BadGraph: A Backdoor Attack Against Latent Diffusion Model for Text-Guided Graph Generation
This work addresses a critical security problem for applications like drug discovery by exposing backdoor risks in latent diffusion models for graph generation, though it is incremental as it extends prior backdoor attack research to a new domain.
The paper tackles the security vulnerability of backdoor attacks in text-guided graph generation by proposing BadGraph, a method that uses textual triggers to poison training data, achieving over 80% attack success rate with a 24% poisoning rate while maintaining normal performance on clean inputs.
The rapid progress of graph generation has raised new security concerns, particularly regarding backdoor vulnerabilities. While prior work has explored backdoor attacks in image diffusion and unconditional graph generation, conditional, especially text-guided graph generation remains largely unexamined. This paper proposes BadGraph, a backdoor attack method against latent diffusion models for text-guided graph generation. BadGraph leverages textual triggers to poison training data, covertly implanting backdoors that induce attacker-specified subgraphs during inference when triggers appear, while preserving normal performance on clean inputs. Extensive experiments on four benchmark datasets (PubChem, ChEBI-20, PCDes, MoMu) demonstrate the effectiveness and stealth of the attack: less than 10% poisoning rate can achieves 50% attack success rate, while 24% suffices for over 80% success rate, with negligible performance degradation on benign samples. Ablation studies further reveal that the backdoor is implanted during VAE and diffusion training rather than pretraining. These findings reveal the security vulnerabilities in latent diffusion models of text-guided graph generation, highlight the serious risks in models' applications such as drug discovery and underscore the need for robust defenses against the backdoor attack in such diffusion models.