Transforming gradient-based techniques into interpretable methods - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2024

Transforming gradient-based techniques into interpretable methods

Résumé

The explication of Convolutional Neural Networks (CNN) through xAI techniques often poses challenges in interpretation. The inherent complexity of input features, notably pixels extracted from images, engenders complex correlations. Gradient-based methodologies, exemplified by Integrated Gradients (IG), effectively demonstrate the significance of these features. Nevertheless, the conversion of these explanations into images frequently yields considerable noise. Presently, we introduce GAD (Gradient Artificial Distancing) as a supportive framework for gradient-based techniques. Its primary objective is to accentuate influential regions by establishing distinctions between classes. The essence of GAD is to limit the scope of analysis during visualization and, consequently reduce image noise. Empirical investigations involving occluded images have demonstrated that the identified regions through this methodology indeed play a pivotal role in facilitating class differentiation.
Fichier principal
Vignette du fichier
main.pdf (9.36 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04414672 , version 1 (24-01-2024)

Identifiants

  • HAL Id : hal-04414672 , version 1

Citer

Caroline Mazini Rodrigues, Nicolas Boutry, Laurent Najman. Transforming gradient-based techniques into interpretable methods. 2024. ⟨hal-04414672⟩
71 Consultations
9 Téléchargements

Partager

Gmail Facebook X LinkedIn More