Darema Lecture Series Begins With Georgia Tech's Judy Hoffman
Speaker:
Judy Hoffman, assistant professor in the School of Interactive Computing at Georgia Tech
Description:
The Illinois Institute of Technology College of Science is pleased to announce the establishment of the Dr. Frederica Darema Lecture Series in Computer Science. This permanent fund will help advance female and minority early-stage computer science researchers at U.S. academic institutions.
The lecture series is designed to encourage women and individuals from under-represented groups to pursue academic careers in computer sciences, and focuses on providing speaking opportunities for tenure track assistant professors (or the equivalent) at U.S. institutions in their fourth to sixth year. Lectureships may also be awarded to exceptional junior researchers in U.S. federal or industrial research laboratories in their third to fifth years of career, following doctoral/postdoctoral studies.
Artificial Intelligence researcher Judy Hoffman, assistant professor in the School of Interactive Computing at Georgia Tech, will be the inaugural guest speaker for the Frederica Darema Lecture Series giving a lecture on “How Dataset Bias Leads to Learned Model Failures.”
Hoffman was a Research Scientist at Facebook AI Research. Hoffman received my PhD in Electrical Engineering and Computer Science from UC Berkeley, and was a Postdoctoral Researcher at UC Berkeley and at Stanford. She was awarded the NSF Graduate Fellowship and the Rosalie M. Stern Fellowship.
“How Dataset Bias Leads to Learned Model Failures” examines the ability to automatically understand what is perceived in visual data is in increasingly high demand. However, despite tremendous performance improvement in recent years, state-of-the-art deep visual models learned using large-scale benchmark datasets still fail to generalize to the diverse visual world. In this talk I will discuss a general purpose semi-supervised learning algorithm, domain adversarial learning, which facilitates transfer of information between different visual environments and across different semantic tasks thereby enabling recognition models to generalize to previously unseen worlds, such as from simulated to real-world driving imagery. I’ll also touch on the pervasiveness of dataset bias and how this bias can adversely affect underrepresented subpopulations.
The lecture is scheduled for 12:30 p.m. December 5, and will be held at The McCormick Tribune Campus Center.
The Darema Lecture Series is named after Dr. Frederica Darema, a Greek American physicist with a career that has spanned physics, and computer and computational sciences. She received her B.S. degree from the School of Physics and Mathematics of the University of Athens – Greece, and M.S. and Ph.D. degrees in theoretical nuclear physics from the Illinois Institute of Technology and the University of California at Davis, respectively.