publications

2026

1. BEIHAI

   

   TIMO-OSS-Search-Preview: A Strong Lightweight Agentic Model for Medical Deep Research

   Zekun Li, Zijiang Yang, Qiran Jiang, Zitai Wang, Xin Xuan Xu Liu, Shuo Wang, Weizhi Ma, and Yang Liu

   In The 2nd Beijing Tsinghua Medical AI Forum (BEIHAI), Oct 2026

   A 4B search agent, built with evidence-driven data synthesis and model–harness co-design, beats much larger models on medical deep research.

   Oral Bib

   Selected for oral presentation and poster at BEIHAI 2026.

   @inproceedings{li2026timo,
     title = {TIMO-OSS-Search-Preview: A Strong Lightweight Agentic Model for Medical Deep Research},
     author = {Li, Zekun and Yang, Zijiang and Jiang, Qiran and Wang, Zitai and Liu, Xin Xuan Xu and Wang, Shuo and Ma, Weizhi and Liu, Yang},
     booktitle = {The 2nd Beijing Tsinghua Medical AI Forum (BEIHAI)},
     year = {2026},
     month = oct,
     tldr = {A 4B search agent, built with evidence-driven data synthesis and model–harness co-design, beats much larger models on medical deep research.},
   }

2. ACL

   

   Beyond Self-Report: Bridging the Intention-Behavior Gap in Critical Thinking Assessment via Interpretable Multi-Agent System

   Zekun Li, Jifan Yu, Haoxuan Li, Ye He, Daniel Zhang-Li, Shangqing Tu, Joy Jia Yin Lim, Yikun Jiang, Jiaxin Yuan, and Yu Zhang

   In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2026

   Replaces self-report questionnaires with multi-agent scenario interviews, matching expert critical-thinking ratings better at $0.41 per person.

   DOI Bib HTML PDF

   @inproceedings{li2026beyond,
     title = {Beyond Self-Report: Bridging the Intention-Behavior Gap in Critical Thinking Assessment via Interpretable Multi-Agent System},
     author = {Li, Zekun and Yu, Jifan and Li, Haoxuan and He, Ye and Zhang-Li, Daniel and Tu, Shangqing and Lim, Joy Jia Yin and Jiang, Yikun and Yuan, Jiaxin and Zhang, Yu},
     booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
     pages = {26849--26871},
     year = {2026},
     doi = {10.18653/v1/2026.acl-long.1236},
     tldr = {Replaces self-report questionnaires with multi-agent scenario interviews, matching expert critical-thinking ratings better at $0.41 per person.},
   }

3. AAAI-W

   

   MASA: Multi-Agent Guided Interview for Scenario-Based Assessment of Critical Thinking

   Zekun Li, Song Li, Ye He, Haoxuan Li, Yikun Jiang, Joy Jia Yin Lim, Daniel Zhang-Li, Shangqing Tu, Yucheng Wang, Yu Zhang, and Jifan Yu

   In AAAI 2026 Workshop on AI for Education (AI4EDU), 2026

   Oral session, Singapore

   Four cooperating agents interview students in micro-scenarios and score critical-thinking dispositions with an auditable assessment chain-of-thought.

   Bib PDF Website

   @inproceedings{li2026masa,
     title = {MASA: Multi-Agent Guided Interview for Scenario-Based Assessment of Critical Thinking},
     author = {Li, Zekun and Li, Song and He, Ye and Li, Haoxuan and Jiang, Yikun and Lim, Joy Jia Yin and Zhang-Li, Daniel and Tu, Shangqing and Wang, Yucheng and Zhang, Yu and Yu, Jifan},
     booktitle = {AAAI 2026 Workshop on AI for Education (AI4EDU)},
     year = {2026},
     note = {Oral session, Singapore},
     tldr = {Four cooperating agents interview students in micro-scenarios and score critical-thinking dispositions with an auditable assessment chain-of-thought.},
   }

4. ACL

   

   SimPBL: A Multi-Agent Framework for Project-Based Learning

   Daniel Zhang-Li^*, Joy Jia Yin Lim^*, Binglin Liu, Shangqing Tu, Zijun Yao, Hao Peng, Jifan Yu, Haoxuan Li, Zhanxin Hao, Ye He, Zekun Li, Jiangyi Wang, Lei Hou, Bin Xu, Xin Cong, Zhiyuan Liu, Huiqin Liu, Yu Zhang, and Juanzi Li

   In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 2026

   An orchestrator agent with boundary-aware collaborator agents scaffolds project-based learning, raising exam scores by 14%.

   DOI Bib HTML PDF Code

   @inproceedings{zhangli2026simpbl,
     title = {SimPBL: A Multi-Agent Framework for Project-Based Learning},
     author = {Zhang-Li, Daniel and Lim, Joy Jia Yin and Liu, Binglin and Tu, Shangqing and Yao, Zijun and Peng, Hao and Yu, Jifan and Li, Haoxuan and Hao, Zhanxin and He, Ye and Li, Zekun and Wang, Jiangyi and Hou, Lei and Xu, Bin and Cong, Xin and Liu, Zhiyuan and Liu, Huiqin and Zhang, Yu and Li, Juanzi},
     booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
     pages = {32755--32772},
     year = {2026},
     doi = {10.18653/v1/2026.acl-long.1512},
     tldr = {An orchestrator agent with boundary-aware collaborator agents scaffolds project-based learning, raising exam scores by 14%.},
   }

5. JCST

   

   From MOOC to MAIC: Reimagine Online Teaching and Learning Through LLM-Driven Agents

   Jifan Yu, Daniel Zhang-Li, Zheyuan Zhang, Yucheng Wang, Haoxuan Li, Joy Jia Yin Lim, Zhanxin Hao, Shangqing Tu, Lu Zhang, Xusheng Dai, Jianxiao Jiang, Shen Yang, Fei Qin, Zekun Li, Binglin Liu, Xin Cong, Bin Xu, Lei Hou, Manli Li, Juanzi Li, Huiqin Liu, Yu Zhang, Zhiyuan Liu, and Maosong Sun

   Journal of Computer Science and Technology, 2026

   LLM-driven multi-agent classrooms make online courses both scalable and adaptive, deployed with 500+ Tsinghua students.

   DOI Bib PDF Website

   @article{yu2026mooc,
     title = {From MOOC to MAIC: Reimagine Online Teaching and Learning Through LLM-Driven Agents},
     author = {Yu, Jifan and Zhang-Li, Daniel and Zhang, Zheyuan and Wang, Yucheng and Li, Haoxuan and Lim, Joy Jia Yin and Hao, Zhanxin and Tu, Shangqing and Zhang, Lu and Dai, Xusheng and Jiang, Jianxiao and Yang, Shen and Qin, Fei and Li, Zekun and Liu, Binglin and Cong, Xin and Xu, Bin and Hou, Lei and Li, Manli and Li, Juanzi and Liu, Huiqin and Zhang, Yu and Liu, Zhiyuan and Sun, Maosong},
     journal = {Journal of Computer Science and Technology},
     volume = {41},
     number = {1},
     pages = {394--414},
     year = {2026},
     doi = {10.1007/s11390-025-6000-0},
     tldr = {LLM-driven multi-agent classrooms make online courses both scalable and adaptive, deployed with 500+ Tsinghua students.},
   }