Using Far from Perfect ML to Help Patients - A Billion Medical Devices

This Stanford seminar explores how machine learning can enhance personal devices to help manage health conditions outside clinical settings. Learn from Carnegie Mellon University Associate Professor Mayank Goel as he discusses building real-time ML systems that measure depression symptoms, fatigue, sleep quality, and hyperactivity. Discover how these systems, despite their inherent imperfections, can provide actionable health information to patients, caretakers, and doctors. The talk addresses the challenges of interpreting noisy data and making it useful despite accuracy limitations. Professor Goel, whose inventions have been incorporated into products by Apple, Google, and Meta, shares insights on technologies currently deployed in clinics worldwide. Part of Stanford's Human-Computer Interaction Seminar series, this presentation examines the growing role of technology in healthcare beyond basic step counters and heart rate monitors.