paper-with-me

Papers

SynGhost: Invisible and Universal Task-agnostic Backdoor Attack via Syntactic Transfer

2024-02-29 · Pengzhou Cheng, Wei Du, Zongru Wu, Fengwei Zhang, Libo Chen, Zhuosheng Zhang, Gongshen Liu

Although pre-training achieves remarkable performance, it suffers from task-agnostic backdoor attacks due to vulnerabilities in data and training mechanisms. These attacks can transfer backdoors to various downstream tasks. In this paper, we introduce $\mathtt{maxEntropy}$, an entropy-based poisoning filter that mitigates such risks. To overcome the limitations of manual target setting and explicit triggers, we propose $\mathtt{SynGhost}$, an invisible and universal task-agnostic backdoor attack via syntactic transfer, further exposing vulnerabilities in pre-trained language models (PLMs). Specifically, $\mathtt{SynGhost}$ injects multiple syntactic backdoors into the pre-training space through corpus poisoning, while preserving the PLM's pre-training capabilities. Second, $\mathtt{SynGhost}$ adaptively selects optimal targets based on contrastive learning, creating a uniform distribution in the pre-training space. To identify syntactic differences, we also introduce an awareness module to minimize interference between backdoors. Experiments show that $\mathtt{SynGhost}$ poses significant threats and can transfer to various downstream tasks. Furthermore, $\mathtt{SynGhost}$ resists defenses based on perplexity, fine-pruning, and $\mathtt{maxEntropy}$. The code is available at https://github.com/Zhou-CyberSecurity-AI/SynGhost.

📄 PDF Abstract BibTeX arXiv:2402.18945

Code (1)

zhou-cybersecurity-ai/synghost 공식 구현 pytorch

Tasks

Backdoor AttackContrastive LearningNatural Language Understanding

Methods 이 논문이 사용한 방법론

Contrastive Learning 설명 없음

Similar Papers 제목 키워드 기반

IU: Imperceptible Universal Backdoor Attack

2026-02-28 · Hsin Lin, Yan-Lun Chen, Ren-Hung Hwang, Chia-Mu Yu arxiv

Backdoor attacks pose a critical threat to the security of deep neural networks, yet existing efforts on universal backdoors often rely on visually salient patterns, making them easier to detect and less practical at sca…

Invisible Backdoor Attacks on Deep Neural Networks via Steganography and Regularization

2019-09-06 · Shaofeng Li, Minhui Xue, Benjamin Zi Hao Zhao, Haojin Zhu 외

Deep neural networks (DNNs) have been proven vulnerable to backdoor attacks, where hidden features (patterns) trained to a normal model, which is only activated by some specific input (called triggers), trick the model i…

GhostEncoder: Stealthy Backdoor Attacks with Dynamic Triggers to Pre-trained Encoders in Self-supervised Learning

2023-10-01 · Qiannan Wang, Changchun Yin, Zhe Liu, Liming Fang 외

Within the realm of computer vision, self-supervised learning (SSL) pertains to training pre-trained image encoders utilizing a substantial quantity of unlabeled images. Pre-trained image encoders can serve as feature ex…

Backdoor AttackImage SteganographySelf-Supervised Learning

UOR: Universal Backdoor Attacks on Pre-trained Language Models

2023-05-16 · Wei Du, Peixuan Li, Boqun Li, Haodong Zhao 외

Backdoors implanted in pre-trained language models (PLMs) can be transferred to various downstream tasks, which exposes a severe security threat. However, most existing backdoor attacks against PLMs are un-targeted and t…

Backdoor AttackContrastive Learningtext-classificationText Classification

Inevitable Encounters: Backdoor Attacks Involving Lossy Compression

2026-03-14 · Qian Li, Yunuo Chen, Yuntian Chen arxiv

Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before being used to compromise deep learning systems. However, in the era of big data, the inevitable use of lossy compression pos…

Image Compression