Knowledge-Rich Natural Language Processing
Kiril Simov & Petya Osenova
Language & Computation
Week One - 11.00-12.30 - Level: I
The goal of the course is to provide basic ideas on the Semantic Web technology necessary for understanding the extraction of linguistic knowledge from world knowledge datasets and its incorporation in various Natural Language Processing (NLP) tasks. The availability of huge amount of world knowledge facts, stored in databases like DBPedia, GeoNames, etc. and of best practices on modeling the stored linguistic knowledge (like LEMON lexicon model or Open Linguistics Group cloud) makes this stream of knowledge-rich NLP quite promising in real applications.
The course introduces the key ideas behind the current active research and extensive use of knowledge-rich resources and technology in the era of linked Semantic Web. Our understanding of knowledge-rich approaches includes various types of formalized linguistic and factual data. In this respect, the relations among lexical databases of various kinds (WordNets, VerbNets, etc.), ontologies as highly formalized common knowledge stores, and Linked Open Data are considered in detail. The course focuses also on the Semantic-based processing and partly – on the Semantic Web Representation, storage and exchange of data (RDF, OWL, SPARQL).
In addition, the important applications of the knowledge-rich resources and technology are discussed. These are: semantic annotation, parsing, machine translation, content search and information extraction.