Xuan Xu
I am a doctoral student in the School of Cyberspace Security at Beijing University of Posts and Telecommunications (BUPT). I am enrolled in the direct Ph.D. program and expect to graduate in June 2027. My advisors are Prof. Linna Zhou and Prof. Zhongliang Yang.
My research broadly spans language models, topic modeling, semantic retrieval, and AI safety & ethics. I am particularly interested in:
- practical and efficient deployment of large language models in domain-specific settings;
- intersections between topic modeling and large language models;
- LLM interpretability and safety; and
- agent product design and industry research.
Education
| Period | Institution | Program |
|---|---|---|
| 2022.09 – 2027.06 | Beijing University of Posts and Telecommunications, School of Cyberspace Security | Direct Ph.D. program; expected graduation: June 2027 |
| 2018.09 – 2022.06 | Nanjing University of Posts and Telecommunications, School of Electronic and Optical Engineering | B.Eng. |
Selected First-Author Publications
-
Xuan Xu, Haolun Li, et al. Towards Modern Topic Models: A Survey of Taxonomies and Paradigm Shifts from Algorithm-Centric to LLM-Centered Topic Analysis. Findings of ACL 2026. [Paper] [PDF] [Code]
-
Xuan Xu, et al. FinCPRG: A Bidirectional Generation Pipeline for Hierarchical Queries and Rich Relevance in Financial Chinese Passage Retrieval. ECML-PKDD 2025. [arXiv] [Code] [Dataset]
-
Xuan Xu, Fufang Wen, et al. FinBERT2: A Specialized Bidirectional Encoder for Bridging the Gap in Finance-Specific Deployment of Large Language Models. KDD 2025. [DOI] [arXiv] [Code]
-
Xu, Xuan, et al. VStego800K: Large-Scale Steganalysis Dataset for Streaming Voice. IWDW 2023. [DOI] [Code]
Contact
- Email: sh22xuxuan@bupt.edu.cn
- ORCID: 0009-0007-1326-4282