Ashwin Pananjady, UC Berkeley

Title: Non-Parametric Models for Comparisons and Choice: Computation and Adaptation
Abstract:  I will introduce a class of non-parametric, "permutation-based" models for comparison and choice data that borrows from the literature on sociology and economics and significantly generalizes other popular approaches in these contexts. The talk will focus on the mathematical statistics of fitting these models from data, and describe a methodological toolbox that is inspired by considerations of adaptation as well as computation. We will cover bits and pieces from two papers: the first is joint work with Cheng Mao and Martin Wainwright, and the second is joint work with Richard Samworth.