Embodied Clinical Intelligence Colloquium (ECLIC)

AI systems are currently siloed into highly-specific clinical tasks and are not suitable to handle the complexities and nuances of clinical care. Embodied Clinical Intelligence establishes a new paradigm in which AI systems ground themselves in clinical context, act within real-world clinical workflows, and interact with clinicians in real-time.

We are a community of researchers at the intersection of AI and medicine that are unapologetically technical and uncompromisingly clinical. In our mission of pushing the frontiers of research in embodied clinical intelligence, we host workshops, fireside chats, and guest lectures, to build the community advancing the next generation of embodied AI systems in medicine.

Workshop Series

Technical
Session 1: Beyond the OpenClaw Hype: From Reinforcement Learning to ReAct
Clinical
Session 2: Why We Can't Just Throw AI Algorithms at the EHR: The Messy Reality of Clinical Data
Technical
Session 3: Vision-Language Models
Clinical
Session 4: A Brief History of Clinical AI Evaluation: The Pitfalls of Automated Evaluation Metrics
Technical
Session 5: Reinforcement Learning with Large Language Models
Clinical
Session 6: Lessons Learned from Running Clinical Reader Studies
Technical
Session 7: Developing Agents from a Reinforcement Learning Perspective
Clinical
Session 8: Why Computer Scientists Need to Shadow Clinicians
  • Case Study: Hallucinated Comparisons in Radiology Report Generation Vision-Language Models
  • Experience shadowing Dr. Jae Ho Sohn in the reading room
Conclusion: Towards Embodied Clinical Intelligence