Work Files Saved Searches
   My Account                                                  Search:   Quick/Number   Boolean   Advanced       Help   


 The Delphion Integrated View

  Buy Now:   Buy PDF- 17pp  PDF  |   File History  |   Other choices   
  Tools:  Citation Link  |  Add to Work File:    
  View:  Expand Details   |  INPADOC   |  Jump to: 
 
 Email this to a friend  Email this to a friend 
       
Title: US5267345: Speech recognition apparatus which predicts word classes from context and words from word classes
[ Derwent Title ]


Country: US United States of America

View Images High
Resolution

 Low
 Resolution

 
17 pages

 
Inventor: Brown, Peter F.; New York, NY
Della Pietra, Stephen A.; Pearl River, NY
Della Pietra, Vincent J.; Blauvelt, NY
Mercer, Robert L.; Yorktown Heights, NY
Resnik, Philip S.; Philadelphia, PA
Chen, Stanley S.; Cambridge, MA

Assignee: International Business Machines Corporation, Armonk, NY
other patents from INTERNATIONAL BUSINESS MACHINES CORPORATION (280070) (approx. 44,393)
 News, Profiles, Stocks and More about this company

Published / Filed: 1993-11-30 / 1992-02-10

Application Number: US1992000834063

IPC Code: Advanced: G06F 9/44; G06F 17/27; G10L 15/10; G10L 15/18; G10L 15/28;
IPC-7: G10L 9/00;

ECLA Code: G10L15/197;

U.S. Class: Current: 704/255; 704/E15.023;
Original: 395/002.64;

Field of Search: 381/041-46 395/2.64

Government Interest:     This invention was made with Government support under Contract Number N00014-91-C-0135 awarded by the Office of Naval Research. The Government has certain rights in this invention.

Priority Number:
1992-02-10  US1992000834063

Abstract: A language generator for a speech recognition apparatus scores a word-series hypothesis by combining individual scores for each word in the hypothesis. The hypothesis score for a single word comprises a combination of the estimated conditional probability of occurrence of a first class of words comprising the word being scored, given the occurrence of a context comprising the words in the word-series hypothesis other than the word being scored, and the estimated conditional probability of occurrence of the word being scored given the occurrence of the first class of words, and given the occurrence of the context. An apparatus and method are provided for classifying multiple series of words for the purpose of obtaining useful hypothesis scores in the language generator and speech recognition apparatus.

Attorney, Agent or Firm: Schechter, Marc D. ;

Primary / Asst. Examiners: Fleming, Michael R.; Doerrler, Michelle

Maintenance Status: E3 Expired  Check current status

INPADOC Legal Status: Show legal status actions          Buy Now: Family Legal Status Report

Designated Country: DE FR GB 

Family: Show 5 known family members

First Claim:
Show all 35 claims
We claim:     1. A language generator comprising:
  • means for generating two or more word-series hypothesis signals, each word-series hypothesis signal representing a word-series hypothesis comprising a series of linguistic units, each linguistic unit comprising one or more words selected from a language;
  • means for generating a single-unit hypothesis score signal for each linguistic unit in each word-series hypothesis, each single-unit hypothesis score signal representing a single-unit hypothesis score comprising an estimate of the probability of occurrence of the individual linguistic unit in the word-series hypothesis;
  • means for generating a word-series hypothesis score signal for each word-series hypothesis, each word-series hypothesis score signal comprising a combination of the single-unit hypothesis scores for the linguistic units in the word-series hypothesis;
  • means for storing one or more word-series hypothesis signals representing candidate word-series hypotheses having the best word-series hypothesis scores; and
  • means for outputting at least one word signal representing at least one word in at least one best candidate word-series hypothesis;
  • characterized in that the means for generating a single-unit hypothesis score signal for a first individual linguistic unit comprises:
  • means for estimating the conditional probability of occurrence of a first class of linguistic units comprising the first individual linguistic unit, given the occurrence of a context comprising the linguistic units in the word-series hypothesis other than the first individual linguistic unit;
  • means for estimating the conditional probability of occurrence of the first individual linguistic unit given the occurrence of the first class of linguistic units, and given the occurrence of the context; and
  • means for generating a single-unit hypothesis score signal for the first individual linguistic unit comprising a first component score comprising a combination of the estimated conditional probability of occurrence of the first class of linguistic units given the occurrence of the context, and the estimated conditional probability of occurrence of the first individual linguistic unit given the occurrence of the first class and given the occurrence of the context.


Background / Summary: Show background / summary

Drawing Descriptions: Show drawing descriptions

Description: Show description

Forward References: Show 69 U.S. patent(s) that reference this one

       
U.S. References: Go to Result Set: All U.S. references   |  Forward references (69)   |   Backward references (3)   |   Citation Link

Buy
PDF
Patent  Pub.Date  Inventor Assignee   Title
Get PDF - 54pp US4748670  1988-05 Bahl et al.  International Business Machines Corporation Apparatus and method for determining a likely word sequence from labels generated by an acoustic processor
Get PDF - 18pp US4831550  1989-05 Katz  International Business Machines Corporation Apparatus and method for estimating, from sparse data, the probability that a particular one of a set of events is the next event in a string of events
Get PDF - 45pp US4980918  1990-09 Bahl et al.  International Business Machines Corporation Speech recognition system with efficient storage and rapid assembly of phonological graphs
       
Foreign References:
Buy
PDF
Publication Date IPC Code Assignee   Title
Get PDF - 27pp EP0242743 1987-10  G10L 5/06 TEXAS INSTRUMENTS INC Speech recognition system 
Get PDF - 18pp EP0299572 1989-01  G10L 5/06 N.V. Philips' Gloeilampenfabrieken Method for connected word recognition 
Get PDF - 8pp EP0418711 1991-03  G10L 5/06 ALCATEL N.V. Method for speaker-adaptive speech recognition 


Other Abstract Info: DERABS G93-259913

Other References:
  • Bahl, L. R., et al. "A Maximum Likeihood Approach to Continuous Speech Recognition," IEEE Transactions on Pattern Analysis and Machine Intelligence, volume PAMI-5, No. 2, pp. 179-180, Mar. 1983. (12 pages) Cited by 42 patents
  • Bahl, L. R., et al. "A Tree-Based Statistical Language Model for Natural Language Speech Recognition." IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 37, No. 7, Jul. 1989, pp. 1001-1008. (8 pages) Cited by 7 patents
  • Bahl, L. R., et al. "Vector Quantization Procedure For Speech Recognition Systems Using Discrete Parameter Phoneme-Based Markov Word Models." , IBM Technical Disclosure Bulletin, vol. 34, No. 7, Dec. 1989, pp. 340 and 341.
  • Brown, P. F., et al. "A Statistical Approach to Machine Translation." Computational Lignuistics, vol. 16, No. 2, Jun. 1990, pp. 79-85.
  • Mays, E., et al. "Context Based Spelling correction." Information Processing & Management, vol. 27, No. 5, 1991, pp. 517-522. (6 pages) Cited by 4 patents [ISI abstract]


  • Inquire Regarding Licensing

    Powered by Verity


    Plaques from Patent Awards      Gallery of Obscure PatentsNominate this for the Gallery...

    Thomson Reuters Copyright © 1997-2014 Thomson Reuters 
    Subscriptions  |  Web Seminars  |  Privacy  |  Terms & Conditions  |  Site Map  |  Contact Us  |  Help