AAAI 2026 | 图基础模型(GFM)&文本属性图(TAG)论文总结

时空探索之旅

发布于 2026-03-10 14:47:44

1070

AAAI 2026将在2026年1月20日到1月27日于新加坡（Singapore）举行。AAAI 2026会议主会共有23, 680篇论文投稿，其中4, 167 篇被接收，接收率为17.6%。

还是尝试拓展一下知识面，试着总结了其中图基础模型（graph foundation model, GFM）和文本图（text-attributed graph, TAG）高分的论文。如有疏漏，欢迎大家补充。

1. Unlocking Multi-Modal Potentials for Link Prediction on Dynamic Text-Attributed Graphs2. Towards Robust Text-Attributed Federated Graph Learning: Multimodal Threats and Defense3. THGB: A Comprehensive Benchmark for Text-attributed Heterogeneous Graphs4. LLM-Enhanced Energy Contrastive Learning for Out-ofDistribution Detection in Text-Attributed Graphs5. GCL-OT: Graph Contrastive Learning with Optimal Transport for Heterophilic Text-Attributed Graphs6. Towards Effective, Stealthy, and Persistent Backdoor Attacks Targeting Graph Foundation Models7. Source-Free Graph Foundation Model Adaptation via Pseudo-Source Reconstruction8. SA?GFM: Enhancing Robust Graph Foundation Models with Structure-Aware Semantic Augmentation

1 Unlocking Multi-Modal Potentials for Link Prediction on Dynamic Text-Attributed Graphs

链接：https://arxiv.org/abs/2502.19651

作者：Yuanyuan Xu; Wenjie Zhang; Ying Zhang; Xuemin Lin; Xiwei Xu

关键词：动态文本图，多模态，链接预测

2 Towards Robust Text-Attributed Federated Graph Learning: Multimodal Threats and Defense

作者：Zitong Shi; Guancheng Wan; Wenke Huang; Yuxin Wu; Quan Zhang; Mang Ye

关键词：联邦图学习，稳健性

3 THGB: A Comprehensive Benchmark for Text-attributed Heterogeneous Graphs

链接：https://github.com/xxxinxxx/THGB

作者：Lixin Zhou; Zemin Liu; Yuan Fang; Dan Niu; Jing Ying

关键词：异质图，benchmark

4 LLM-Enhanced Energy Contrastive Learning for Out-ofDistribution Detection in Text-Attributed Graphs

作者：Xiaoxu Ma; Dong Li; Minglai Shao; Xintao Wu; Chen Zhao

关键词：分布外检测，LLM增强

5 GCL-OT: Graph Contrastive Learning with Optimal Transport for Heterophilic Text-Attributed Graphs

链接：https://arxiv.org/abs/2511.16778

作者：Yating Ren; Yikun Ban; Huobin Tan

关键词：最优传输，对比学习，异质图

6 Towards Effective, Stealthy, and Persistent Backdoor Attacks Targeting Graph Foundation Models

链接：https://www.arxiv.org/abs/2511.17982

作者：Jiayi Luo; Qingyun Sun; Lingjuan Lyu; Ziwei Zhang; Haonan Yuan; Xingcheng Fu; Jianxin Li

关键词：图基础模型，后门攻击

7 Source-Free Graph Foundation Model Adaptation via Pseudo-Source Reconstruction

作者：Liang Yang; Hui Ning; Jiaming Zhuo; Ziyi Ma; Chuan Wang; Wenning Wu; Zhen Wang

关键词：图基础模型，适配

58 SGFM: Enhancing Robust Graph Foundation Models with Structure-Aware Semantic Augmentation

作者：Junhua Shi; Qingyun Sun; Haonan Yuan; Xingcheng Fu

关键词：稳健图基础模型，结构感知语义增强