Hang Li

Hang Li

Ph.D. Candidate, Michigan State University
LLMs · Multi-Agent Systems · AI for Education

Hang Li

I am a Ph.D. candidate in Computer Science at Michigan State University, advised by Prof. Jiliang Tang in the Data Science and Engineering (DSE) Lab. I also work closely with Prof. Hui Liu. My research is on large language models (LLMs) and LLM-based multi-agent systems, with a focus on making them reliable for education. I study how LLMs reason and where they fail, and I build systems for automated grading, error diagnosis and teacher professional development.

I have published in venues including ACL, EMNLP, AAAI/IAAI, IJCAI, WWW, TMLR, ICASSP, INTERSPEECH, AIED and EDM, and my work has been cited more than 2,000 times (Google Scholar). Our work on LLM-based knowledge tagging received the IAAI 2025 Innovative Application Award. I helped design and write three grant proposals with Prof. Tang: two funded by the NSF and one by the Gates Foundation.

I was a research intern at Capital One (2026) and Squirrel AI Learning (2024). Before my Ph.D., I was a Machine Learning Scientist at TAL Education Group, building multimodal learning systems for online classrooms. I hold an M.S. in Statistics from the University of Illinois Urbana-Champaign and a B.S. in Information and Computing Science from Beijing Jiaotong University.

lihang4 [at] msu [dot] edu  ·  CV (PDF)

Research Interests

Reliable LLMs for Assessment

LLM-based automated grading that is accurate, calibrated and robust: rubric and demonstration optimization, uncertainty quantification, human-in-the-loop grading, and prompt injection attacks on graders.

LLM Reasoning & Multi-Agent Systems

How LLMs solve problems and detect errors: solution divergence, conformity bias and prompt optimization for long-context reasoning. Multi-agent systems for knowledge tagging and teacher professional development.

Machine Learning on Graphs

Diffusion models on graphs, LLMs for learning on graphs, graph representation learning, and graph-based retrieval and generation.

Multimodal Learning

Audio-and-language pre-training and multimodal representation learning for speech, emotion and classroom understanding.

News

  • Oct 2026 Invited talk on LLM-driven multi-agent systems for teacher professional development at the Department of Educational Psychology, University of Georgia.
  • 2026 Two first-author papers accepted to Findings of EMNLP 2026: prompt injection attacks on LLM graders and solution divergence in LLM problem solving.
  • 2026 MEMO, on prompt optimization for long-context reasoning, accepted to the NeurIPS 2026 Workshop on Test-time Continual Learning Agents.
  • Jun 2026 Started a research internship at Capital One (June–August 2026).
  • Jun 2026 Two oral presentations at AIED 2026, including our multi-agent system i-VIP.
  • 2025 Our knowledge tagging paper received the IAAI 2025 Innovative Application Award.
  • 2025 A proposal I co-designed and drafted on AI long-form grading was funded by the Gates Foundation.
  • Jun 2025 Gave a tutorial on Bringing GenAI to Education and co-organized the GenAI4EDU workshop at EDM 2025.
  • 2025 Ask-Before-Detection accepted to ACL 2025.

Selected Publications

* denotes equal contribution. Full list on the publications page.

  • EMNLP Findings 2026
    Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
    Hang Li, Kaiqi Yang, Yucheng Chu, Hui Liu, Jiliang Tang
    Findings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026
    arXivPDF
  • EMNLP Findings 2026
    “**Important** You Should Give Me Full Credits!”: Exploring Prompt Injection Attacks on LLM-Based Automatic Grading Systems
    Hang Li, Fedor Filippov, Yuling Lin, Pengfei He, Kaiqi Yang, Yucheng Chu, Yingqian Cui, Hui Liu, Jiliang Tang
    Findings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2026
    arXivPDF
  • AIED 2026
    i-VIP: A LLM-Driven Multi-Agent System for Professional Development of Mathematics Teachers
    Hang Li, Kaiqi Yang, Yucheng Chu, Ahreum Han, Rui Meng, Yasemin Copur-Gencturk, Hui Liu
    International Conference on Artificial Intelligence in Education (AIED), 2026 Oral
  • ACL 2025
    Ask-Before-Detection: Identifying and Mitigating Conformity Bias in LLM-Powered Error Detector for Math Word Problem Solutions
    Hang Li, Tianlong Xu, Kaiqi Yang, Yucheng Chu, Yanling Chen, Yichi Song, Qingsong Wen, Hui Liu
    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 2025
  • IAAI 2025
    Knowledge Tagging with Large Language Model Based Multi-Agent System
    Hang Li, Tianlong Xu, Ethan Chang, Qingsong Wen
    Proceedings of the 37th Annual Conference on Innovative Applications of Artificial Intelligence (IAAI), 2025 Innovative Application Award
  • IEEE SPM 2025
    Large Language Models for Education: A Survey and Outlook
    Shen Wang*, Tianlong Xu*, Hang Li*, Chaoli Zhang, Joleen Liang, Jiliang Tang, Philip S. Yu, Qingsong Wen
    IEEE Signal Processing Magazine (SPM), 2025 500+ citations
    arXivPDF
  • EDM 2025
    A LLM-Powered Automatic Grading Framework with Human-Level Guidelines Optimization
    Yucheng Chu*, Hang Li*, Kaiqi Yang, Harry Shomer, Hui Liu, Yasemin Copur-Gencturk, Jiliang Tang
    International Conference on Educational Data Mining (EDM), 2025
    arXivPDF
  • SIGKDD Explor. 2024
    Exploring the Potential of Large Language Models (LLMs) in Learning on Graphs
    Zhikai Chen, Haitao Mao, Hang Li, Wei Jin, Hongzhi Wen, Xiaochi Wei, Shuaiqiang Wang, Dawei Yin, Wenqi Fan, Hui Liu, Jiliang Tang
    ACM SIGKDD Explorations Newsletter, 2024 450+ citations
    arXivPDF
  • EMNLP 2021
    CTAL: Pre-training Cross-modal Transformer for Audio-and-Language Representations
    Hang Li, Yu Kang, Tianqiao Liu, Wenbiao Ding, Zitao Liu
    Conference on Empirical Methods in Natural Language Processing (EMNLP), 2021

Honors & Awards

  • IAAI Innovative Application Award, AAAI/IAAI 2025
  • 2nd Place, OGB-LSC Node Classification Track, NeurIPS 2022
  • 1st Place (Task 3) and 2nd Place (Task 4 & overall), NeurIPS 2020 Education Challenge
  • 2nd Place, ClariQ Challenge (Clarifying Questions for Dialogue), EMNLP 2020
  • 1st Place, STABILO Time-Series Classification Challenge, ACM UbiComp 2020

Media Coverage