TriGuard: Testing Model Safety with Attribution Entropy, Verification, and Drift

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Mahato, Dipesh Tharu, Poudel, Rohan, Dhungana, Pramod
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911009129627648
author Mahato, Dipesh Tharu
Poudel, Rohan
Dhungana, Pramod
author_facet Mahato, Dipesh Tharu
Poudel, Rohan
Dhungana, Pramod
contents Deep neural networks often achieve high accuracy, but ensuring their reliability under adversarial and distributional shifts remains a pressing challenge. We propose TriGuard, a unified safety evaluation framework that combines (1) formal robustness verification, (2) attribution entropy to quantify saliency concentration, and (3) a novel Attribution Drift Score measuring explanation stability. TriGuard reveals critical mismatches between model accuracy and interpretability: verified models can still exhibit unstable reasoning, and attribution-based signals provide complementary safety insights beyond adversarial accuracy. Extensive experiments across three datasets and five architectures show how TriGuard uncovers subtle fragilities in neural reasoning. We further demonstrate that entropy-regularized training reduces explanation drift without sacrificing performance. TriGuard advances the frontier in robust, interpretable model evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2506_14217
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle TriGuard: Testing Model Safety with Attribution Entropy, Verification, and Drift
Mahato, Dipesh Tharu
Poudel, Rohan
Dhungana, Pramod
Machine Learning
Artificial Intelligence
Deep neural networks often achieve high accuracy, but ensuring their reliability under adversarial and distributional shifts remains a pressing challenge. We propose TriGuard, a unified safety evaluation framework that combines (1) formal robustness verification, (2) attribution entropy to quantify saliency concentration, and (3) a novel Attribution Drift Score measuring explanation stability. TriGuard reveals critical mismatches between model accuracy and interpretability: verified models can still exhibit unstable reasoning, and attribution-based signals provide complementary safety insights beyond adversarial accuracy. Extensive experiments across three datasets and five architectures show how TriGuard uncovers subtle fragilities in neural reasoning. We further demonstrate that entropy-regularized training reduces explanation drift without sacrificing performance. TriGuard advances the frontier in robust, interpretable model evaluation.
title TriGuard: Testing Model Safety with Attribution Entropy, Verification, and Drift
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2506.14217