A Statistical Model for Parsing and Word-Sense Disambiguation

A Statistical Model for Parsing and Word-Sense Disambiguation Daniel M. Bikel Dept. of Computer & Information Science University of Pennsylvania 200 South 33rd Street, Philadelphia, PA 19104-6389, U.S.A. dbikel@cis, upenn, edu Abstract This paper describes a first attempt at a sta- tistical model for simultaneous syntactic pars- ing and generalized word-sense disambigna- tion. On a new data set we have constructed for the task, while we were disappointed not to find parsing improvement over a traditional parsing model, our model achieves a recall of 84.0% and a precision of 67.3% of exact synset matches on our test corpus, where the gold standard has a reported inter-annotator agree- ment of 78.6%. 1 Introduction In this paper we describe a generative, statis- tical model for simultaneously producing syn- tactic parses and word senses in sentences. We begin by motivating this new approach to these two, previously-separate problems, then, after reviewing previous work in these areas, we describe our model in detail. Finally, we will present the promising results of this, our first attempt, and the direction of future work. 2 Motivation for the Approach 2.1 Motivation from examples Consider the following examples: 1. IBM bought Lotus for $200 million. 2. Sony widened its product line with per- sonal computers. 3. The bank issued a check for $100,000. 4. Apple is expecting [NP strong results]. 5. IBM expected [SBAa each employee to wear a shirt and tie]. With Example 1, the reading [IBM bought [Lotus for $200 million]] is nearly impossi- ble, for the simple reason that a monetary amount is a likely instrument for buying and not for describing a company. Similarly, there is a reasonably strong preference in Example 2 for [pp with personal computers] to attach to widened, because personal computers are products with which a product line could be widened. As pointed out by (Stetina and Na- gao, 1997), word sense information can be a proxy for the semantic- and world-knowledge we as humans bring to bear on attachment decisions such as these. This proxy effect is due to the "lightweight semantics" that word senses--in particular WordNet word senses-- convey. Conversely, both the syntactic and semantic context in Example 3 let us know that bank is not a river bank and that check is not a restaurant bill. In Examples 4 and 5, knowing that the complement of expect is an NP or an SBAR provides information as to whether the sense is "await" or "require". Thus, Examples 3-5 illustrate how the syntactic context of a word can help determine its meaning. 2.2 Motivation from previous work 2.2.1 Parsing In recent years, the success of statistical pars- ing techniques can be attributed to several fac- tors, such as the increasing size of comput- ing machinery to accommodate larger models, the availability of resources such as the Penn Treebank (Marcus et al., 1993) and the suc- cess of machine learning techniques for lower- level NLP problems, such as part-of-speech tagging (Church, 1988; Brill, 1995), and PP- attachment (Brill and Resnik, …

Content truncated to 500 words