Spoken Language Processing: A Guide to Theory, Algorithm, and System Development
Spoken Language Processing: A Guide to Theory, Algorithm, and System Development
A robust system for natural spoken dialogue
ACL '96 Proceedings of the 34th annual meeting on Association for Computational Linguistics
Hi-index | 0.00 |
In this paper, we present a mandarin spoken dialogue system--STRQS (Shanghai Traffic Route Querying System), which is used for querying best traffic route between any two locations in Shanghai. A series of language processing strategies is used to understand speech utterances. The understanding processing is done in three steps: First, word segmentation and part-of-speech tagging module splits the utterance into words and labels them with semantic categories. The second step is a robust partial parsing process. Parsing is based on Unification Grammar (UG). An augmented chart algorithm with feature computing is implemented. Finally, the parsed utterance is associated with a semantic interpreter by a frame module. Semantic based analysis method we developed can directly extract information from the output of a speech recognizer, which contains errors and ill-formed components. The testing results demonstrate the robustness of our approach.