Data processing: speech signal processing – linguistics – language – Linguistics
Reexamination Certificate
2001-07-26
2004-06-15
Edouard, Patrick N. (Department: 2654)
Data processing: speech signal processing, linguistics, language
Linguistics
C704S255000
Reexamination Certificate
active
06751584
ABSTRACT:
BACKGROUND
The present invention relates to an application that builds linguistic models from a corpus of speech.
For a machine to comprehend speech, not only must the machine identify spoken (or typed) words, but it also must understand language grammar to comprehend the meaning of commands. Accordingly, much research has been devoted to the construction of language models that a machine may use to ascibe meaning to spoken commands. Often, language models are preprogrammed. However, such predefined models increase the costs of a speech recognition system. Also, the language models obtained therefrom have narrow applications. Unless a programmer predefines the language model to recognize a certain command, the speech recognition system that uses the model may not recognize the command. What is needed is a training system that automatically extracts grammatical relationships from a predefined corpus of speech.
SUMMARY
An embodiment of the present invention provides a method of learning grammar from a corpus, in which context words are identified from a corpus. For the other non-context words, the method counts the occurrence of predetermined relationships with the context words, and maps the counted occurrences to a multidimensional frequency space. Clusters are grown from the frequency vectors. The clusters represent classes of words; words in the same cluster posses the same lexical significancy and provide an indicator of grammatical structure.
REFERENCES:
patent: 5325298 (1994-06-01), Gallant et al.
patent: 5619709 (1997-04-01), Caid et al.
patent: 5835893 (1998-11-01), Ushioda
patent: 5839106 (1998-11-01), Bellegarda
patent: 5860063 (1999-01-01), Gorin et al.
patent: 6052657 (2000-04-01), Yamron
patent: 6073091 (2000-06-01), Kanevsky et al.
patent: 6094653 (2000-07-01), Li et al.
“Dimensions of Meaning,” Hinrich Schutze, Center for the Study of Language and Information, Ventura Hall.
“Grammar Fragment Acquisition using Syntactic and Semantic Clustering,” Jeremy H. Wright, Giuseppe Riccardi, Allen L. Gorin & Kazuhiro Arai.
“Improved Clustering Techniques for Class-Based Statistical Language Modelling,” Reinhard Kneser and Hermann Ney.
“Aggregate and Mixed Order Markov Models for Statistical Language Processing,” Lawrence Saul and Fernando Pereira.
“Empirical Acquisition of Word and Phrase Classes in the Atis Domain,” Michael K. McCandless and James R. Glass.
“Distributional Clustering of English Words,” Fernando Pereria, Naftali Tishby and Lillian Lee.
Bangalore Srinivas
Riccardi Giuseppe
AT&T Corp.
Edouard Patrick N.
LandOfFree
Automatic clustering of tokens from a corpus for grammar... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Automatic clustering of tokens from a corpus for grammar..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Automatic clustering of tokens from a corpus for grammar... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3310214