analiza aiqbench
This research addresses the critical challenge of domain-specific limitations in defending Graph Neural Networks against adversarial attacks by introducing ProGAP, a transferable purification scheme. By leveraging vulnerability-aware graph prompt learning, the method effectively bridges defense knowledge across different data distributions without requiring extensive retraining or parameter updates. The approach demonstrates significant improvements in robustness and efficiency, offering a practical solution for securing complex relational data structures in multimedia tasks.
Let Prompts Bridge Defense Knowledge: Transferable Graph Purification via Vulnerability-Aware GPL
arXiv:2608.29054v1 Announce Type: new Abstract: Graph Neural Networks (GNNs) have emerged as a cornerstone for representing complex relational dependencies in diverse multimedia tasks, particularly in cross-platform user interest modeling and cross-modal semantic alignment. In the real world, a practical defense against graph adversarial perturbations is needed. However, we observe that the prevailing adversarial purification methods are essentially domain-restricted defenses, which leads to the following shortcomings: (1) single-domain data provides insufficient structural and semantic diversity for learning robust purification criteria; (2) training of domain-specific defense strategies from scratch consumes substantial computational cost. To address the above limitations, we propose a transferable graph purification scheme, named ProGAP, to bridge adversarial defense knowledge via vulnerability-aware graph prompt learning. Firstly, to capture universal adversarial patterns, a perturbation-capture edge detector is pretrained on data-rich graphs by jointly modeling topological and semantic information. Subsequently, to achieve more knowledge transfer w.r.t. robustness, vulnerability-aware prompts are designed that inject targeted purification guidance into biased nodes, during which the pretrained detector adapts to distribution shifts in downstream graphs without parameter-laborious updates. Experimental results demonstrate that compared with state-of-the-art baselines, our ProGAP achieves 1%-9% improvement, and reduces the time consumption by up to 2.2x. The code for ProGAP is available at https://github.com/Lieyoufffff/ProGAP.
Abstract:Graph Neural Networks (GNNs) have emerged as a cornerstone for representing complex relational dependencies in diverse multimedia tasks, particularly in cross-platform user interest modeling and cross-modal semantic alignment. In the real world, a practical defense against graph adversarial perturbations is needed. However, we observe that the prevailing adversarial purification methods are essentially domain-restricted defenses, which leads to the following shortcomings: (1) single-domain data provides insufficient structural and semantic diversity for learning robust purification criteria; (2) training of domain-specific defense strategies from scratch consumes substantial computational cost. To address the above limitations, we propose a transferable graph purification scheme, named ProGAP, to bridge adversarial defense knowledge via vulnerability-aware graph prompt learning. Firstly, to capture universal adversarial patterns, a perturbation-capture edge detector is pretrained on data-rich graphs by jointly modeling topological and semantic information. Subsequently, to achieve more knowledge transfer w.r.t. robustness, vulnerability-aware prompts are designed that inject targeted purification guidance into biased nodes, during which the pretrained detector adapts to distribution shifts in downstream graphs without parameter-laborious updates. Experimental results demonstrate that compared with state-of-the-art baselines, our ProGAP achieves 1%-9% improvement, and reduces the time consumption by up to 2.2x. The code for ProGAP is available at this https URL.
To appear in the Proceedings of the 34th ACM International Conference on Multimedia (MM '26). 10 pages, 7 figures, and 4 tables. Shuomin Xue and Jingyuan Li contributed equally to this work
Artificial Intelligence (cs.AI)
arXiv:2608.29054 [cs.AI]
(or arXiv:2608.29054v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.29054 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 249 cuvinte
