Wu Xin is currently an Assistant Professor at the School of Software, South China University of Technology. Her main research focuses on natural language processing and large models, with in-depth exploration around logic-enhanced semantic modeling and content generation. Specifically: (1) She proposed the structure-aware generation model SLEtoNL, which significantly improves the semantic mapping accuracy from logical expressions to natural language through the introduction of tree-graph structural encoding mechanisms (IEEE TASLP); (2) She proposed the Logic Control Framework (LCF), achieving explicit control over the logical validity of generated content via contrastive learning (AAAI 2025); (3) She designed the Abstract Layer Reasoning model (ADR), effectively enhancing the stability of generation models (COLING 2024); (4) She constructed a collaborative mechanism between symbolic reasoning and large models, enabling logical consistency generation for time-sensitive knowledge (EMNLP 2024). In the past five years, she has published more than 20 international papers in top conferences and journals such as AAAI, ACL, EMNLP, and IEEE TASLP. She holds 3 authorized invention patents and has won awards such as the Runner-Up for Best Paper at the CCF-recommended conference APWeb-WAIM 2021, Second Prize in the National Technical Evaluation of the Chinese College Computer Competition Big Data Challenge, and CCF BDCI. As a core member, she participated in the intelligent beautification function of WPS Presentation, which covers over 38 million users and creates economic benefits of approximately 3.343 billion yuan. She also participated as a core member in Huawei's \Challenge Hard Problems\ project, which was awarded the "Spark Award".
E-mail: xinwu@scut.edu.cn
Research Direction:Natural Language Processing, Large Model Logic Enhancement, Fast and Slow Thinking Reasoning, Knowledge Editing
Representative Papers
1. Wu, Xin, Yuqi Bu, Yi Cai, Yifei Chen, and Ho-Fung Leung. Type-agnostic and form-oriented deductive conclusion generation. Neural Networks (2025): 107968.
2. Bu, Yuqi, Xin Wu, Zirui Zhao, Yi Cai, David Hsu, and Qiong Liu. Walk in Others’ Shoes with a Single Glance: Human-Centric Visual Grounding with Top-View Perspective Transformation. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 26904-26923. 2025.
3. Guo, Sile, Hongfei Liu, Xin Wu, and Yi Cai. Retrieval-Augmented Classification for Financial User Profiling: A Lightweight Approach. In 2025 IEEE 41st International Conference on Data Engineering Workshops (ICDEW), pp. 259-265. IEEE, 2025.
4. Gong, Songwen, Hongfei Liu, Xin Wu, and Yi Cai. RBRTI: Retrieval-Based Risk Type Identification from Financial News. In 2025 IEEE 41st International Conference on Data Engineering Workshops (ICDEW), pp. 246-252. IEEE, 2025.
5. Huang, Hailin, Hongfei Liu, Xin Wu, and Yi Cai. Risk-Aware Automatic Text Summarization with Pre-Identification of Risk Categories and Emphasis. In 2025 IEEE 41st International Conference on Data Engineering Workshops (ICDEW), pp. 239-245. IEEE, 2025.
