Feature Compression May Be the Root Cause of Adversarial Fragility in Neural Network Classifiers (Student Abstract)

Jingchao Gao; Ziqing Lu; Raghu Mudumbai; Xiaodong Wu; Jirong Yi; Myung Cho; Catherine Xu; Hui Xie

AAAI 2026technical0 citations

Feature Compression May Be the Root Cause of Adversarial Fragility in Neural Network Classifiers (Student Abstract)

Jingchao Gao, Ziqing Lu, Raghu Mudumbai, Xiaodong Wu, Jirong Yi, Myung Cho, Catherine Xu, Hui Xie

Abstract

In this paper, we study the adversarial robustness of deep neural networks (DNN) for classification against optimal classifiers. We look at the smallest magnitude of possible additive perturbations that can change a classifier

BibTeX

@inproceedings{aaai2026_featurecompressi,
  title = {Feature Compression May Be the Root Cause of Adversarial Fragility in Neural Network Classifiers (Student Abstract)},
  author = {Jingchao Gao and Ziqing Lu and Raghu Mudumbai and Xiaodong Wu and Jirong Yi and Myung Cho and Catherine Xu and Hui Xie and Weiyu Xu},
  booktitle = {AAAI 2026},
  year = {2026}
}