ESBL: An integrated method for learning from partial information

Research output: Contribution to journalArticlepeer-review

Abstract

A general framework for representing incomplete domain knowledge through the use of predicate variables is given. Explanation-based learning (EBL) and similarity-based learning (SBL) are reinterpreted in terms of Horn-clause resolution. It is shown that, thus reinterpreted, EBL and SBL can be naturally integrated to produce an "abductive" method for refining incomplete domain knowledge.

Original languageEnglish
Pages (from-to)323-343
Number of pages21
JournalAnnals of Mathematics and Artificial Intelligence
Volume4
Issue number3-4
DOIs
StatePublished - Sep 1991

Fingerprint

Dive into the research topics of 'ESBL: An integrated method for learning from partial information'. Together they form a unique fingerprint.

Cite this