Building Commons for Clinical LLMs

A two-day international conference on large language models in clinical settings, organised by the Health Data Hub, France's national health data platform, with the LISN, LIMICS and LIG research laboratories.

The organising idea is in the title: rather than every hospital system and every country building isolated medical models, the conference asks what a shared commons for clinical language models would look like, and what it would take to build one.

Day one is six keynote spotlights plus two high-level panels, Building Commons for Clinical LLMs and Deploying LLM-Based Solutions into Clinical Practice, followed by poster and networking sessions. Confirmed keynote subjects include evaluating and implementing language models for medical tasks from Stanford Medicine, the development and evaluation of a Japanese medical model from RIKEN iTHEMS, and open-source digital commons for generative AI in healthcare from the French PARTAGES consortium.

Day two is three collaborative workshops.

Around 120 participants are expected from research, healthcare and industry. Attendance is free but pre-registration is required and capacity is genuinely limited. Session clock times were not published at the time of listing.