Attain Optimal Performance by Shifting Your Goals

Abstract

In the problem of domain adaptation for binary classification, the learner is presented with labeled examples from a source domain, and must correctly classify unlabeled examples from a target domain, which may differ from the source. Previous work on this problem has assumed that the performance measure of interest is the expected value of some loss function. We argue that for different optimality criteria, stronger domain adaptation results are possible than what has previously been established. In particular, we study a class of domain adaptation problems that generalizes both the covariate shift assumption and a model for feature-dependent label noise, and establish optimal classification on the target domain despite not having access to labelled data from this domain. The class of criteria to which our work applies includes some existing criteria designed to ensure nondiscrimination.

Biography

Clay Scott received his PhD in Electrical Engineering from Rice University in 2004, and joined the University of Michigan in 2006 with a primary appointment in EECS. His research interests focus on statistical machine learning theory and algorithms, with an emphasis on nonparametric methods for supervised and unsupervised learning. He has also worked on a number of applications stemming from various scientific disciplines, including brain imaging, nuclear threat detection, environmental monitoring, and computational biology. In 2010, he received the Career Award from the National Science Foundation.