Concept learning in description logics using refinement operators

29^th November 2018, 10:00
Hadrien Pulcini
University of Liverpool

Abstract

I will discuss the 2010 paper "Concept learning in description logics using refinement operators" by J.Lehmann and P.Hitzler.
In that paper the authors design an algorithm that, given some background knowledge as well as a set of positive and negative examples, outputs a description logic concept satisfied by all of the positive examples and by none of the negative ones. This is done by iteratively refining weaker concepts using a refinement operator.

I will explain the learning algorithm and some theoretical results they provided about refinement operators in general.