paper-with-me

홈 › Papers

Adversarial Training Reduces Information and Improves Transferability

2020-07-22 · Matteo Terzi, Alessandro Achille, Marco Maggipinto, Gian Antonio Susto

Recent results show that features of adversarially trained networks for classification, in addition to being robust, enable desirable properties such as invertibility. The latter property may seem counter-intuitive as it is widely accepted by the community that classification models should only capture the minimal information (features) required for the task. Motivated by this discrepancy, we investigate the dual relationship between Adversarial Training and Information Theory. We show that the Adversarial Training can improve linear transferability to new tasks, from which arises a new trade-off between transferability of representations and accuracy on the source task. We validate our results employing robust networks trained on CIFAR-10, CIFAR-100 and ImageNet on several datasets. Moreover, we show that Adversarial Training reduces Fisher information of representations about the input and of the weights about the task, and we provide a theoretical argument which explains the invertibility of deterministic networks without violating the principle of minimality. Finally, we leverage our theoretical insights to remarkably improve the quality of reconstructed images through inversion.

📄 PDF Abstract BibTeX arXiv:2007.11259

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Adversarial Pixel Restoration as a Pretext Task for Transferable Perturbations

2022-07-18 · Hashmat Shadab Malik, Shahina K Kunhimon, Muzammal Naseer, Salman Khan 외

Transferable adversarial attacks optimize adversaries from a pretrained surrogate model and known label space to fool the unknown black-box models. Therefore, these attacks are restricted by the availability of an effect…

object-detectionObject DetectionVideo SegmentationVideo Semantic Segmentation

Reducing Adversarial Example Transferability Using Gradient Regularization

2019-04-16 · George Adam, Petr Smirnov, Benjamin Haibe-Kains, Anna Goldenberg

Deep learning algorithms have increasingly been shown to lack robustness to simple adversarial examples (AdvX). An equally troubling observation is that these adversarial examples transfer between different architectures…

Early Stop And Adversarial Training Yield Better surrogate Model: Very Non-Robust Features Harm Adversarial Transferability

2021-09-29 · Chaoning Zhang, Gyusang Cho, Philipp Benz, Kang Zhang 외

The transferability of adversarial examples (AE); known as adversarial transferability, has attracted significant attention because it can be exploited for TransferableBlack-box Attacks (TBA). Most lines of works attribu…

Attribute

CAG: A Real-time Low-cost Enhanced-robustness High-transferability Content-aware Adversarial Attack Generator

2019-12-16 · Huy Phan, Yi Xie, Siyu Liao, Jie Chen 외

Deep neural networks (DNNs) are vulnerable to adversarial attack despite their tremendous success in many AI fields. Adversarial attack is a method that causes the intended misclassfication by adding imperceptible pertur…

Adversarial Attack

A Unified Approach to Interpreting and Boosting Adversarial Transferability

2020-10-08 · Xin Wang, Jie Ren, Shuyun Lin, Xiangming Zhu 외

In this paper, we use the interaction inside adversarial perturbations to explain and boost the adversarial transferability. We discover and prove the negative correlation between the adversarial transferability and the …