Assistant Professor
College of Letters & Science | School of Computer, Data & Information Sciences
Hometown: Hershey, PA
Ben Lengerich is an Assistant Professor of Statistics at UW-Madison, with expertise in AI and machine learning for healthcare. His research develops context-adaptive models to better understand complex diseases and improve precision medicine. He earned his PhD in Computer Science from Carnegie Mellon and was a postdoctoral researcher at MIT CSAIL and the Broad Institute. His work bridges cutting-edge machine learning with real-world medical applications.
Prof. Lengerich prefers virtual talks only.
Talks:
LLMs in Medicine: What Works, What Doesn’t?
Large language models (LLMs) like GPT-4 are being integrated into healthcare, but do they truly improve clinical decision-making? This talk explores where LLMs succeed, where they fail, and why context matters in medical AI. Using insights from my research on dataset debugging and interpretability, I’ll discuss how LLMs handle errors in clinical data, when their predictions can be misleading, and what’s needed to build more trustworthy, context-aware AI for medicine.
Genomics Meets AI: Learning Personalized Biology
AI is revolutionizing precision medicine, but can it truly personalize treatment? This talk explores how machine learning, particularly context-adaptive models, helps uncover patient-specific gene regulatory networks. Drawing from my research in genomic AI, I’ll discuss how context shapes gene expression, how models can predict disease risk and drug response, and what’s next for AI-driven precision medicine.
Debugging Healthcare Data with AI
Bad data leads to bad decisions—especially in medicine. This talk examines how AI can be used to identify and correct errors in medical datasets, improving both research and clinical decision-making. Using findings from my NeurIPS dataset debugging work, I’ll highlight common pitfalls in medical AI, why models make systematic errors, and how AI-assisted quality control can lead to more reliable, interpretable, and actionable insights in healthcare.