When Clinical Notes Cannot Move: Designing NLP for Multi-institutional Research
Attend this webinar to learn about the architectural options that could resolve challenges with multi-institutional clinical natural language processing (NLP). NCI CBIIT’s Dr. Umit Topaloglu will address:
- the reality of how a multi-institutional environment could mean the inability to move clinical notes, the failure to pool annotations, and a lack of portability due to differences in documentation, semantics, etc. across research sites.
- how you can resolve this problem by using centralized, federated, or hybrid architectures, depending upon what should move across sites (data, models, prompts, etc.) and what must remain local.
- how this topic extends to large language models and agentic workflows, where the unit of validation is not a single model but a versioned, multi-step system instead.
Consider attending if you’re interested in topics like federated learning and how we can use such networks to create interoperable, trustworthy systems for multi-institutional cancer research. By attending this webinar, you’ll gain practical approaches to semantic portability, site-specific and continuous validation, subgroup-aware monitoring, provenance, and governance for systems that must remain reliable despite ever-evolving changes.
If you’re curious to learn more about NCI’s federated learning activities, visit our project page.
The American Medical Informatics Association’s (AMIA) NLP Working Group is hosting this webinar.
About the Speaker
Dr. Topaloglu is the branch chief of NCI CBIIT’s Clinical & Translational Research Informatics Branch. In this position, he leads NCI’s clinical research informatics strategy across precision medicine, clinical trials reporting, semantic infrastructure, real-world data, and artificial intelligence (AI). He is also the founder and leader of the Federated Learning-based AI in Multimodal Exploration (FLAIMME) consortium—a national consortium helping NCI-designated cancer centers share AI models while protecting data privacy.