Working Group Webinar Library

Clinical NLP works well when the data stays in one place. But what happens when oncology research spans institutions where notes can't move and annotations can't be pooled? This talk explores how to choose between centralized, federated, and hybrid approaches, and what it takes to keep LLM and agentic workflows reliable as models, documentation, and environments change.

This session features four distinguished leaders representing entrepreneurship, industry, and early- and mid-career academic pathways in biomedical informatics and clinical NLP and LLM. The interactive session offers candid insights, practical guidance, and diverse perspectives to support trainees, early-career researchers, and anyone interested in shaping a long-term career in AI, NLP/LLM and biomedical informatics. Our panelists bring a uniquely broad range of real-world experiences. Presenters

Get an overview of current NLM funding opportunities for AI and biomedical informatics research, plus an update on NIH's unified funding strategy - with practical guidance on aligning your project with NIH/NLM priorities and identifying the right funding path.

Dr. Yin presents ADSMoE, a novel transformer autoencoder that integrates multi-modal data—neuroimaging, genetic, protein, and clinical—to identify four biologically distinct Alzheimer's disease subtypes, validated across three major cohorts (ADNI, NACC, and UKBB).

Learn how AI-powered "digital twin" patient models could replace traditional clinical trials, accelerating medical discovery at a fraction of the time and cost.