Training Robust Neural Networks Using Lipschitz Bounds

• Published in: IEEE control systems letters Vol. 6; pp. 121 - 126
• Main Authors: Pauli, Patricia, Koch, Anne, Berberich, Julian, Kohler, Paul, Allgower, Frank
• Format: Journal Article
• Language: English
• Published: IEEE 2022
• Subjects: Artificial neural networks; Estimation; Linear matrix inequalities; neural networks; Neurons; Perturbation methods; Robustness; Training; Upper bound

Due to their susceptibility to adversarial perturbations, neural networks (NNs) are hardly used in safety-critical applications. One measure of robustness to such perturbations in the input is the Lipschitz constant of the input-output map defined by an NN. In this letter, we propose a framework to...