paper-with-me

홈 › Papers

Adversarial Vulnerability of Neural Networks Increases with Input Dimension

2019-05-01 · ICLR 2019 5 · Carl-Johann Simon-Gabriel, Yann Ollivier, Léon Bottou, Bernhard Schölkopf, David Lopez-Paz

Over the past four years, neural networks have been proven vulnerable to adversarial images: targeted but imperceptible image perturbations lead to drastically different predictions. We show that adversarial vulnerability increases with the gradients of the training objective when viewed as a function of the inputs. For most current network architectures, we prove that the L1-norm of these gradients grows as the square root of the input size. These nets therefore become increasingly vulnerable with growing image size. Our proofs rely on the network’s weight distribution at initialization, but extensive experiments confirm that our conclusions still hold after usual training.

📄 PDF Abstract BibTeX

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

First-order Adversarial Vulnerability of Neural Networks and Input Dimension

2018-02-05 · ICLR 2019 5 · Carl-Johann Simon-Gabriel, Yann Ollivier, Léon Bottou, Bernhard Schölkopf 외

Over the past few years, neural networks were proven vulnerable to adversarial images: targeted but imperceptible image perturbations lead to drastically different predictions. We show that adversarial vulnerability incr…

The Role of Input Dimensionality in the Emergence and Targeted Control of Adversarial Examples

2026-06-24 · Nasrin Malekzadeh Goradel, Niccolo Pancino, Yaser Gholizade Atani, Benedetta Tondi 외 arxiv

Several theoretical works have tried to explain the adversarial vulnerability of deep neural networks through properties of high-dimensional geometry. However, the assumptions underlying these works are rarely examined e…

On the relationship between class selectivity, dimensionality, and robustness

2020-07-08 · Matthew L. Leavitt, Ari S. Morcos

While the relative trade-offs between sparse and distributed representations in deep neural networks (DNNs) are well-studied, less is known about how these trade-offs apply to representations of semantically-meaningful i…

Defending Against Adversarial Examples by Regularized Deep Embedding

2019-09-25 · Yao Li, Martin Renqiang Min, Wenchao Yu, Cho-Jui Hsieh 외

Recent studies have demonstrated the vulnerability of deep convolutional neural networks against adversarial examples. Inspired by the observation that the intrinsic dimension of image data is much smaller than its pixel…

Adversarial AttackAdversarial Robustness

Linking average- and worst-case perturbation robustness via class selectivity and dimensionality

2020-09-28 · Matthew L Leavitt, Ari S. Morcos

Representational sparsity is known to affect robustness to input perturbations in deep neural networks (DNNs), but less is known about how the semantic content of representations affects robustness. Class selectivity—the…