Clean Label

Clean-label attacks represent a sophisticated form of data poisoning targeting machine learning models, where malicious actors subtly manipulate training data without altering labels, making detection challenging. Current research focuses on developing increasingly stealthy clean-label attacks across various domains, including image classification, natural language processing, and graph neural networks, often employing generative adversarial networks or genetic algorithms to create realistic poisoned samples. These attacks highlight vulnerabilities in model robustness and training procedures, underscoring the need for improved defenses and more rigorous evaluation methodologies to ensure the reliability and security of machine learning systems in real-world applications.

Papers

October 14, 2024

The Implicit Bias of Structured State Space Models Can Be Poisoned With Clean Labels
Yonatan Slutzky, Yotam Alexander, Noam Razin, Nadav Cohen
Strong Generalization Adversarial Learning State Space Model Neural Network Model Implicit Bias Clean Label Label Poisoning

December 30, 2023

A clean-label graph backdoor attack method in node classification task
Xiaogang Xing, Ming Xu, Yujing Bai, Dongdong Yang
Backdoor Attack Node Classification Stealthy Backdoor Attack Graph Backdoor Attack Clean Label

December 3, 2023

Towards Sample-specific Backdoor Attack with Clean Labels via Attribute Trigger
Yiming Li, Mingyan Zhu, Junfeng Guo, Tao Wei, Shu-Tao Xia, Zhan Qin
Backdoor Attack Backdoor Defense Attribute Manipulation Clean Label Attack Paradigm Sample Pair

November 28, 2023

Clean Label Disentangling for Medical Image Segmentation with Noisy Labels
Zicheng Wang, Zhen Zhao, Erjian Guo, Luping Zhou
Medical Image Segmentation Noisy Label Label Noise Clean Label

May 31, 2023

March 20, 2023

PASS: Peer-Agreement based Sample Selection for training with Noisy Labels
Arpit Garg, Cuong Nguyen, Rafael Felix, Thanh-Toan Do, Gustavo Carneiro
Training Data Noisy Label Decision Boundary Peer Agent Sample Selection Clean Label Noisy Label Sample Noisy Label Detection

July 1, 2022

BadHash: Invisible Backdoor Attacks against Deep Hashing with Clean Label
Shengshan Hu, Ziqi Zhou, Yechao Zhang, Leo Yu Zhang, Yifeng Zheng, Yuanyuan HE, Hai Jin
Adversarial Attack Backdoor Attack Deep Hashing Adversarial Backdoor Invisible Backdoor Attack Clean Label Image Backdoor Attack

June 3, 2022

Kallima: A Clean-label Framework for Textual Backdoor Attacks
Xiaoyi Chen, Yinpeng Dong, Zeyu Sun, Shengfang Zhai, Qingni Shen, Zhonghai Wu
Backdoor Attack Textual Backdoor Attack Backdoor Sample Clean Label

November 15, 2021

Triggerless Backdoor Attack for NLP Tasks with Clean Labels
Leilei Gan, Jiwei Li, Tianwei Zhang, Xiaoya Li, Yuxian Meng, Fei Wu, Yi Yang, Shangwei Guo, Chun Fan
Backdoor Attack NLP Field NLP Task NLP Model Textual Backdoor Attack Multimodal Backdoor Semantic Anomaly Clean Label