Verification of Neural Networks' Global Robustness

arXiv:2402.19322v1 Announce Type: new
Abstract: Neural networks are successful in various applications but are also susceptible to adversarial attacks. To show the safety of network classifiers, many verifiers have been introduced to reason about the local robustness of a given input to a given perturbation. While successful, local robustness cannot generalize to unseen inputs. Several works analyze global robustness properties, however, neither can provide a precise guarantee about the cases where a network classifier does not change its classification. In …

abstract adversarial adversarial attacks applications arxiv attacks classifiers cs.cr cs.lg cs.pl global inputs network networks neural networks reason robustness safety show type verification