paper-with-me

Papers

BadCLIP: Dual-Embedding Guided Backdoor Attack on Multimodal Contrastive Learning

2023-11-20 · CVPR 2024 1 · Siyuan Liang, Mingli Zhu, Aishan Liu, Baoyuan Wu, Xiaochun Cao, Ee-Chien Chang

Studying backdoor attacks is valuable for model copyright protection and enhancing defenses. While existing backdoor attacks have successfully infected multimodal contrastive learning models such as CLIP, they can be easily countered by specialized backdoor defenses for MCL models. This paper reveals the threats in this practical scenario that backdoor attacks can remain effective even after defenses and introduces the \emph{\toolns} attack, which is resistant to backdoor detection and model fine-tuning defenses. To achieve this, we draw motivations from the perspective of the Bayesian rule and propose a dual-embedding guided framework for backdoor attacks. Specifically, we ensure that visual trigger patterns approximate the textual target semantics in the embedding space, making it challenging to detect the subtle parameter variations induced by backdoor learning on such natural trigger patterns. Additionally, we optimize the visual trigger patterns to align the poisoned samples with target vision features in order to hinder the backdoor unlearning through clean fine-tuning. Extensive experiments demonstrate that our attack significantly outperforms state-of-the-art baselines (+45.3% ASR) in the presence of SoTA backdoor defenses, rendering these mitigation and detection strategies virtually ineffective. Furthermore, our approach effectively attacks some more rigorous scenarios like downstream tasks. We believe that this paper raises awareness regarding the potential threats associated with the practical application of multimodal contrastive learning and encourages the development of more robust defense mechanisms.

📄 PDF Abstract BibTeX arXiv:2311.12075

Code (1)

LiangSiyuan21/BadCLIP pytorch

Tasks

Backdoor AttackContrastive Learning

Methods 이 논문이 사용한 방법론

CLIP Contrastive Language-Image Pre-training (CLIP), consisting of a simplified version of ConVIRT trained from scratch, is an efficient method of image representation learning…
ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…
Contrastive Learning 설명 없음

Similar Papers 제목 키워드 기반

BadCLIP: Trigger-Aware Prompt Learning for Backdoor Attacks on CLIP

2023-11-26 · CVPR 2024 1 · Jiawang Bai, Kuofeng Gao, Shaobo Min, Shu-Tao Xia 외

Contrastive Vision-Language Pre-training, known as CLIP, has shown promising effectiveness in addressing downstream image recognition tasks. However, recent works revealed that the CLIP model can be implanted with a down…

Prompt Learning

BadCLIP++: Stealthy and Persistent Backdoors in Multimodal Contrastive Learning

2026-02-19 · Siyuan Liang, Yongcheng Jing, Yingjie Wang, Jiaxing Huang 외 arxiv

Research on backdoor attacks against multimodal contrastive learning models faces two key challenges: stealthiness and persistence. Existing methods often fail under strong detection or continuous fine-tuning, largely du…

Contrastive Learning

Adversarial Backdoor Defense in CLIP

2024-09-24 · Junhao Kuang, Siyuan Liang, Jiawei Liang, Kuanrong Liu 외

Multimodal contrastive pretraining, exemplified by models like CLIP, has been found to be vulnerable to backdoor attacks. While current backdoor defense methods primarily employ conventional data augmentation to create a…

backdoor defenseData Augmentation

CleanerCLIP: Fine-grained Counterfactual Semantic Augmentation for Backdoor Defense in Contrastive Learning

2024-09-26 · Yuan Xun, Siyuan Liang, Xiaojun Jia, Xinwei Liu 외

Pre-trained large models for multimodal contrastive learning, such as CLIP, have been widely recognized in the industry as highly susceptible to data-poisoned backdoor attacks. This poses significant risks to downstream …

backdoor defenseContrastive LearningcounterfactualText Augmentation+2

Defending Our Privacy With Backdoors

2023-10-12 · Dominik Hintersdorf, Lukas Struppek, Daniel Neider, Kristian Kersting

The proliferation of large AI models trained on uncurated, often sensitive web-scraped data has raised significant privacy concerns. One of the concerns is that adversaries can extract information about the training data…