no code implementations • 1 Jan 2021 • Max Revay, Ruigang Wang, Ian Manchester
In image classification experiments we show that the Lipschitz bounds are very accurate and improve robustness to adversarial attacks.
Generalization Bounds Image Classification