85e83ecfd6a6d310VgnVCM100000c2b1d38dRCRDapproved/UMICH/stats/Home/News & Events/Archived Events/2009-2010 EventsHao Zhou###@###(Mon, 14 Sep 2009)Hao Zhou###@###(Mon, 14 Sep 2009)438 West HallPreliminary Examination: Generalized Linear Model Approach to Multi-Task Learningstats125295480000012529548000003:00 PM<p><b>Title: </b>Generalized Linear Model Approach to Multi-Task Learning<br> <b>Advisor: </b>Professor George Michailidis<br> <b>Committee Members: </b>Associate Professor Kerby Shedden, Associate Professor Ji Zhu</p> <p><b>Abstract:</b> Multi-task learning is a machining learning method that improves learning of signal task by using information from other train sets. In this proposal, we explain how multi-task learning works and give an overview of existing multi-task learning methods and some theoretical backgrounds. We propose a new multi-task learning method based on generalized linear models and discuss the possible steps of statistical inference.</p>Njjsantosjjsantos136683087325755e83ecfd6a6d310VgnVCM100000c2b1d38d____once11112newnewEvent Flyer/UMICH/stats/Home/Events/Dissertations and Oral Preliminary Examinations/hao zhou prelim flyer.pdfHao Zhou