Faster Verified Explanations for Neural Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: De Palma, Alessandro, Dolcetti, Greta, Urban, Caterina
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911659913641984
author De Palma, Alessandro
Dolcetti, Greta
Urban, Caterina
author_facet De Palma, Alessandro
Dolcetti, Greta
Urban, Caterina
contents Verified explanations are a principled way to explain the decisions taken by neural networks, which are otherwise black-box in nature. However, these techniques face significant scalability challenges, as they require multiple calls to neural network verifiers, each of them with an exponential worst-case complexity. We present FaVeX, a novel algorithm to compute verified explanations. FaVeX accelerates the computation by dynamically combining batch and sequential processing of input features, and by reusing information from previous queries, both when proving invariances with respect to certain input features, and when searching for feature assignments altering the prediction. Furthermore, we present a novel and hierarchical definition of verified explanations, termed verifieroptimal robust explanations, that explicitly factors the incompleteness of network verifiers within the explanation. Our comprehensive experimental evaluation demonstrates the superior scalability of both FaVeX, and of verifier-optimal robust explanations, which together can produce meaningful formal explanation on networks with hundreds of thousands of non-linear activations.
format Preprint
id arxiv_https___arxiv_org_abs_2512_00164
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Faster Verified Explanations for Neural Networks
De Palma, Alessandro
Dolcetti, Greta
Urban, Caterina
Machine Learning
Programming Languages
Verified explanations are a principled way to explain the decisions taken by neural networks, which are otherwise black-box in nature. However, these techniques face significant scalability challenges, as they require multiple calls to neural network verifiers, each of them with an exponential worst-case complexity. We present FaVeX, a novel algorithm to compute verified explanations. FaVeX accelerates the computation by dynamically combining batch and sequential processing of input features, and by reusing information from previous queries, both when proving invariances with respect to certain input features, and when searching for feature assignments altering the prediction. Furthermore, we present a novel and hierarchical definition of verified explanations, termed verifieroptimal robust explanations, that explicitly factors the incompleteness of network verifiers within the explanation. Our comprehensive experimental evaluation demonstrates the superior scalability of both FaVeX, and of verifier-optimal robust explanations, which together can produce meaningful formal explanation on networks with hundreds of thousands of non-linear activations.
title Faster Verified Explanations for Neural Networks
topic Machine Learning
Programming Languages
url https://arxiv.org/abs/2512.00164