简介:

Abstract

If infinite amounts of labeled data are provided, many machine-learning algorithms become perfect. With finite amounts of data, regularization or priors have to be used to introduce bias into a classifier. We propose a third option: learning with marginalized corrupted features. We (implicitly) corrupt existing data as a means to generate additional, infinitely many, training samples from a slightly different data distribution — this is computationally tractable, because the corruption can be marginalized out in closed form. Our framework leads to machine learning algorithms that are fast, generalize well and naturally scale to very large data sets. We showcase this technology as regularization for general risk minimization and for marginalized deep learning to learn document representations. We further show that marginalized corruption is not limited to features. Marginalized corrupted labels can be used to create dense representations of image tags, and we show that these are significantly better suited for applications such as tag prediction.

 

Short Bio:

Kilian Q. Weinberger is an Associate Professor in the Department of Computer Science & Engineering at Washington University in St. Louis. He received his Ph.D. from the University of Pennsylvania in Machine Learning under the supervision of Lawrence Saul and his undergraduate degree in Mathematics and Computer Science from the University of Oxford.  During his career he has won several best paper awards at ICML, CVPR and AISTATS. In 2011 he was awarded the AAAI senior program chair award and in 2012 he received the NSF CAREER award. Kilian Weinberger’s research is in Machine Learning and its applications. In particular, he focuses on high dimensional data analysis, metric learning, machine learned web-search ranking, transfer- and multi-task learning as well as biomedical applications. Before joining Washington University in St. Louis, Kilian worked as a research scientist at Yahoo! Research in Santa Clara.