Foundation Models for Cancer Workshop
March 26, 2026 | 12:00 PM – 1:00 PM
Online
If you’re attending this three-day workshop, be sure to join NCI CBIIT staff during Session 8 as they co-host the conversation around federated learning (FL) for foundational models.
In this session, Dr. Shannon Silkensen, Dr. Umit Topaloglu, and their cancer center colleagues will highlight:
- the rationale for why cancer trial matching remains difficult and why federated AI is the right approach.
- the necessary technical and institutional groundwork to stand up a functioning FL network across four academic medical centers.
- how foundational models can extract structured, computer-readable information from both clinical text and medical imaging.
- a training data pipeline and early performance findings (including a binary classification task—interventional vs. observational trial type—with results showing baseline accuracy).
NCI has built an FL network to give researchers a collaborative environment for building and operating robust foundational models across several cancer centers without forfeiting data privacy. NCI CBIIT has played a convening and fiduciary role in this project; supports the FL network’s project management activities; and ensures that the cancer centers contribute to the processes for building, testing, and training foundational models.
Register today! And visit the workshop website for information on the other sessions, speaker biographies, and more.
This three-day workshop is hosted by NCI’s Division of Cancer Treatment & Diagnosis.