Advances in Gaussian Processes
Abstract
I present some of our group’s recent work on Gaussian processes (GP) both from a theoretical and applied perspective. I first introduce the Renye GP, which is an alternative objective for GPs capable of improving generalization through tuning the induced regularization. I then highlight the ability of mini-batch stochastic gradient descent to perform inference in GPs (a correlated setting) and hence scaling them far beyond what has been thought possible. From an applied perspective, I introduce predictive GP models that can be used for joint event data, weakly supervised settings, and state-of-the-art multi-output regression.
Bio
Raed Al Kontar is an Assistant Professor in the Department of Industrial & Operations Engineering at the University of Michigan. His main research interest is data science using probabilistic models, with a focus on applications within Internet of Things (IoT) enabled systems, specifically in tele-service settings. He has received several awards, including being a best paper finalist in INFORMS QSR section and Data Mining section.
