Jiatong Li

profile.jpeg

I am a second-year Ph.D. student in the School of Computer, Data & Information Sciences at University of Wisconsin-Madison where I am fortunate to be advised by Dr. Sharon Li. My current research focuses on long-term LLM agents and agentic memory.

Previously, I obtained my M.S. and B.S. degrees in Data Science from University of Science and Technology of China (USTC) in 2025 and 2022, respectively, under the supervision of Dr. Qi Liu. My prior research experience spans perturbation-driven LLM robustness evaluation, data mining, and cognitive diagnosis for intelligent education. Throughout my academic journey, I have had the pleasure of collaborating with Dr. Mengxiao Zhu at USTC, Dr. Xiang Liu at Educational Testing Service, Dr. Renjun Hu at Alibaba Cloud, and Dr. Yifei Yuan at University of Copenhagen.

News

Oct 08, 2026 Our paper about LLM agents’ information disclosure, Thinking Is Not Telling: Information Disclosure in User-Service LLM Agents, is accepted at TMLR. Thank you all collaborators!
Sep 24, 2026 Our paper about long context agent memory, Multi-Head Recurrent Memory Agents, is accepted at the NeurIPS 2026. Thank you all collaborators!
May 11, 2026 Start my research internship at CoreAI group, Microsoft.
Aug 17, 2025 Start my PhD study at UW-Madison.

Selected Publications

  1. NeurIPS ’26
    mhm-preview.png
    Multi-Head Recurrent Memory Agents
    Jiatong Li, Samuel Yeh, and Sharon Li
    The Annual Conference on Neural Information Processing Systems, 2026
  2. TMLR
    thinkagent-preview.png
    Thinking Is Not Telling: Information Disclosure in User-Service LLM Agents
    Jiatong Li, Changdae Oh, Hyeong Kyu Choi, and 2 more authors
    Transactions on Machine Learning Research, 2026
  3. NeurIPS ’24
    perteval-preview.png
    PertEval: Unveiling Real Knowledge Capacity of LLMs with Knowledge-Invariant Perturbations (Spotlight)
    Jiatong Li, Renjun Hu, Kunzhe Huang, and 5 more authors
    In Advances in Neural Information Processing Systems, 2024