paper-with-me

Papers

Investigating Catastrophic Overfitting in Fast Adversarial Training: A Self-fitting Perspective

2023-02-23 · Zhengbao He, Tao Li, Sizhe Chen, Xiaolin Huang

Although fast adversarial training provides an efficient approach for building robust networks, it may suffer from a serious problem known as catastrophic overfitting (CO), where multi-step robust accuracy suddenly collapses to zero. In this paper, we for the first time decouple single-step adversarial examples into data-information and self-information, which reveals an interesting phenomenon called "self-fitting". Self-fitting, i.e., the network learns the self-information embedded in single-step perturbations, naturally leads to the occurrence of CO. When self-fitting occurs, the network experiences an obvious "channel differentiation" phenomenon that some convolution channels accounting for recognizing self-information become dominant, while others for data-information are suppressed. In this way, the network can only recognize images with sufficient self-information and loses generalization ability to other types of data. Based on self-fitting, we provide new insights into the existing methods to mitigate CO and extend CO to multi-step adversarial training. Our findings reveal a self-learning mechanism in adversarial training and open up new perspectives for suppressing different kinds of information to mitigate CO.

📄 PDF Abstract BibTeX arXiv:2302.11963

Code (0)

등록된 구현이 없습니다.

Tasks

Self-Learning

Methods 이 논문이 사용한 방법론

Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Self-Learning 설명 없음

Similar Papers 제목 키워드 기반

Prior-Guided Adversarial Initialization for Fast Adversarial Training

2022-07-18 · Xiaojun Jia, Yong Zhang, Xingxing Wei, Baoyuan Wu 외

Fast adversarial training (FAT) effectively improves the efficiency of standard adversarial training (SAT). However, initial FAT encounters catastrophic overfitting, i.e.,the robust accuracy against adversarial attacks s…

Adversarial AttackAdversarial Attack on Video Classification

Understanding Catastrophic Overfitting in Single-step Adversarial Training

2020-10-05 · Hoki Kim, Woojin Lee, Jaewook Lee

Although fast adversarial training has demonstrated both robustness and efficiency, the problem of "catastrophic overfitting" has been observed. This is a phenomenon in which, during single-step adversarial training, the…

Revisiting and Exploring Efficient Fast Adversarial Training via LAW: Lipschitz Regularization and Auto Weight Averaging

2023-08-22 · Xiaojun Jia, Yuefeng Chen, Xiaofeng Mao, Ranjie Duan 외

Fast Adversarial Training (FAT) not only improves the model robustness but also reduces the training cost of standard adversarial training. However, fast adversarial training often suffers from Catastrophic Overfitting (…

Adversarial RobustnessData Augmentation

Improving Fast Adversarial Training Paradigm: An Example Taxonomy Perspective

2024-07-22 · Jie Gui, Chengze Jiang, Minjing Dong, Kun Tong 외

While adversarial training is an effective defense method against adversarial attacks, it notably increases the training cost. To this end, fast adversarial training (FAT) is presented for efficient training and has beco…

CoLA

Fast Adversarial Training with Adaptive Step Size

2022-06-06 · Zhichao Huang, Yanbo Fan, Chen Liu, Weizhong Zhang 외

While adversarial training and its variants have shown to be the most effective algorithms to defend against adversarial attacks, their extremely slow training process makes it hard to scale to large datasets like ImageN…