Pré-Publication, Document De Travail Année : 2023

Exploring the Connection between Neuron Coverage and Adversarial Robustness in DNN Classifiers

Résumé

The lack of robustness in neural network classifiers, especially when facing adversarial attacks, is a significant limitation for critical applications. While some researchers have suggested a connection between neuron coverage during training and vulnerability to adversarial perturbations, concrete experimental evidence supporting this claim is lacking. This paper empirically investigates the impact of maximizing neuron coverage during training and assess the effectiveness of adversarial attacks on under-covered neurons. Additionally, we explore the potential of leveraging coverage for designing more efficient attacks. Our experiments reveal no clear correlation between neuron coverage, adversarial robustness, or attack effectiveness.

Fichier principal
Vignette du fichier
EXPLORING THE CONNECTION BETWEEN NEURON COVERAGE AND ADVERSARIAL ROBUSTNESS IN DNN CLASSIFIERS.pdf (338.26 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04151746 , version 1 (05-07-2023)

Licence

Identifiants

  • HAL Id : hal-04151746 , version 1

Citer

William Piat, Jalal Fadili, Frédéric Jurie. Exploring the Connection between Neuron Coverage and Adversarial Robustness in DNN Classifiers. 2023. ⟨hal-04151746⟩
203 Consultations
325 Téléchargements

Partager

  • More