Word Sense Disambiguation Using Label Propagation Based Semi-Supervised Learning

Proceedings of the 43rd Annual Meeting of the ACL, pages 395–402, Ann Arbor, June 2005. c⃝2005 Association for Computational Linguistics Word Sense Disambiguation Using Label Propagation Based Semi-Supervised Learning Zheng-Yu Niu, Dong-Hong Ji Institute for Infocomm Research 21 Heng Mui Keng Terrace 119613 Singapore {zniu, dhji}@i2r.a-star.edu.sg Chew Lim Tan Department of Computer Science National University of Singapore 3 Science Drive 2 117543 Singapore tancl@comp.nus.edu.sg Abstract Shortage of manually sense-tagged data is an obstacle to supervised word sense dis- ambiguation methods. In this paper we in- vestigate a label propagation based semi- supervised learning algorithm for WSD, which combines labeled and unlabeled data in learning process to fully realize a global consistency assumption: simi- lar examples should have similar labels. Our experimental results on benchmark corpora indicate that it consistently out- performs SVM when only very few la- beled examples are available, and its per- formance is also better than monolingual bootstrapping, and comparable to bilin- gual bootstrapping. 1 Introduction In this paper, we address the problem of word sense disambiguation (WSD), which is to assign an appro- priate sense to an occurrence of a word in a given context. Many methods have been proposed to deal with this problem, including supervised learning al- gorithms (Leacock et al., 1998), semi-supervised learning algorithms (Yarowsky, 1995), and unsuper- vised learning algorithms (Sch¨utze, 1998). Supervised sense disambiguation has been very successful, but it requires a lot of manually sense- tagged data and can not utilize raw unannotated data that can be cheaply acquired. Fully unsupervised methods do not need the definition of senses and manually sense-tagged data, but their sense cluster- ing results can not be directly used in many NLP tasks since there is no sense tag for each instance in clusters. Considering both the availability of a large amount of unlabelled data and direct use of word senses, semi-supervised learning methods have re- ceived great attention recently. Semi-supervised methods for WSD are character- ized in terms of exploiting unlabeled data in learning procedure with the requirement of predefined sense inventory for target words. They roughly fall into three categories according to what is used for su- pervision in learning process: (1) using external re- sources, e.g., thesaurus or lexicons, to disambiguate word senses or automatically generate sense-tagged corpus, (Lesk, 1986; Lin, 1997; McCarthy et al., 2004; Seo et al., 2004; Yarowsky, 1992), (2) exploit- ing the differences between mapping of words to senses in different languages by the use of bilingual corpora (e.g. parallel corpora or untagged monolin- gual corpora in two languages) (Brown et al., 1991; Dagan and Itai, 1994; Diab and Resnik, 2002; Li and Li, 2004; Ng et al., 2003), (3) bootstrapping sense- tagged seed examples to overcome the bottleneck of acquisition of large sense-tagged data (Hearst, 1991; Karov and Edelman, 1998; Mihalcea, 2004; Park et al., 2000; Yarowsky, 1995). As a commonly used semi-supervised learning method for WSD, bootstrapping algorithm works by iteratively classifying unlabeled examples and adding confidently classified examples into labeled dataset using a model learned …

Content truncated to 500 words