NeurIPS 2026 · Main Track · Poster
Incentivizing Agentic Retrieval for Disease-Centric Clinical Case Search via Trajectory Memory
Advances in Neural Information Processing Systems, 2026
I am an undergraduate student majoring in Artificial Intelligence at Xiamen University and a research intern at Stony Brook University.
My research focuses on AI for medicine and healthcare, particularly medical foundation models, clinical reasoning, agentic retrieval, and multimodal biomedical representation learning. My recent work spans disease-centric clinical case retrieval, medical terminology interoperability, and foundation models for neuroimaging data.
In the long term, I am interested in fundamental machine learning methods that enable reliable representation learning, reasoning, and generalization across biomedical tasks and modalities.
* Equal contribution.
NeurIPS 2026 · Main Track · Poster
Advances in Neural Information Processing Systems, 2026
Department of Applied Mathematics & Statistics /
Department of Computer Science
Research Intern
Advisor: Prof. Chenyu You · Mentor: Yifan Wang
September 2026 – Present
Studying biomedical and neuroimaging foundation models, with a focus on fMRI representation learning, scalable architectures, preprocessing pipelines, and generalization across healthcare tasks.
Institute of Health Big Data, School of Informatics
Undergraduate Research Intern
Advisor: Prof. Liansheng Wang · Mentor: Jie Lin
July 2025 – Present
Working on clinical reasoning, medical agents, and evidence-grounded retrieval, including disease-centric representations, long-horizon reasoning over clinical narratives, temporal modeling, and medical concept normalization.
B.Eng. in Artificial Intelligence
September 2024 – June 2028 (Expected)
Entered in an electronic information-related major in September 2024; transferred to Artificial Intelligence in May 2025.
A trajectory-memory-based framework for improving agentic retrieval in disease-centric clinical case search.
My contribution. Research, method development, and experimental evaluation for disease-centric clinical case retrieval and trajectory memory.
A framework for predicting whether nested text spans contain complete evidence before verification. Online Evidence-Containment Prediction (OECP) connects span containment to online isotonic regression and uses Expected Search-space Reduction (ESR) to guide efficient evidence localization.
Research focus. Online learning, structured prediction, and budgeted evidence verification for question answering and fact verification.
Exploring scalable foundation model architectures for learning representations from functional MRI time-series data.
Current focus. fMRI data structures and preprocessing, large-scale neuroimaging representation learning, Transformer-based temporal models, cross-subject generalization, and downstream evaluation.