Word Sense Disambiguation Based on Word Similarity Calculation Using Word Vector Representation from a Knowledge-based Graph

Proceedings of the 27th International Conference on Computational Linguistics, pages 2704–2714 Santa Fe, New Mexico, USA, August 20-26, 2018. 2704 Word Sense Disambiguation Based on Word Similarity Calculation Using Word Vector Representation from a Knowledge-based Graph Dongsuk O*1, Sunjae Kwon*2, Kyungsun Kim1 and Youngjoong Ko3 1R&D Center, Diquest Inc., Seoul 08390, Republic of Korea 2School of Elec. & Comp. Engineering, UNIST, Ulsan 44919, Republic of Korea 3Dept. of Computer Engineering, Dong-A University, Busan 49315, Republic of Korea 1ods@diquest.com, 2soon91jae@unist.ac.kr, 1kksun@diquest.com, 3youngjoong.ko@gmail.com Abstract Word sense disambiguation (WSD) is the task to determine the sense of an ambiguous word according to its context. Many existing WSD studies have been using an external knowledge- based unsupervised approach because it has fewer word set constraints than supervised ap- proaches requiring training data. In this paper, we propose a new WSD method to generate the context of an ambiguous word by using similarities between an ambiguous word and words in the input document. In addition, to leverage our WSD method, we further propose a new word similarity calculation method based on the semantic network structure of BabelNet. We evaluate the proposed methods on the SemEval-2013 and SemEval-2015 for English WSD dataset. Ex- perimental results demonstrate that the proposed WSD method significantly improves the base- line WSD method. Furthermore, our WSD system outperforms the state-of-the-art WSD sys- tems in the Semeval-13 dataset. Finally, it has higher performance than the state-of-the-art un- supervised knowledge-based WSD system in the average performance of both datasets. 1 Introduction In natural language text, it is very common for a word to have more than one sense. For example, in a sentence “An airline hires new cabin crews” of Figure 1, the words ‘airline,’ ‘hires,’ ‘new,’ ‘cabin’ and ‘crews’ have more than two senses. In this case, we can map ‘airline’ to ‘airline Noun#2’, ‘hires’ to ‘hire Verb#1,’ ‘new’ to ‘new Adjective#6,’ ‘cabin’ to ‘cabin Noun#3’ and ‘crews’ to ‘crew Noun #7’ by the context of this sentence. The word sense disambiguation (WSD) is the task to determine the correct meaning of an ambiguous word in a given context. WSD is being used as a key step for performance improvement in many natural language processing tasks such as machine translation (Vickrey et al., 2005; Chan et al., 2007), information retrieval (Sanderson 1994; Stokoe et al., 2003) and so on. WSD can be divided into supervised approaches and knowledge-based unsupervised approaches. In the supervised approach, machine learning models are trained by a corpus, in which the correct senses of ambiguous words are already annotated by human annotator (Weissenborn et al., 2015; Melamud et al., 2016; Raganato et al., 2017). However, constructing training corpus for all languages and words is tremendously expensive, so the supervised approaches generally have some limitations on the set of the words that can be disambiguated. On the other hand, the knowledge-based unsupervised approaches utilize lexical knowledge bases (LKBs) such as a Wordnet (Banerjee and Pederson., 2003; Chaplot et al., 2015). These approaches have performed WSD by combining contextual information …

Content truncated to 500 words