---
OA_place: repository
OA_type: green
_id: '19668'
abstract:
- lang: eng
  text: Learning-based methods provide a promising approach to solving highly non-linear
    control tasks that are often challenging for classical control methods. To ensure
    the satisfaction of a safety property, learning-based methods jointly learn a
    control policy together with a certificate function for the property. Popular
    examples include barrier functions for safety and Lyapunov functions for asymptotic
    stability. While there has been significant progress on learning-based control
    with certificate functions in the white-box setting, where the correctness of
    the certificate function can be formally verified, there has been little work
    on ensuring their reliability in the black-box setting where the system dynamics
    are unknown. In this work, we consider the problems of certifying and repairing
    neural network control policies and certificate functions in the black-box setting.
    We propose a novel framework that utilizes runtime monitoring to detect system
    behaviors that violate the property of interest under some initially trained neural
    network policy and certificate. These violating behaviors are used to extract
    new training data, that is used to re-train the neural network policy and the
    certificate function and to ultimately repair them. We demonstrate the effectiveness
    of our approach empirically by using it to repair and to boost the safety rate
    of neural network policies learned by a state-of-the-art method for learning-based
    control on two autonomous system control tasks.
acknowledgement: This work was supported in part by the ERC project ERC2020-AdG 101020093
article_processing_charge: No
arxiv: 1
author:
- first_name: Zhengqi
  full_name: Yu, Zhengqi
  id: 20aa2ae8-f2f1-11ed-bbfa-8205053f1342
  last_name: Yu
- first_name: Dorde
  full_name: Zikelic, Dorde
  id: 294AA7A6-F248-11E8-B48F-1D18A9856A87
  last_name: Zikelic
  orcid: 0000-0002-4681-1699
- first_name: Thomas A
  full_name: Henzinger, Thomas A
  id: 40876CD8-F248-11E8-B48F-1D18A9856A87
  last_name: Henzinger
  orcid: 0000-0002-2985-7724
citation:
  ama: 'Yu E, Zikelic D, Henzinger TA. Neural control and certificate repair via runtime
    monitoring. In: <i>Proceedings of the 39th AAAI Conference on Artificial Intelligence</i>.
    Vol 39. Association for the Advancement of Artificial Intelligence; 2025:26409-26417.
    doi:<a href="https://doi.org/10.1609/aaai.v39i25.34840">10.1609/aaai.v39i25.34840</a>'
  apa: 'Yu, E., Zikelic, D., &#38; Henzinger, T. A. (2025). Neural control and certificate
    repair via runtime monitoring. In <i>Proceedings of the 39th AAAI Conference on
    Artificial Intelligence</i> (Vol. 39, pp. 26409–26417). Philadelphia, PA, United
    States: Association for the Advancement of Artificial Intelligence. <a href="https://doi.org/10.1609/aaai.v39i25.34840">https://doi.org/10.1609/aaai.v39i25.34840</a>'
  chicago: Yu, Emily, Dorde Zikelic, and Thomas A Henzinger. “Neural Control and Certificate
    Repair via Runtime Monitoring.” In <i>Proceedings of the 39th AAAI Conference
    on Artificial Intelligence</i>, 39:26409–17. Association for the Advancement of
    Artificial Intelligence, 2025. <a href="https://doi.org/10.1609/aaai.v39i25.34840">https://doi.org/10.1609/aaai.v39i25.34840</a>.
  ieee: E. Yu, D. Zikelic, and T. A. Henzinger, “Neural control and certificate repair
    via runtime monitoring,” in <i>Proceedings of the 39th AAAI Conference on Artificial
    Intelligence</i>, Philadelphia, PA, United States, 2025, vol. 39, no. 25, pp.
    26409–26417.
  ista: 'Yu E, Zikelic D, Henzinger TA. 2025. Neural control and certificate repair
    via runtime monitoring. Proceedings of the 39th AAAI Conference on Artificial
    Intelligence. AAAI: Conference on Artificial Intelligence vol. 39, 26409–26417.'
  mla: Yu, Emily, et al. “Neural Control and Certificate Repair via Runtime Monitoring.”
    <i>Proceedings of the 39th AAAI Conference on Artificial Intelligence</i>, vol.
    39, no. 25, Association for the Advancement of Artificial Intelligence, 2025,
    pp. 26409–17, doi:<a href="https://doi.org/10.1609/aaai.v39i25.34840">10.1609/aaai.v39i25.34840</a>.
  short: E. Yu, D. Zikelic, T.A. Henzinger, in:, Proceedings of the 39th AAAI Conference
    on Artificial Intelligence, Association for the Advancement of Artificial Intelligence,
    2025, pp. 26409–26417.
conference:
  end_date: 2025-03-04
  location: Philadelphia, PA, United States
  name: 'AAAI: Conference on Artificial Intelligence'
  start_date: 2025-02-25
corr_author: '1'
date_created: 2025-05-11T22:02:40Z
date_published: 2025-04-11T00:00:00Z
date_updated: 2025-05-12T09:49:25Z
day: '11'
department:
- _id: ToHe
doi: 10.1609/aaai.v39i25.34840
ec_funded: 1
external_id:
  arxiv:
  - '2412.12996'
intvolume: '        39'
issue: '25'
language:
- iso: eng
main_file_link:
- open_access: '1'
  url: https://doi.org/10.48550/arXiv.2412.12996
month: '04'
oa: 1
oa_version: Preprint
page: 26409-26417
project:
- _id: 62781420-2b32-11ec-9570-8d9b63373d4d
  call_identifier: H2020
  grant_number: '101020093'
  name: Vigilant Algorithmic Monitoring of Software
publication: Proceedings of the 39th AAAI Conference on Artificial Intelligence
publication_identifier:
  eissn:
  - 2374-3468
  issn:
  - 2159-5399
publication_status: published
publisher: Association for the Advancement of Artificial Intelligence
quality_controlled: '1'
scopus_import: '1'
status: public
title: Neural control and certificate repair via runtime monitoring
type: conference
user_id: 2DF688A6-F248-11E8-B48F-1D18A9856A87
volume: 39
year: '2025'
...