NAT: Learning to Attack Neurons for Enhanced Adversarial Transferability

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nakka, Krishna Kanth, Alahi, Alexandre
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912551300759552
author Nakka, Krishna Kanth
Alahi, Alexandre
author_facet Nakka, Krishna Kanth
Alahi, Alexandre
contents The generation of transferable adversarial perturbations typically involves training a generator to maximize embedding separation between clean and adversarial images at a single mid-layer of a source model. In this work, we build on this approach and introduce Neuron Attack for Transferability (NAT), a method designed to target specific neuron within the embedding. Our approach is motivated by the observation that previous layer-level optimizations often disproportionately focus on a few neurons representing similar concepts, leaving other neurons within the attacked layer minimally affected. NAT shifts the focus from embedding-level separation to a more fundamental, neuron-specific approach. We find that targeting individual neurons effectively disrupts the core units of the neural network, providing a common basis for transferability across different models. Through extensive experiments on 41 diverse ImageNet models and 9 fine-grained models, NAT achieves fooling rates that surpass existing baselines by over 14\% in cross-model and 4\% in cross-domain settings. Furthermore, by leveraging the complementary attacking capabilities of the trained generators, we achieve impressive fooling rates within just 10 queries. Our code is available at: https://krishnakanthnakka.github.io/NAT/
format Preprint
id arxiv_https___arxiv_org_abs_2508_16937
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle NAT: Learning to Attack Neurons for Enhanced Adversarial Transferability
Nakka, Krishna Kanth
Alahi, Alexandre
Computer Vision and Pattern Recognition
The generation of transferable adversarial perturbations typically involves training a generator to maximize embedding separation between clean and adversarial images at a single mid-layer of a source model. In this work, we build on this approach and introduce Neuron Attack for Transferability (NAT), a method designed to target specific neuron within the embedding. Our approach is motivated by the observation that previous layer-level optimizations often disproportionately focus on a few neurons representing similar concepts, leaving other neurons within the attacked layer minimally affected. NAT shifts the focus from embedding-level separation to a more fundamental, neuron-specific approach. We find that targeting individual neurons effectively disrupts the core units of the neural network, providing a common basis for transferability across different models. Through extensive experiments on 41 diverse ImageNet models and 9 fine-grained models, NAT achieves fooling rates that surpass existing baselines by over 14\% in cross-model and 4\% in cross-domain settings. Furthermore, by leveraging the complementary attacking capabilities of the trained generators, we achieve impressive fooling rates within just 10 queries. Our code is available at: https://krishnakanthnakka.github.io/NAT/
title NAT: Learning to Attack Neurons for Enhanced Adversarial Transferability
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2508.16937