RESCHEDULED DUE TO SNOW
2/10/2017 | 12:00pm | E62-350
Reception to follow.

Nathan Kallus

Assistant Professor
Cornell University

Abstract: Personalization has long been central in machine learning, with successful applications to recommendation systems in electronic settings. A question of growing urgency is how to translate this success to emergent contexts in medicine and business where the data available has an inherent observational or behavioral nature. In the first part of the talk, I will present a new approach to building personalization models based on purely observational data, such as hospitals' electronic medical records, where the isolated effect of a treatment may be hidden by confounding factors. This is important because, unlike electronic settings, in medicine and other settings, experimentation can be prohibitively small-scale, costly, dangerous, and unethical in comparison to passive data collection, which can be massive. Based on a new reformulation of the personalization problem as a single learning task, the new learning algorithms are significantly more efficient and interpretable than standard approaches. I will demonstrate their power in specific personalized medicine and policymaking applications. In the second part of the talk, I will address the problem of dynamic assortment personalization in the face of a highly heterogeneous population and many items and using only behavioral purchase data. I formulate the problem as a new kind of bandit problem, a discrete-contextual stochastic bandit. I show how to use a new low-rank consumer choice model and convex optimization to tractably achieve regret that is orders of magnitude smaller than before possible.

Bio: Nathan Kallus is an Assistant Professor in the School of Operations Research and Information Engineering at Cornell University and Cornell Tech in NYC. His research revolves around data-driven decision making, the interplay of optimization and statistics in decision making and inference, and the analytical capacities and challenges of observational, large-scale, and web-driven data. Prof. Kallus holds a PhD in Operations Research from MIT as well as a BA in Mathematics and BS in Computer Science from UC Berkeley. Before coming to Cornell, Nathan was a Visiting Scholar at USC's Department of Data Sciences and Operations and a Postdoctoral Associate at MIT's Operations Research and Statistics group.