Multimodal Representation Learning
Representation, alignment, editing, and computation methods for heterogeneous multimodal data in recommendation and knowledge-rich information systems.
Ph.D. Candidate, Shanghai Jiao Tong University
Gu Tang
I work on multimodal representation learning, recommender systems, and vertical foundation models for information systems.
I am a Ph.D. candidate in Electronic Information at Shanghai Jiao Tong University, advised by Prof. Xiaoying Gan and working with Prof. Xinbing Wang's team. My research studies how heterogeneous multimodal data can be represented, aligned, edited, and computed for trustworthy recommendation and data-driven knowledge services.
3
CCF-A conference papers as independent first author
1
CAS Q1 / Top journal paper as independent first author
2
Major research projects with student-lead responsibilities
Research
Learning from heterogeneous multimodal data
Multimodal Domain Foundation Models
Foundation model construction and computational methods tailored to vertical information systems, with emphasis on data quality, modality utility, and controllable reasoning.
Data-Literature Knowledge Services
Knowledge graphs, multimodal parsing, and large-model question answering for scientific data discovery, understanding, and reuse.
Education
Academic Training
-
Shanghai Jiao Tong University
Ph.D. in Electronic Information, expected June 2027.
Advisor: Prof. Xiaoying Gan. Research team: Prof. Xinbing Wang's group.
-
Chongqing University of Technology
M.Eng. in Computer Technology.
Advisor: Prof. Xiaofei Zhu. Honors include National Scholarship, Outstanding Master's Graduate of Chongqing, and Top Ten Student of the university.
-
Chongqing University of Technology
B.S. in Information and Computing Science.
Honors include Second Prize in Contemporary Undergraduate Mathematical Contest in Modeling and First-class Academic Scholarship.
Publications
Selected Papers
TargetMR: Learning Modality Target for Multimodal Recommendation
Gu Tang, Xiaoying Gan, Luoyi Fu, Xinbing Wang, et al.
R2MR: Review and Rewrite Modality for Recommendation
Gu Tang, Xiaoying Gan, Luoyi Fu, Xinbing Wang, et al.
EditKG: Editing Knowledge Graph for Recommendation
Gu Tang, Xiaoying Gan, Luoyi Fu, Xinbing Wang, et al.
Time Enhanced Graph Neural Networks for Session-based Recommendation
Gu Tang, Xiaofei Zhu, Jiafeng Guo, et al.
SeqEditor: Editing User Behavior Sequence for Recommendation
Gu Tang, Xiaoying Gan, Luoyi Fu, Xinbing Wang, et al. Submitted to NeurIPS 2026.
CleanMR: Cleaning Redundancy for Multimodal Recommendation
Gu Tang, Xiaoying Gan, Luoyi Fu, Xinbing Wang, et al. Submitted to ACM Transactions on Information Systems.
Projects and Platforms
Research Systems
Spatiotemporal Knowledge Retrieval and Computational Reasoning
Project No. 2022YFB3904204, RMB 3.2M, Dec. 2022 - Nov. 2025. Served as the first student lead for a completed national-level project.
Mountain Disaster Risk Warning Platform
Application demonstration project CAS-WX2021SF-0106, RMB 300K. Served as student lead for platform construction and application demonstration.
DataExpo
A one-stop data navigation system integrating million-scale multimodal geoscience data. DataExpo has supported data-driven research for major programs such as Deep-time Digital Earth and more than ten domestic and international institutions.
Visit DataExpoCNS Data Bulletin
A data-literature integrated knowledge service system focused on high-quality research from CNS and related venues. It combines multimodal knowledge parsing, knowledge graphs, and large-model question answering to help researchers discover, understand, and reuse scientific data.
Visit CNS Data BulletinCompetitions
Selected Awards
- 2025 Tencent Advertising Algorithm Competition Full-modal generative recommendation, Rank 52 / 2800, top 2%.
- 2022 Global Artificial Intelligence Algorithm Challenge Rank 12 / 1473, top 1%.
- 2021 Huawei DIGIX Global Campus AI Algorithm Elite Competition Rank 4 / 4017, top 1%.
- 2021 WeChat Big Data Challenge Second Prize nationwide, Rank 21 / 6054, top 1%.
- 2021 Meituan Waimai Recommendation Technology Evaluation Rank 6 / 408, top 2%.
- 2021 Huawei Cup China Graduate Mathematical Contest in Modeling Second Prize nationwide.