Ontologies and Natural Language Processing

Abstract

Ontologies are widely used in Knowledge Engineering, Artificial Intelligence and Computer Science, in applications related to knowledge management, natural language processing, e-commerce, intelligent integration information, information retrieval, database design and integration, bio-informatics, education, and in new emerging fields like the Semantic Web. Most of these applications involve people from different fields, some of which are not familiar with this term or are only acquainted with the classical notion of ontology, and, therefore, cannot exploit all the possibilities included in an ontology.
This article aims to overcome this situation at least for people interested in one of those fields; namely, natural language processing (NLP). First, we will introduce the basics of ontologies. We will explain what an ontology is, its main components, and the kind of ontologies that there are. We will relate the concept of ontology to those of taxonomy and thesaurus, which are two well known resources in linguistics and other fields related to NLP. We will use them to make ontologies better understandable to that community.
We will also present some interactions between both disciplines. Ontologies and NLP can benefit from each other. On the one hand, ontologies can be used in different areas of NLP, such as Information Retrieval. On the other hand, NLP techniques can be useful for ontologies, especially in Ontology Learning. There is also another issue worth mentioning: ontologies can be used to model NLP-related knowledge -e.g. vocabulary, grammar, linguistic features, etc. -, just as they can model any other domain. Dealing with the large amount of topics relating both areas would be beyond the scope of a reasonable size article. That is why we will focus on two aspects that seem to be among the most relevant nowadays: ontologies about grammatical units, usually referred as 'linguistic ontologies' in literature, and ontology learning (semiautomatic ontology construction) using NIP techniques.