ECE  KORHAMMER  SEMINAR  SERIES

Abstract:

In this talk I will first describe our work on developing new tools for screening and intervention in developmental disorders, autism spectrum disorder and eating disorders in particular. I will show how equipped with computer vision and machine learning, we deployed scalable, phone/tablet-based tools in pediatric clinics and homes in the US and Africa. This work resulted in the largest collection of videos of child behavior in their natural environments, and led to the discovery of new biomarkers as well as the replication of known ones for the first time at low cost and scale. Such rich data opens the door for personalized behavioral analysis and interventions as we will illustrate during the presentation. The unique challenges of this field also motivate new foundational work in computer vision and machine learning, and we will exemplify a few in the areas of personalized privacy, do-not-harm Pareto fairness, observation-based studies of the causal effects of behavioral nudges, and personalized machine learning. The works described in this talk are the result of collaborations with clinicians, app developers, neuroscientists, therapists, colleagues, and numerous outstanding students.

Bio:

Guillermo Sapiro was born in Uruguay and educated at the Technion in Israel. After postdoctoral studies at MIT he joined HP Labs, where he co-developed the JPEG-LS standard for lossless image compression. He then spent 15 years at the University of Minnesota before joining Duke University where he is a James B. Duke University Professor. He was awarded the ONR Young Investigator Award, the NSF CAREER Award, the Presidential Early Career Award for Scientists and Engineers, and the National Security Science and Engineering Faculty Fellowship. He received the Test-of-Time awards both in machine learning (ICML) and computer vision (ICCV). He is a Fellow of IEEE and of SIAM, and a member of the American Academy of Arts and Sciences and the National Academy of Engineering.

This seminar is partially supported with funds from the Korhammer Lecture Series