He Cheng
He Cheng
University of Colorado Anschutz Medical Campus
My research spans large language model reasoning, biomedical natural language processing, knowledge graphs, multi-agent systems, reinforcement learning, interpretable machine learning, and model robustness.
About me
He Cheng, Ph.D., is a Postdoctoral Researcher in the Department of Biomedical Informatics at the University of Colorado Anschutz Medical Campus. His current research centers on large language model reasoning, biomedical natural language processing (BioNLP), knowledge graphs, and orchestrator-centered multi-agent systems. He develops scalable knowledge graph retrieval methods and reinforcement learning approaches for reliable, cost-aware clinical reasoning.
Before that, he obtained his Ph.D. in Computer Science from Utah State University, where he conducted research on anomaly detection, explainability, and backdoor attacks in machine learning, publishing multiple first-author papers in top data mining and machine learning venues. He also holds an M.S. in Electrical and Computer Engineering from the State University of New York at Binghamton and a B.E. in Mechanical Engineering from the China University of Petroleum. Beyond research, he is passionate about applying AI to healthcare, and enjoys hiking, coding new tools, and exploring interdisciplinary applications of machine learning.
My research
I am a Postdoctoral Researcher in the Department of Biomedical Informatics at the University of Colorado Anschutz Medical Campus.
My current research focuses on LLM reasoning, biomedical natural language processing (BioNLP), knowledge graphs, and multi-agent systems. I develop reinforcement learning methods for orchestrating specialized agents in clinical reasoning while balancing diagnostic quality and inference cost.
I am the lead developer of LogosKG, a hardware-optimized framework for scalable and interpretable multi-hop retrieval on large biomedical knowledge graphs. I have also worked extensively on anomaly detection, including counterfactual explanations (CFDet) and backdoor attack frameworks (BLOG, BA-OCAD, BadSAD).
Beyond research, I enjoy hiking, coding new tools, and exploring interdisciplinary applications of AI in healthcare.
Research interests
Language models and BioNLP
Large language model reasoning, biomedical natural language processing, retrieval-augmented generation, and clinical decision support.
Knowledge graph reasoning
Efficient, scalable, and interpretable multi-hop retrieval over large biomedical knowledge graphs.
Multi-agent systems
Orchestrator-centered agent collaboration and reinforcement learning for cost-aware clinical reasoning.
Interpretable and robust ML
Model transparency, anomaly detection, and robustness against backdoor attacks.
Recent highlights
LogosKG accepted to ACL 2026
Our paper, “LogosKG: Hardware-Optimized Scalable and Interpretable Knowledge Graph Retrieval,” was accepted to the 64th Annual Meeting of the Association for Computational Linguistics.
Paper · Code · Online demo
BadSAD at IEEE Big Data 2025
“BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection” appears in the proceedings of the 2025 IEEE International Conference on Big Data.