An Empirical Comparison of Unknown Word Prediction Methods

Proceedings of the 5th International Joint Conference on Natural Language Processing, pages 767–775, Chiang Mai, Thailand, November 8 – 13, 2011. c⃝2011 AFNLP An Empirical Comparison of Unknown Word Prediction Methods Kostadin Cholakov†, Gertjan van Noord†, Valia Kordoni‡, Yi Zhang‡ † University of Groningen, The Netherlands ‡ Saarland University and DFKI GmbH, Germany {k.cholakov,g.j.m.van.noord}@rug.nl {kordoni,yzhang}@coli.uni-sb.de Abstract We compare two types of methods which deal with unknown words in the context of computational grammars. Methods of the first type are based on the idea of su- pertagging and use a tagger to predict lex- ical descriptions for unknown tokens in a given input. The second type of methods perform lexical acquisition (LA) which, in the context of this paper, refers to the au- tomatic acquisition of new lexical entries for the lexicon of a given grammar. The methods are compared based on the effect their application has on the parsing cov- erage and accuracy of the GG grammar of German (Crysmann, 2003). In particu- lar, we adapt the LA method of Cholakov and van Noord (2010) which was origi- nally developed for the Dutch Alpino sys- tem to be used with the GG. Its impact on coverage and accuracy on a test corpus of German newspaper texts is compared to the results reported previously on the same corpus for methods which employed a tag- ger. Furthermore, in a smaller experiment, we show that the linguistic knowledge this LA method provides can also be used for sentence realisation. 1 Introduction Computational grammars of natural language which rely on hand-crafted lexicons containing elaborate linguistic descriptions usually have low lexical coverage. This is due to the fact that it is impossible to list every word in a language in the lexicon. If a word is unknown to the grammar, or its description in the lexicon is wrong or incom- plete, the grammar might fail to produce a (cor- rect) analysis. In this paper, we compare two types of ap- proaches for dealing with unknown words. The first type is based on the concept of supertagging while the second one performs LA. Generally, su- pertagging refers to the process of applying a se- quential tagger to assign lexical descriptions asso- ciated with each word in an input string, relative to a given grammar. It was introduced as a means to reduce parsing ambiguity of LTAG grammars (Bangalore and Joshi, 1999), and has since been applied within CCG (Clark, 2002; Clark and Cur- ran, 2004) and HPSG (Dridan et al., 2008; Zhang et al., 2010) grammars. Supertagging has also been employed for dealing with unknown words. However, in such methods, the tagger is used to as- sign lexical descriptions only to the unknown to- kens in a given sentence. It is important to note here that henceforth, we will use the term sup- pertagging in this narrow sense of tagging un- known words only. Supertagging methods often work online. The unknown words are assigned lexical entries when they are encountered in the input during parsing. Therefore, …

Content truncated to 500 words