Date(s) - 25 Aug 2017
1:00 PM - 2:00 PM
3043 ECpE Building Addition
Title: Online Constrained Ranking for Large Scale Content Recommendation
Abstract: Personalized content recommendation is at the core of many of the user facing products at Yahoo and in almost all other online content serving companies. Users need to be matched with interesting content in near real-time while simultaneously achieving business targets. In this talk, I will describe the details of a large-scale recommendation system and involved challenges. We will talk about how constraints are put on the recommendation and ranking algorithms via either quality or business requirements. This constraint ranking is formulated as a Linear Program and solved in an online fashion every time a user is served content. Finally, we will go over some theoretical guarantees when we solve the proposed LP approximately via a primal-dual formulation.
Bio: Akshay Soni received his B.Tech. degree in Information and Communication technology from DA-IICT, Gujarat, India, in 2010, and the M.S. and Ph.D. degrees in Electrical Engineering from the University of Minnesota, Minneapolis, USA, in 2011 and 2015 respectively. He is currently a Scientist at Yahoo Research working on content recommendation and personalization. His research interests include personalization, search, recommendation, adaptive compressive sensing, machine learning for big data, image processing and statistical learning theory.
More information to come.