Mingchen Li
Ph.D. Student · Computer Science · UMass Amherst
I am a third-year Ph.D. student in Computer Science at the University of Massachusetts Amherst. My research focuses on large language models, agentic retrieval, deep research, reasoning, and AI for biomedicine.
General AI: Agents, Retrieval, and Reasoning
I am interested in building LLM-based agents for reasoning, planning, retrieval, tool use, and self-evolution across diverse tasks. My research focuses on improving the effectiveness, efficiency, reliability, and adaptability of agentic AI systems, including how they use external tools and knowledge, reflect on their own decisions, and adjust their strategies during problem solving.
AI for Biomedicine and Healthcare
My research interests also lie in AI for biomedicine and healthcare, including large language models, retrieval-augmented generation, clinical prediction, biomedical information extraction, medical reasoning, and knowledge graph learning. My work includes applications in cancer, opioid overdose risk prediction, and other clinical prediction tasks, with the goal of developing reliable and efficient AI systems to support biomedical research, clinical diagnosis, and healthcare decision-making.
Publications
Complete list: Google Scholar profile.
General AI
-
RICE-PO: Turning Retrieval Interactions into Credit Signals for Reasoning Agents
NeurIPS 2026pdf -
In-Context Optimization for Retrieval-Augmented Generation: A Gradient-Descent Perspective
NeurIPS 2026pdf -
A condensed transition graph framework for zero-shot link prediction with large language models
-
Semantic Structure based Query Graph Prediction for Question Answering over Knowledge Graph
-
A Hierarchical N-Gram Framework for Zero-Shot Link Prediction
-
Counterfactual Graph for Multi-Agent LLM Calibration
arXiv preprint 2026pdf
Biomedicine & Healthcare
-
Efficient and Effective Internal Memory Retrieval for LLM-Based Healthcare Prediction.
ACL 2026 Findingpdf -
BiomedRAG: A retrieval augmented large language model for biomedicine
-
RT: a Retrieving and Chain-of-Thought framework for few-shot medical named entity recognition
-
LEAP: LLM instruction-example adaptive prompting framework for biomedical relation extraction
Journal of the American Medical Informatics Association (JAMIA) 31 (9), 2010 ... 2024pdf -
Journal of Biomedical Informatics (JBI) 2024pdf
-
A Review of Reinforcement Learning for Natural Language Processing, and Applications in Healthcare
Journal of the American Medical Informatics Association (JAMIA) 2023pdf
Experience
- PhD Student, Computer Science, UMass Amherst. 09.2024 - Present
- Researcher, University of Minnesota Twin Cities. 01.2023 - 08.2024
- Research Assistant, Georgia State University. 01.2021 - 12.2022
Service
- Reviewer — Journals: Artificial Intelligence in Medicine, Engineering Applications of Artificial Intelligence, Expert Systems with Applications, Information Processing & Management, International Journal of Medical Informatics, Journal of Artificial Intelligence Research, Knowledge-Based Systems, Npj Health Systems
- Reviewer — Conferences: EMNLP 2022, ACL 2023, EMNLP 2023, EMNLP Industry Track 2023, IEEE ICHI 2023, NAACL 2024, EACL 2026, EMNLP 2026, NIPS 2026, AAAI 2027, ICLR 2027
Contact
Google Scholar: scholar profile
GitHub: ToneLi
Email: mingchenli@umass.edu