Attaching Multiple Prepositional Phrases: Backed-off Estimation Generalized

Attaching Multiple Prepositional Phrases: Generalized Backed-off Estimation Paola Merlo IRCS-U. of Pennsylvania LATL-University of Geneva 2 rue de Candolle 1204 Geneva, Switzerland merlo~latl, unige, ch Matthew W. Crocker Centre for Cognitive Science University of Edinburgh 2 Buccleuch Place, Edinburgh, UK EH8 9LW mwc©cogsci, ed. ac. uk Cathy Berthouzoz LATL University of Geneva 2 rue de Candolle 1204 Geneva, Switzerland berthouzoz~lat i. unige, ch Abstract There has recently been considerable inter- est in the use of lexically-based statistical techniques to resolve prepositional phrase attachments. To our knowledge, however, these investigations have only considered the problem of attaching the first PP, i.e., in a IV NP PP] configuration. In this paper, we consider one technique which has been successfully applied to this prob- lem, backed-off estimation, and demon- strate how it can be extended to deal with the problem of multiple PP attach- ment. The multiple PP attachment intro- duces two related problems: sparser data (since multiple PPs are naturally rarer), and greater syntactic ambiguity (more at- tachment configurations which must be dis- tinguished). We present and algorithm which solves this problem through re-use of the relatively rich data obtained from first PP training, in resolving subsequent PP attachments. 1 Introduction Ambiguity is the most specific feature of natural lan- guages, which sets them aside from programming languages, and which is at the root of the difficulty of the parsing enterprise, pervading languages at all levels: lexical, morphological, syntactic, semantic and pragmatic. Unless clever techniques are devel- oped to deal with ambiguity, the number of possible parses for an average sentence (20 words) is simply intractable. In the case Of prepositional phrases, the expansion of the number of possible analysis is the Catalan number series, thus the number of possible analyses grows with a function that is exponential in the number of Prepositional Phrase (Church and Patil, 1982). One of the most interesting topics of debate at the moment, is the use of frequency in- formation for automatic syntactic disambiguation. As argued in many pieces of work in the AI tra- dition (Marcus, 1980; Crain and Steedman, 1985; Altmann and Steedman, 1988; Hirst, 1987), the ex- act solution of the disambiguation problem requires complex reasoning and high level syntactic and se- mantic knowledge. However, current work in part- of-speech tagging has succeeded in showing that it is possible to carve one particular subproblem and solve it by approximation -- using statistical tech- niques -- independently of the other levels of com- putation. In this paper we consider the problem of prepo- sitional phrase (PP) ambiguity. While there have been a number of recent studies concerning the use of statistical techniques for resolving single PP at- tachments, i.e. in constructions of the form [V NP PP], we are unaware of published work which applies these techniques to the more general, and patho- logical, problem of multiple PPs, e.g. IV NP PP1 PP2 ...]. In particular, the multiple PP attachment problem results in sparser data which must be used to resolve greater ambiguity: a …

Content truncated to 500 words