Online Learning with Model Selection

Current online learning methods suffer issues such as lower convergence rates and limited capability to recover the support of true features compared to their offline counterparts. In this talk, we present a novel framework for online learning based on running averages and introduce a series of online versions of some popular existing offline methods such as Elastic Net, Minimax Concave Penalty and Feature Selection with Annealing. We prove the equivalence between our online methods and their offline counterparts and give theoretical true feature recovery and convergence guarantees for some of them. In contrast to the existing online methods, the proposed methods can extract models with any desired sparsity level at any time. Numerical experiments indicate that our new methods enjoy high accuracy of true feature recovery and a fast convergence rate, compared with standard online and offline algorithms. We also show how the running averages framework can be used for model adaptation in the presence of varying- coefficient models. Finally, we present some applications to large datasets where again the proposed framework shows competitive results compared to popular online and offline algorithms.

Adrian Barbu

Professor, Florida State University on February 1, 2019 at 11:45 AM in EB2 1230

Adrian Barbu received his Ph.D. in Mathematics from the Ohio State University in 2000 and his Ph.D. in Computer Science from the University of California, Los Angeles in 2005 (advised by Dr. Song-Chun Zhu). From 2005 to 2007 he worked in Siemens Corporate Research, first as a research scientist and later as a project manager, with the focus on medical imaging problems. He received the 2011 Thomas A. Edison Patent Award with his co-authors from Siemens for their work on Marginal Space Learning. In 2007 he joined the Statistics department at Florida State University, first as an assistant professor, and since 2013 as an associate professor. He has published more than 70 papers in computer vision, machine learning and medical imaging and he has more than 25 patents related to medical imaging and image denoising.

The Department of Electrical and Computer Engineering hosts a regularly scheduled seminar series with preeminent and leading reseachers in the US and the world, to help promote North Carolina as a center of innovation and knowledge and to ensure safeguarding its place of leading research.