5th International Conference on Spoken Language Processing

Sydney, Australia
November 30 - December 4, 1998

Grammar Fragment Acquisition using Syntactic and Semantic Clustering

Kazuhiro Arai (1), Jeremy H. Wright (2), Giuseppe Riccardi (2), Allen L. Gorin (2)

(1) NTT Human Interface Laboratories, Japan
(2) AT&T Laboratories-Research, USA

A new method is proposed for automatically acquiring Fragments to understand fluent speech. The goal of this method is to generate a collection of Fragments, each representing a set of syntactically and semantically similar phrases. First, phrases frequently observed in the training set are selected as candidates. Each candidate phrase has three associated probability distributions of : following contexts, preceding contexts, and associated semantic actions. The similarity between candidate phrases is measured by applying the Kullback-Leibler distance to these three probability distributions. Candidate phrases that are close in all three distances are clustered into a Fragment. Salient sequences of these Fragments are then automatically acquired, and exploited by a spoken language understanding to classify calls in AT&T's ``How May I Help You?'' task. The experimental results show that the average and maximum improvements in call-type classification performance of 2.2% and 2.8% are respectively achieved by introducing the Fragments.

Full Paper

Bibliographic reference.  Arai, Kazuhiro / Wright, Jeremy H. / Riccardi, Giuseppe / Gorin, Allen L. (1998): "Grammar fragment acquisition using syntactic and semantic clustering", In ICSLP-1998, paper 0063.