Increasing Interpretability of Neural Networks By Approximating Human Visual Saliency

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Boyd, Aidan, Trabelsi, Mohamed, Uzunalioglu, Huseyin, Kushnir, Dan
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909357784956928
author Boyd, Aidan
Trabelsi, Mohamed
Uzunalioglu, Huseyin
Kushnir, Dan
author_facet Boyd, Aidan
Trabelsi, Mohamed
Uzunalioglu, Huseyin
Kushnir, Dan
contents Understanding specifically where a model focuses on within an image is critical for human interpretability of the decision-making process. Deep learning-based solutions are prone to learning coincidental correlations in training datasets, causing over-fitting and reducing the explainability. Recent advances have shown that guiding models to human-defined regions of saliency within individual images significantly increases performance and interpretability. Human-guided models also exhibit greater generalization capabilities, as coincidental dataset features are avoided. Results show that models trained with saliency incorporation display an increase in interpretability of up to 30% over models trained without saliency information. The collection of this saliency information, however, can be costly, laborious and in some cases infeasible. To address this limitation, we propose a combination strategy of saliency incorporation and active learning to reduce the human annotation data required by 80% while maintaining the interpretability and performance increase from human saliency. Extensive experimentation outlines the effectiveness of the proposed approach across five public datasets and six active learning criteria.
format Preprint
id arxiv_https___arxiv_org_abs_2410_16115
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Increasing Interpretability of Neural Networks By Approximating Human Visual Saliency
Boyd, Aidan
Trabelsi, Mohamed
Uzunalioglu, Huseyin
Kushnir, Dan
Computer Vision and Pattern Recognition
Understanding specifically where a model focuses on within an image is critical for human interpretability of the decision-making process. Deep learning-based solutions are prone to learning coincidental correlations in training datasets, causing over-fitting and reducing the explainability. Recent advances have shown that guiding models to human-defined regions of saliency within individual images significantly increases performance and interpretability. Human-guided models also exhibit greater generalization capabilities, as coincidental dataset features are avoided. Results show that models trained with saliency incorporation display an increase in interpretability of up to 30% over models trained without saliency information. The collection of this saliency information, however, can be costly, laborious and in some cases infeasible. To address this limitation, we propose a combination strategy of saliency incorporation and active learning to reduce the human annotation data required by 80% while maintaining the interpretability and performance increase from human saliency. Extensive experimentation outlines the effectiveness of the proposed approach across five public datasets and six active learning criteria.
title Increasing Interpretability of Neural Networks By Approximating Human Visual Saliency
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2410.16115