paper-with-me

홈 › Papers

Superclass Adversarial Attack

2022-05-29 · Soichiro Kumano, Hiroshi Kera, Toshihiko Yamasaki

Adversarial attacks have only focused on changing the predictions of the classifier, but their danger greatly depends on how the class is mistaken. For example, when an automatic driving system mistakes a Persian cat for a Siamese cat, it is hardly a problem. However, if it mistakes a cat for a 120km/h minimum speed sign, serious problems can arise. As a stepping stone to more threatening adversarial attacks, we consider the superclass adversarial attack, which causes misclassification of not only fine classes, but also superclasses. We conducted the first comprehensive analysis of superclass adversarial attacks (an existing and 19 new methods) in terms of accuracy, speed, and stability, and identified several strategies to achieve better performance. Although this study is aimed at superclass misclassification, the findings can be applied to other problem settings involving multiple classes, such as top-k and multi-label classification attacks.

📄 PDF Abstract BibTeX arXiv:2205.14629

Code (0)

등록된 구현이 없습니다.

Tasks

Adversarial AttackMulti-Label ClassificationMUlTI-LABEL-ClASSIFICATION

Methods 이 논문이 사용한 방법론

SPEED The monocular depth estimation (MDE) is the task of estimating depth from a single frame. This information is an essential knowledge in many computer vision tasks such as scene…

Similar Papers 제목 키워드 기반

Hierarchically Robust Zero-shot Vision-language Models

2026-04-20 · Junhao Dong, Yifei Zhang, Hao Zhu, Yew-Soon Ong 외 arxiv

Vision-Language Models (VLMs) can perform zero-shot classification but are susceptible to adversarial attacks. While robust fine-tuning improves their robustness, existing approaches align fixed text embeddings with an i…

Adversarial Robustness

Sample Complexity Bounds for Robustly Learning Decision Lists against Evasion Attacks

2022-05-12 · Pascale Gourdeau, Varun Kanade, Marta Kwiatkowska, James Worrell

A fundamental problem in adversarial machine learning is to quantify how much training data is needed in the presence of evasion attacks. In this paper we address this issue within the framework of PAC learning, focusing…

PAC learning

Superclass Learning With Representation Enhancement

2023-01-01 · CVPR 2023 1 · Jinlong Kang, Liyuan Shang, Suyun Zhao, Hong Chen 외

In many real scenarios, data are often divided into a handful of artificial super categories in terms of expert knowledge rather than the representations of images. Concretely, a superclass may contain massive and va…

Classification Done Right for Vision-Language Pre-Training

2024-11-05 · Zilong Huang, Qinghao Ye, Bingyi Kang, Jiashi Feng 외

We introduce SuperClass, a super simple classification method for vision-language pre-training on image-text data. Unlike its contrastive counterpart CLIP who contrast with a text encoder, SuperClass directly utilizes to…

Classification

What can we learn from misclassified ImageNet images?

2022-01-20 · Shixian Wen, Amanda Sofie Rios, Kiran Lekkala, Laurent Itti

Understanding the patterns of misclassified ImageNet images is particularly important, as it could guide us to design deep neural networks (DNN) that generalize better. However, the richness of ImageNet imposes difficult…

GPUQuantization