UNSPECIFIED (2001) When can two unsupervised learners achieve PAC separation? In: 14th Annual Conference on Computational Learning Theory (COLT 2001)/5th European Conference on Computational Learning Theory (EuroCOLT 2001), JUL 16-19, 2001, AMSTERDAM, NETHERLANDS.

Full text not available from this repository.

Abstract

In this paper we study a new restriction of the PAC learning framework, in which each label class is handled by an unsupervised learner that aims to fit an appropriate probability distribution to its own data. A hypothesis is derived by choosing, for any unlabeled instance, the label whose distribution assigns it the higher likelihood. The motivation for the new learning setting is that the general approach of fitting separate distributions to each label class, is often used in practice for classification problems. The set of probability distributions that is obtained is more useful than a collection of decision boundaries. A question that arises, however, is whether it is ever more tractable (in terms of computational complexity or sample-size required) to find a simple decision boundary than to divide the problem up into separate unsupervised learning problems and find appropriate distributions. Within the framework, we give algorithms for learning various simple geometric concept classes. In the boolean domain we show how to learn parity functions, and functions having a constant upper bound on the number of relevant attributes. These results distinguish the new setting from various other well-known restrictions of PAC-learning. We give an algorithm for learning monomials over input vectors generated by an unknown product distribution. The main open problem is whether monomials (or any other concept class) distinguish learnability in this framework from standard PAC-learnability.

Item Type:	Conference Item (UNSPECIFIED)
Subjects:	Q Science > QA Mathematics > QA76 Electronic computers. Computer science. Computer software
Series Name:	LECTURE NOTES IN ARTIFICIAL INTELLIGENCE
Journal or Publication Title:	COMPUTATIONAL LEARNING THEORY, PROCEEDINGS
Publisher:	SPRINGER-VERLAG BERLIN
ISBN:	3-540-42343-5
ISSN:	0302-9743
Editor:	Helmbold, D and Williamson, B
Date:	2001
Volume:	2111
Number of Pages:	17
Page Range:	pp. 303-319
Publication Status:	Published
Title of Event:	14th Annual Conference on Computational Learning Theory (COLT 2001)/5th European Conference on Computational Learning Theory (EuroCOLT 2001)
Location of Event:	AMSTERDAM, NETHERLANDS
Date(s) of Event:	JUL 16-19, 2001
URI:	http://wrap.warwick.ac.uk/id/eprint/11072

Data sourced from Thomson Reuters' Web of Knowledge

Request changes to a record

Actions (login required)

View Item