Towards Vietnamese entity disambiguation

Abstract

Entity Disambiguation (ED) is a fundamental task in Natural Language Processing (NLP). The term Entity is used to mean either a Named Entity or an Abstract Concept. Although there have been many works on the ED task for English and some for Vietnamese, this is the first time this paper tackles the general ED task for Vietnamese that deal with both named entities and abstract concepts. In this paper, we propose a method for linking named entities and abstract concepts in Vietnamese documents to the corresponding articles in the Vietnamese Wikipedia. In particular, it first has to recognize Vietnamese entity mentions, i.e., phrases that represent named entities or abstract concepts. Experimental evaluation is also presented to demonstrate the performance of the proposed method.