Abstract: Distribution shift is a naturally occurring phenomenon that manifests in various forms of real-world data when the underlying distribution of the data changes over time or between different contexts. This is especially important in the area of human-centered AI where distribution shifts could have important implications in terms of performance, robustness, and bias, among others. In this talk, I will discuss our recent works on addressing distribution shifts in representation learning for different types of human-centered data and their implications.
Bio: Dr. Etemad is an Associate Professor at the Department of Electrical and Computer Engineering, Queen's University. He holds an endowed professorship of Mitchell Professor in AI for Human Sensing & Understanding, and leads the Ambient Intelligence and Interactive Machines (Aiim) lab. His main area of research is human-centered machine learning and deep learning. Dr. Etemad received his M.A.Sc. and Ph.D. degrees in Electrical and Computer Engineering from Carleton University, Ottawa, Canada, in 2009 and 2014, respectively. Prior to joining Queen’s, he held several industry positions as lead scientist. He has published over 170 papers in top venues in the area (e.g., NeurIPS, ICLR, AAAI, CVPR, T-PAMI, ECCV, ICCV, etc.), is a co-inventor of 10 patents, and has given over 25 invited talks. Dr. Etemad is an Associate Editor for IEEE Transactions on Affective Computing and IEEE Transactions on Artificial Intelligence. He has served as a PC member/reviewer, and has held organizing roles at various venues. He has received a number of awards including Supervisor of the Year Award (at Queen’s), Instructor of the Year Award (at Queen’s), and several Best Paper Awards (e.g., at ACM ICMI'23). Dr. Etemad’s lab and research program have been funded by the Natural Sciences and Engineering Research Council (NSERC) of Canada, Ontario Centers of Excellence (OCE), Canadian Foundation for Innovation (CFI), Mitacs, and other organizations, as well as the private sector. He is currently a Visiting Fellow at the University of Cambridge and has previously been a Visiting Faculty at Google Research.
Link to join virtually: https://cam-ac-uk.zoom.us/j/81322468305
This talk is not being recorded.
