苏涛宇 Taoyu Su

I am Taoyu Su. I received my Ph.D. from the Institute of Information Engineering, Chinese Academy of Sciences and the University of Chinese Academy of Sciences, Beijing, China. I am advised by Prof. Tingwen Liu.

苏涛宇，博士毕业于中国科学院信息工程研究所，导师为柳厅文研究员，当前研究方向包括多模态实体对齐、智能体应用安全、社会网络分析。作为项目研发负责人参与国家重点研发计划（青年科学家）、某部委网络分析系统等项目建设。

Email Google Scholar GitHub

News

[2026-01] One Corresponding-author oral paper and One co-author paper accepted to WWW 2026 (CCF-A).

[2025-11] One Corresponding-author oral paper accepted to AAAI 2026 (CCF-A, oral).

[2025-07] One Corresponding-author paper accepted to NLPCC 2025 (CCF-C).

[2025-05] One co-author paper accepted to ICML 2025 (CCF-A), thanks to all co-authors!

[2025-04] One first-author paper accepted to SIGIR 2025 (CCF-A), thanks to all co-authors!

[2024-07] One first-author paper accepted to ACM MM 2024 (CCF-A), thanks to all co-authors!

[2024-07] One first-author paper accepted to ECAI 2024 (CCF-B), thanks to all co-authors!

Research Highlights

Interested in the Multi-modal Entity Alignment, Agentic AI Security, Social Network Analysis.

研究方向包括多模态实体对齐、智能体应用安全、社会网络分析。

Selected Publications
	Conditional Diffusion Guided Knowledge Transfer for Multi-Domain Knowledge Graph Completion Jiawei Sheng, Taoyu Su, Xixun Lin, Xiaodong Li, Tingwen Liu WWW (CCF-A, Oral, 317/3370=9.4%)*, 2026 paper code DMKGC pioneers a generation-based MKGC approach using conditional diffusion to transfer knowledge across KGs without suppressing domain-specific information.
	Information-Theoretic Minimal Sufficient Representation for Multi-Domain Knowledge Graph Completion Jiawei Sheng, Taoyu Su, Weiyi Yang, Linghui Wang, Yongxiu Xu, Tingwen Liu AAAI (CCF-A, Oral, 529/23680=2.2%)*, 2026 paper code IMKGC proposes an information-theoretic framework that learns minimal sufficient representations by suppressing redundant information for multi-domain knowledge graph completion.
	Unlocking the Power of Large Language Models for Multi-table Entity Matching Yingkai Tang, Taoyu Su, Wenyuan Zhang, Xiaoyang Guo, Tingwen Liu NLPCC (CCF-C)*, 2025 paper arxiv code LLM4MEM leverages Large Language Models to resolve semantic inconsistencies and efficiency bottlenecks in multi-table entity matching.
	Mitigating Modality Bias in Multi-modal Entity Alignment from a Causal Perspective Taoyu Su, Jiawei Sheng, Duohe Ma, Xiaodong Li, Juwei Yue, Mengxiao Song, Yingkai Tang, Tingwen Liu SIGIR (CCF-A, Oral), 2025 paper arxiv code CDMEA mitigates visual modality bias in multi-modal entity alignment from a causal perspective by suppressing the direct effect of visual features while leveraging both visual and graph modalities.
	IBMEA: Exploring Variational Information Bottleneck for Multi-modal Entity Alignment Taoyu Su, Jiawei Sheng, Shicheng Wang, Xinghua Zhang, Hongbo Xu, Tingwen Liu ACM MM (CCF-A), 2024 paper arxiv code IBMEA emphasizes the alignment-relevant information and suppress the alignment-irrelevant information in generating entity representations.
	LoginMEA: Local-to-Global Interaction Network for Multi-modal Entity Alignment Taoyu Su, Xinghua Zhang, Jiawei Sheng, Zhenyu Zhang, Tingwen Liu ECAI (CCF-B), 2024 paper arxiv code LoginMEA fuses local multi-modal interactions to generate holistic entity semantics and then refine them with global relational interactions of entity neighbors.

Services

Conference Program Committee / Reviewer
- ACMMM 2026/2025/2024
- WWW 2026/2025
- KDD 2026
- CIKM 2026
- ECML-PKDD 2026
- ECAI 2025
Journal Reviewer
- IEEE TKDD
- IEEE TMM
- PeerJ

Mentorship & Supervision

I am honored to mentor the following talented students.

已毕业
- 张同学（24届硕士）指导发表期刊论文1篇，毕业论文撰写。毕业时加入比亚迪公司，从事大模型应用研发
- 林同学（25届硕士）指导发表CCF-C类论文1篇，毕业论文撰写。毕业时加入上海电气集团，从事AI产品管理
- 唐同学（26届硕士）指导发表CCF-C类论文1篇，毕业论文撰写。毕业时加入华为公司（成都），从事AI应用研发
在学
- 张同学（25级硕士）正在研究多模态跨域推荐
- 付同学（25级硕士）正在研究多模态实体对齐

Research Aspiration

以屈原《离骚》名句自勉：科研之路漫长而修远，愿上下求索、笃行不怠。