CVPR Findings2026
BadRSSD: Backdoor Attacks on Regularized Self-Supervised Diffusion Models
Jiayao Wang, Yiping Zhang, Mohammad Maruf Hasan, Xiaoying Lei, Jiale Zhang, Junwu Zhu, Qilin Wu, Dongfang Zhao
Abstract
Self-supervised diffusion models learn high-quality visual representations via latent space denoising. However, their representation layer poses a distinct threat: unlike traditional attacks targeting generative outputs, its unconstrained latent semantic space allows for stealthy backdoors, permitting malicious control upon triggering. In this paper, we propose BadRSSD, the first backdoor attack targeting the representation layer of self-supervised diffusion models. Specifically, it hijacks the semantic representations of poisoned samples with triggers in Principal Component Analysis (PCA) space toward those of a target image, then controls the denoising trajectory during diffusion by applying coordinated constraints across latent, pixel, and feature distribution spaces to steer the model toward generating the specified target. Additionally, we integrate representation dispersion regularization into the constraint framework to maintain feature space uniformity, significantly enhancing attack stealth. This approach preserves normal model functionality (high utility) while achieving precise target generation upon trigger activation (high specificity). Experiments on multiple benchmark datasets demonstrate that BadRSSD substantially outperforms existing attacks in both FID and MSE metrics, reliably establishing backdoors across different architectures and configurations, and effectively resisting state-of-the-art backdoor defenses.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
- GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models cites An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models cites Denoising Diffusion Probabilistic Models
- Scalable Diffusion Models with Transformers cites GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models
- Scalable Diffusion Models with Transformers cites An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
- Scalable Diffusion Models with Transformers cites Denoising Diffusion Probabilistic Models
- Directly Denoising Diffusion Models cites GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models
- Directly Denoising Diffusion Models cites Denoising Diffusion Probabilistic Models
Builds on22
- Denoising Diffusion Probabilistic ModelsJonathan Ho, Ajay Jain, Pieter AbbeelNeurIPS 2020 · 35,902 citations
- An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleAlexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn et al.ICLR 2021 · 21,477 citations
- Directly Denoising Diffusion ModelsDan Zhang, Jingjing Wang, Feng LuoICML 2024 · 11,724 citations
- Scalable Diffusion Models with TransformersWilliam Peebles, Saining XieICCV 2023 · 5,568 citations
- GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion ModelsAlexander Quinn Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav Shyam et al.ICML 2022 · 4,691 citations
Related papers
- Semantic-level Backdoor Attack against Text-to-Image Diffusion ModelsTianxin Chen, Wenbo Jiang, Hongqiao Chen, Zhirun Zheng et al.ICML 2026 · 1 citation
- Text-to-Image Diffusion Models can be Easily Backdoored through Multimodal Data PoisoningShengfang Zhai, Yinpeng Dong, Qingni Shen, Shi Pu et al.ACM MM 2023 · 46 citations
- Distribution Preserving Backdoor Attack in Self-supervised LearningGuanhong Tao, Zhenting Wang, Shiwei Feng, Guangyu Shen et al.S&P 2024 · 32 citations
- Invisible Backdoor Attack against Self-supervised LearningHanrong Zhang, Zhenting Wang, Boheng Li, Fulin Lin et al.CVPR 2025
- TERD: A Unified Framework for Safeguarding Diffusion Models Against BackdoorsYichuan Mo, Hui Huang, Mingjie Li, Ang Li et al.ICML 2024 · 31 citations
