Learning Ontologies from Data: The Story So Far
- Speaker: Dr Slava Sazonau (University of Manchester)
- Host: Uli Sattler
- 26th June 2017 at 14:00 in Kilburn L.T 1.5
In this talk, I will discuss the problem of ontology learning, i.e. acquiring rich ontological knowledge from data. This is an emerging field of AI research and a new type of symbolic machine learning aimed at bridging the gap between knowledge and data. I will focus on learning expressive OWL ontologies from labelled graph data (e.g. RDF), discuss challenges and difficulties, describe respective approaches, and present recent promising results including case studies. I will describe DL-Miner - a tool capable of mining rich hypotheses that can be used to (i) build/complete an ontology, (ii) reveal new interesting domain knowledge, (iii) indicate modelling errors or poor data quality. The talk should be of interest to ontology and knowledge engineers, Semantic Web and Linked Data experts, machine learning and data mining researchers.