verification under these bounds. However, these bounds can lead to drastic conservatism in the resulting end-to-end safety guarantee. This paper proposes an approach to synthesize symbolic perception error bounds that serve as an optimal interface between perception performance and control verification. The key idea is to consider our error bounds to be heteroskedastic with respect to the system’s state — not time like in previous approaches. These bounds can be obtained with two gradient-free optimization algorithms. We demonstrate that our bounds lead to tighter safety guarantees than the state-of-the-art in a case study on a mountain car.} }
Endnote
%0 Conference Paper %T State-Dependent Conformal Perception Bounds for Neuro-Symbolic Verification of Autonomous Systems %A Thomas Waite %A Yuang Geng %A Trevor Turnquist %A Ivan Ruchkin %A Radoslav Ivanov %B Proceedings of the International Conference on Neuro-symbolic Systems %C Proceedings of Machine Learning Research %D 2025 %E George Pappas %E Pradeep Ravikumar %E Sanjit A. Seshia %F pmlr-v288-waite25a %I PMLR %P 127--143 %U https://proceedings.mlr.press/v288/waite25a.html %V 288 %X It remains a challenge to provide safety guarantees for autonomous systems with neural perception and control. A typical approach obtains symbolic bounds on perception error (e.g., using conformal prediction) and performs verification under these bounds. However, these bounds can lead to drastic conservatism in the resulting end-to-end safety guarantee. This paper proposes an approach to synthesize symbolic perception error bounds that serve as an optimal interface between perception performance and control verification. The key idea is to consider our error bounds to be heteroskedastic with respect to the system’s state — not time like in previous approaches. These bounds can be obtained with two gradient-free optimization algorithms. We demonstrate that our bounds lead to tighter safety guarantees than the state-of-the-art in a case study on a mountain car.
APA
Waite, T., Geng, Y., Turnquist, T., Ruchkin, I. & Ivanov, R.. (2025). State-Dependent Conformal Perception Bounds for Neuro-Symbolic Verification of Autonomous Systems. Proceedings of the International Conference on Neuro-symbolic Systems, in Proceedings of Machine Learning Research 288:127-143 Available from https://proceedings.mlr.press/v288/waite25a.html.

Related Material