The Journey from AI to Bedside: Translating Academic Research to Serving Patients Using Ray
summary
Builds a web app that uses RLlib and reinforcement learning to convert EMR notes, vitals and twice‑daily patient uploads into sequential disease states, trains supervised detectors for ~20 treatments, and deploys an RL policy to optimize rewards like oxygenation, hospitalization and readmission. Pilot in India labels hundreds of patients a day, manages 14‑day twice‑daily follow‑up, and frees about five minutes per encounter; effectiveness metrics (AUCs improving with more data) are promising though final clinical outcomes are still pending.
AI-written summary · about
chapters (4)
related talks
36mResponsible AI
17mAI Innovation in Health
23mEngineering AI-Powered Predictive Healthcare Systems
26mFrom Pixel to Patient: Making Medical AI Work in the Real World
25mAI for healthcare: Scaling Access and Quality of Care for Everyone
58mHealthcare with AI of the Future
25mLifesaving AI and JavaScript
20mProject Bonsai: How Microsoft is Using Ray and RLlib Inside a Machine Teaching Service