Convex neural network synthesis for robustness in the 1-norm

Drummond, Ross; Guiver, Chris; Turner, Matthew C.

Abstract

With neural networks being used to control safety-critical systems, they increasingly have to be both accurate (in the sense of matching inputs to outputs) and robust. However, these two properties are often at odds with each other and a trade-off has to be navigated. To address this issue, this paper proposes a method to generate an approximation of a neural network which is certifiably more robust. Crucially, the method is fully convex and posed as a semi-definite programme. An application to robustifying model predictive control is used to demonstrate the results. The aim of this work is to introduce a method to navigate the neural network robustness/accuracy trade-off.

Citation

Drummond, R., Guiver, C., & Turner, M. C. (2024, July). Convex neural network synthesis for robustness in the 1-norm. Presented at 6th Annual Learning for Dynamics & Control Conference, Oxford, England

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 6th Annual Learning for Dynamics & Control Conference
• Start Date: Jul 15, 2024
• End Date: Jul 17, 2024
• Acceptance Date: Mar 28, 2024
• Publication Date: 2024
• Deposit Date: May 21, 2024
• Publicly Available Date: Jan 1, 2026
• Print ISSN: 1532-4435
• Electronic ISSN: 1533-7928
• Peer Reviewed: Peer Reviewed
• Volume: 242
• Pages: 1388-1399
• Series Number: Proceedings of Machine Learning Research
• Series ISSN: 2640-3498
• Keywords: Neural network robustness, convex synthesis, accuracy vs. robustness trade-off
• Public URL: http://researchrepository.napier.ac.uk/Output/3646219
• Publisher URL: https://proceedings.mlr.press/
• External URL: https://l4dc.web.ox.ac.uk/home