MaLGa Colloquia - Some non-parametric Results

Title

MaLGa Colloquia - Some non-parametric Results

Speaker

Yoav Freund - University of California San Diego (UCSD)

Abstract

Non-parametric classifiers are much more flexible than parametric ones. This allows them to compete with the Bayes optimal classifier, and not just with the best rule in a parametric class. While parametric approaches enjoy well understood bounds on their convergence rates, existing convergence bounds for non-parametric methods require making uncheckable assumptions on the underlying distribution. In this talk I will present two results, the first gives convergence rate bounds for kNN. The second describes a non-parametric active learning algorithm. Both results require no a-priori assumptions on the underlying distributions.

Bio

Yoav Freund is a professor of Computer Science and Engineering at UC San Diego. His work is in the area of machine learning, computational statistics and their applications. Dr. Freund is an internationally known researcher in the field of machine learning, a field which bridges computer science and statistics. He is best known for his joint work with Dr. Robert Schapire on the Adaboost algorithm. For this work they were awarded the 2003 Gödel prize in Theoretical Computer Science, as well as the Kanellakis Prize in 2004, an ensemble learning algorithm which is used to combine many “weak” learning machines to create a more robust one.

When

Friday, May 31st, 2.30pm

Where

DIBRIS/DIMA, Via Dodecaneso 35, Room 322