Dartmouth Events

Abstract:  Data heterogeneity is common across many high-impact application domains, ranging from financial fraud detection to healthcare, from manufacturing to traffic analytics. It is at the core of the 'Variety' aspect of big data. Such heterogeneity can be presented in a variety of forms, including task heterogeneity, where multiple related learning tasks may form a hierarchical structure; view heterogeneity, where information is being collected from various sources; oracle heterogeneity, where multiple oracles may have different opinions regarding the true label of an example; instance heterogeneity, where a single example can be decomposed into a set of instances with heterogeneous labels; motif heterogeneity in large networks, where various motifs may lead to local clusters bearing heterogeneous physical meanings; etc. In this talk, I will present our recent work on learning from such data heterogeneity, focusing on two major challenges, i.e., how to unveil the common characteristics underneath the facade of data heterogeneity, and how to model the co-existence of various types of data heterogeneity so that the learning model could enjoy the best of all possible worlds. In particular, I will hinge on multiple applications, discuss the problem-specific data heterogeneity, and our proposed techniques for modeling such data heterogeneity, together with empirical results on real applications. Finally, I will conclude the talk by sharing my vision for modeling data heterogeneity.

Bio:  Dr. Jingrui He is an assistant professor in School of Computing, Informatics, and Decision Systems Engineering at Arizona State University. She received her Ph.D in Computer Science from Carnegie Mellon University in 2010, and joined ASU in 2014. Her research focuses on heterogeneous machine learning, rare category analysis, active learning and semi-supervised learning, with applications in social network analysis, healthcare, financial fraud detection, and manufacturing processes. Dr. He is the recipient of the 2016 NSF CAREER Award, 2 times recipient of the IBM Faculty Award in 2015 and 2014 respectively, and was selected as IJCAI 2017 Early Career Spotlight. She has published more than 60 refereed articles, and is the author of the book on Analysis of Rare Categories (Springer-Verlag, 2011). Her papers have been selected as Bests of the Conference by ICDM 2016, ICDM 2010, and SDM 2010. Dr. He has served on the senior program committee/program committee at KDD, IJCAI, AAAI, SDM, ICML, etc.