CA2643930A1 - Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer - Google Patents

Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer Download PDF

Info

Publication number
CA2643930A1
CA2643930A1 CA002643930A CA2643930A CA2643930A1 CA 2643930 A1 CA2643930 A1 CA 2643930A1 CA 002643930 A CA002643930 A CA 002643930A CA 2643930 A CA2643930 A CA 2643930A CA 2643930 A1 CA2643930 A1 CA 2643930A1
Authority
CA
Canada
Prior art keywords
collected
phrases
semantic
semantic concepts
vector
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
CA002643930A
Other languages
French (fr)
Inventor
Kenneth Todd Reed
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Call Genie Inc
Original Assignee
Call Genie Inc.
Kenneth Todd Reed
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Call Genie Inc., Kenneth Todd Reed filed Critical Call Genie Inc.
Publication of CA2643930A1 publication Critical patent/CA2643930A1/en
Abandoned legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/18—Speech classification or search using natural language modelling
    • G10L15/1815—Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/18—Speech classification or search using natural language modelling
    • G10L15/183—Speech classification or search using natural language modelling using context dependencies, e.g. language models
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/06—Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice
    • G10L15/063—Training
    • G10L2015/0631—Creating reference templates; Clustering

Abstract

A method and system for building a grammar module for a speech application. The method includes the step of clustering phrases having a semantic similarity. The grammar module comprises phrases in a machine-readable format and semantic concepts associated with the phrases. According to another aspect, the grammar module includes embedded semantic interpretations associated with the semantic concepts.

Claims (23)

1. A method for creating a grammar module for a speech application, said method comprising the steps of:

collecting phrases associated with one or more voice responses;
transcribing said collected phrases into a machine-readable format;

clustering selected ones of said collected phrases into one or more semantic concepts, and wherein said selected collected phrases in each of said semantic concepts have a related meaning;

building a grammar module based on said collected phrases and said semantic concepts.
2. The method as claimed in claim 1, wherein said step of clustering comprises the step of identifying one or more words in each of said collected phrases and associated said collected phrases with a semantic concept when one or more of said words have a meaning which is similar or the same.
3. The method as claimed in claim 2, wherein said step of identifying one or more words comprises generating a vector for said collected phrase, said vector having an element for each of a plurality of words in said collected phrase, and comparing the vector for said collected phrase to a vector for one of said semantic concepts, and associating said collected phrase with said semantic concept if said vector has a number of elements exceeding a predefined threshold.
4. The method as claimed in claim 3, wherein said step of building a grammar module comprises converting a plurality of grammar elements into a machine-readable format and converting said semantic concepts into a machine-readable format, and storing said machine-readable grammar elements and semantic concepts in a computer file.
5. The method as claimed in claim 3, wherein one or more of said vector elements includes an indicator, said indicator providing information about said associated vector element.
6. The method as claimed in claim 5, wherein said indicator comprises a content indicator providing a probability indicator for the occurrence of a word.
7. The method as claimed in claim 5, wherein said indicator comprises a word sense indicator providing an intended meaning for a word.
8. The method as claimed in claim 3, further including the step of inserting one or more synonymous terms for one or more words in said collected phrases wherein said one or more words have a synonymous term, and said vector including a corresponding element for at least some of said synonymous terms.
9. The method as claimed in claim 3, further including the step of inserting one or more hypernyms into said vector, and said one or more hypernyms having a weighting.
10. A system for building a grammar module for a speech application, said system comprising:

means for collecting phrases associated with one or more of said voice responses;

means for transcribing said collected phrases into a machine-readable format;
means for clustering selected ones of said collected phrases into a plurality of semantic concepts, wherein each of said semantic concepts comprises one or more collected phrases having a similar meaning;

means for creating a grammar module based on said collected phrases and said semantic concepts.
11. The system as claimed in claim 10, wherein said means for clustering includes means for characterizing each of said selected collected phrases as a vector, said vector having one or more elements corresponding to one or more words comprising said collected phrase, and each of said semantic concepts including one or more vectors having an element for each of a plurality of words associated with said semantic concept.
12. The system as claimed in claim 11, further including means for comparing each of said collected phrase vectors to one or more of said semantic concept vectors based on a similarity measure, and means for grouping one or more of said collected phrases when said similarity measure exceeds a predetermined threshold.
13. The system as claimed in claim 12, further including means for inserting one or more synonymous terms for one or more words in said collected phrases wherein said one or more words have a synonymous term, and said vector including a corresponding element for at least some of said synonymous terms.
14. The system as claimed in claim 12, further including means for inserting one or more hypernyms into said vector, and said one or more hypernyms each having an associated weighting.
15. A method for creating a grammar module suitable for use with a speech application, said method comprising the steps of:

collecting phrases associated with one or more voice responses;
transcribing said collected phrases into a machine-readable format;

grouping one or more of said collected phrases into a plurality of groups, wherein each of said groups has an associated semantic, said one or more collected phrases being grouped based on a similarity between said collected phrase and the associated semantic concept for said group; and building a grammar module based on said collected phrases and said semantic concepts.
16. The method as claimed in claim 15, wherein said step of grouping comprises determining a similarity between said collected phrase and the associated semantic concept for said group, and comparing said similarity to a predefined threshold, and adding said collected phrase to the group associated with said semantic concept if said predefined threshold is satisfied.
17. The method as claimed in claim 16, further including the step utilizing said collected phrase not satisfying said predefined threshold for a new semantic concept.
18. The method as claimed in claim 17, wherein said semantic concepts comprise a plurality of semantically equivalent words or phrases.
19 19. The method as claimed in claim 16, wherein said similarity is determined according to a similarity function.
20. A method for generating a grammar module for a speech application, said method comprising the steps of:

collecting one or more phrases associated with one or more voice responses;
transcribing said collected phrases into a machine-readable format;

clustering selected ones of said collected phrases into one or more semantic concepts, and wherein said selected collected phrases in each of said semantic concepts have a similar meaning;

interpreting at least some of said semantic concepts;

building a grammar module based on said collected phrases, said semantic concepts and said interpreted semantic concepts.
21. The method as claimed in claim 20, wherein said step of building a grammar module comprises creating a machine-readable grammar file.
22. The method as claimed in claim 21, further including converting said interpreted semantic concepts into a machine-readable format and embedding said interpreted semantic concepts in said machine-readable grammar file.
23. The method as claimed in claim 20, wherein said step of interpreting each of said semantic concepts comprises converting said interpreted semantic concepts into a machine-readable format
CA002643930A 2006-04-17 2007-04-17 Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer Abandoned CA2643930A1 (en)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US79235006P 2006-04-17 2006-04-17
US60/792,350 2006-04-17
PCT/CA2007/000634 WO2007118324A1 (en) 2006-04-17 2007-04-17 Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer

Publications (1)

Publication Number Publication Date
CA2643930A1 true CA2643930A1 (en) 2007-10-25

Family

ID=38609002

Family Applications (1)

Application Number Title Priority Date Filing Date
CA002643930A Abandoned CA2643930A1 (en) 2006-04-17 2007-04-17 Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer

Country Status (3)

Country Link
EP (1) EP2008268A4 (en)
CA (1) CA2643930A1 (en)
WO (1) WO2007118324A1 (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9690771B2 (en) 2014-05-30 2017-06-27 Nuance Communications, Inc. Automated quality assurance checks for improving the construction of natural language understanding systems
US10515150B2 (en) 2015-07-14 2019-12-24 Genesys Telecommunications Laboratories, Inc. Data driven speech enabled self-help systems and methods of operating thereof
US10382623B2 (en) 2015-10-21 2019-08-13 Genesys Telecommunications Laboratories, Inc. Data-driven dialogue enabled self-help systems
US10455088B2 (en) 2015-10-21 2019-10-22 Genesys Telecommunications Laboratories, Inc. Dialogue flow optimization and personalization

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5794193A (en) * 1995-09-15 1998-08-11 Lucent Technologies Inc. Automated phrase generation
US6173261B1 (en) * 1998-09-30 2001-01-09 At&T Corp Grammar fragment acquisition using syntactic and semantic clustering
US5860063A (en) * 1997-07-11 1999-01-12 At&T Corp Automated meaningful phrase clustering
US6317707B1 (en) * 1998-12-07 2001-11-13 At&T Corp. Automatic clustering of tokens from a corpus for grammar acquisition
US6415248B1 (en) * 1998-12-09 2002-07-02 At&T Corp. Method for building linguistic models from a corpus
AU5451800A (en) * 1999-05-28 2000-12-18 Sehda, Inc. Phrase-based dialogue modeling with particular application to creating recognition grammars for voice-controlled user interfaces
ATE405918T1 (en) * 1999-12-20 2008-09-15 British Telecomm LEARNING DIALOGUE STATES AND LANGUAGE MODELS OF THE SPOKEN INFORMATION SYSTEM

Also Published As

Publication number Publication date
WO2007118324A1 (en) 2007-10-25
EP2008268A1 (en) 2008-12-31
EP2008268A4 (en) 2010-12-22

Similar Documents

Publication Publication Date Title
US20110224982A1 (en) Automatic speech recognition based upon information retrieval methods
US20080172224A1 (en) Position-dependent phonetic models for reliable pronunciation identification
US8316004B2 (en) Speech retrieval apparatus and speech retrieval method
Hakkinen et al. N-gram and decision tree based language identification for written words
Chen et al. Retrieval of broadcast news speech in Mandarin Chinese collected in Taiwan using syllable-level statistical characteristics
CN104750677A (en) Speech translation apparatus, speech translation method and speech translation program
CA2643930A1 (en) Method and apparatus for building grammars with lexical semantic clustering in a speech recognizer
Ghannay et al. Acoustic Word Embeddings for ASR Error Detection.
KR20230099269A (en) Named entity recognizing system and method of recognizing named entity
US6963832B2 (en) Meaning token dictionary for automatic speech recognition
CN103548015B (en) A method and an apparatus for indexing a document for document retrieval
Lin et al. Learning methods in multilingual speech recognition
Bohac et al. Post-processing of the recognized speech for web presentation of large audio archive
Staš et al. Slovak language model from internet text data
Kuo et al. Discriminative training for call classification and routing
CN111489742B (en) Acoustic model training method, voice recognition device and electronic equipment
Li et al. Towards unsupervised semantic retrieval of spoken content with query expansion based on automatically discovered acoustic patterns
CN1421846A (en) Phonetic recognizing system
Castan et al. Indexing multimedia documents with acoustic concept recognition lattices.
Nair et al. Pair-wise language discrimination using phonotactic information
Soutner et al. Neural network language model with cache
Xu et al. Robust and Fast Lyric Search based on Phonetic Confusion Matrix.
Batista et al. Extending the punctuation module for European Portuguese
Kumaran et al. Attention shift decoding for conversational speech recognition.
TWI830385B (en) Apparatus and method for emotion prediction of audio sentence and computer program product thereof

Legal Events

Date Code Title Description
EEER Examination request
FZDE Dead

Effective date: 20160201