Automatic Extraction Of Systematic Polysemy Using Tree-Cut

This paper describes an automat ic method for extracting systematic polysemy from a hierarchically organized semantic lexicon (WordNet). Systematic polysemy is a set of word senses that are related in systematic and predictable ways. Our method uses a modification of a tree generalization technique used in (Li and Abe, 1998), and generates a tree-cut, which is a list of clusters that parti t ion a tree. We compare the systematic relations extracted by our automatic method to manually extracted WordNet cousins. 1 I n t r o d u c t i o n In recent years, several on-line broad-coverage semantic lexicons became available, including L D O C E (Procter, 1978), WordNet (Miller, 1990) and H E C T O R .(Kilgarriff, 1998a). These lexicons have been used as a domainindependent semantic resource as well as an evaluation criteria in various Natural Language Processing (NLP) tasks, such as Information Retrieval (IR), Information Extraction (IE) and Word Sense Disambiguation (WSD). However, those lexicons are rather complex. For instance, WordNet (version 1.6) contains a total of over 120,000 words and 170,000 word senses, which are grouped into around 100,000 synsets (synonym sets). In addit ion to the size, word entries in those lexicon are often polysemous. For instance, 20% of the words in Wordnet have more than one sense, and the average number of senses of those polysemous words is around 3. Also, the distinction between word senses tends to be ambiguous and arbitrary. For example, the following 6 senses are listed in WordNet for the noun "door": 1. d o o r a swinging or sliding barrier 2. d o o r the space in a wall 3. d o o r anything providing a means of access (or escape) 4. d o o r a swinging or sliding barrier that will close off access into a car 5. d o o r a house that is entered via a door 6. d o o r a room that is entered via a door Because of the high degree of ambiguity, using such complex semantic lexicons brings some serious problems to the performance of NLP systems. The first, obvious problem is the computat ional intractability: increased processing time needed to disambiguate multiple possibilities will necessarily slow down the system. Another problem, which has been receiving attention in the past few years, is the inaccuracy: when there is more than one sense applicable in a given context, different systems (or human individuals) may select different senses as the correct sense. Indeed, recent studies in WSD show that, when sense definitions are fine-grained, similar senses become indistinguishable to human annotators and often cause disagreement on the correct tag (Ng et al., 1999; Veronis, 1998; Kilgarriff, 1998b). Also in IR and IE tasks, difference in the correct sense assignment will surely degrade recall and precision of the systems. Thus, it is apparent that , in order for a lexicon to be useful as an evaluation criteria for NLP systems, it must represent word senses at the level of granularity that captures human intuition. In Lexical Semantics, several approaches have been proposed which organize a lexicon based on systematic polysemy: 1 a set of word senses that are related in systematic and predictable ISystematic polysemy (in the sense we use in this paper) is also referred to as regular polysemy (Apresjan, 1973) or logical polyseray (Pustejovsky, 1995).