Paper Info
Title
One-sided algorithms for integrating empirical and explanation-based learning
Abstract
The purpose of this paper is to describe a framework for integrating empirical learning with explarxxtion- based learning (EBL)(DeJong & Mooney 1986; Mitchell, Keller & Kedar-Cabelli 19861 and to present an algorithm which does this with both pure conjunctive concepts and rE-CNF concepts. Our framework involves using an empirical and an explanation-based method to form separate hypotheses and then com- bining the hypotheses from the separate sources to form a composite hypothesis. An additional important complication arises because the system is required to learn the domain theory (via an empirical method) at the same time it is using the domain theory to support the explanation-based method. The empirical methods that we use are one-sided algorithms that next generate a hypothesis that is more general than the correct hypothesis (assuming that the hypothesis can be represented in t,he hypothesis representation language). In addition, the empirical algorithms that we consider maintain a single hypothesis that is generalized as little as possible to accommodate positive examples. The hypotheses produced by explanation-based learning with a domain theory acquired with such a one-sided empirical learning method will also never be more general than the correct hypothesis. Since both the empirical and explanation-based hypotheses are not more general than the correct hypothesis, they can be combined by finding the least general hypothesis consistent with both hypotheses. ln this manner, the integrated hypothesis will be the least general hypothesis that is consistent with both the observed data and the domain knowledge. This hypothesis may be more general than either the empirical or explanation-based hypotheses. Some regularities may be ruled out because they are not consistent with the data. Other regularities may be ruled out because they are not consistent with the domain theory. although they may be supported by the data.
Year
DOI
Venue
1989
10.1016/B978-1-55860-036-2.50012-6
ML
Keywords
Field
DocType
one-sided algorithm,explanation-based learning,empirical method,domain knowledge,domain theory
Hypothesis,Empirical learning,Domain knowledge,Explanation-based learning,Computer science,Domain theory,Algorithm,Artificial intelligence,Machine learning,Hypothesis Theory
Conference
ISBN
Citations 
PageRank 
1-55860-036-1
6
5.37
References 
Authors
5
2
Name
Order
Citations
PageRank
Wendy Sarrett13544.00
Michael J. Pazzani281021055.78