PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Using Unlabeled Data in Generalization Error Bounds
Matti Kääriäinen
In: (Ab)use of Bounds, 18 Dec 2004, Whistler, Canada.

Abstract

We discuss two new methods for obtaining generalization error bounds in a semi-supervised setting. The first method works in the realizable case and has provable optimality guarantees. The second method requires no extra assumptions and provides bounds that seem to be tight on real world learning domains.

EPrint Type:Conference or Workshop Item (Oral)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Learning/Statistics & Optimisation
Theory & Algorithms
ID Code:955
Deposited By:Matti Kääriäinen
Deposited On:07 March 2005