We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: A Robust Optimisation Perspective on Counterexample-Guided Repair of Neural Networks

Abstract: Counterexample-guided repair aims at creating neural networks with mathematical safety guarantees, facilitating the application of neural networks in safety-critical domains. However, whether counterexample-guided repair is guaranteed to terminate remains an open question. We approach this question by showing that counterexample-guided repair can be viewed as a robust optimisation algorithm. While termination guarantees for neural network repair itself remain beyond our reach, we prove termination for more restrained machine learning models and disprove termination in a general setting. We empirically study the practical implications of our theoretical results, demonstrating the suitability of common verifiers and falsifiers for repair despite a disadvantageous theoretical result. Additionally, we use our theoretical insights to devise a novel algorithm for repairing linear regression models, surpassing existing approaches.
Comments: 22 pages + 9 pages references and appendix, 4 figures
Subjects: Machine Learning (cs.LG); Optimization and Control (math.OC)
Cite as: arXiv:2301.11342 [cs.LG]
  (or arXiv:2301.11342v1 [cs.LG] for this version)

Submission history

From: David Boetius [view email]
[v1] Thu, 26 Jan 2023 19:00:02 GMT (393kb,D)

Link back to: arXiv, form interface, contact.