Generating speech recognition grammars from a large corpus of data

• US 20040054530A1
• Filed: 09/18/2002
• Published: 03/18/2004
• Est. Priority Date: 09/18/2002
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: International Business Machines Corporation
• Inventors: Davis, Brent L., Reich, David E.
• Granted Patent: US 7,567,902 B2
• Time in Patent Office: Days
• US Class Current: 704/231
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules

Abstract

A method of generating a speech recognition grammar for use with a speech recognition system can include parsing the corpus of data to identify grammatical structures within the corpus of data. The identified grammatical structures can be compared with grammar generation rules to determine particular ones of the identified grammatical structures to include within the speech recognition grammar. The grammar generation rules can designate which grammatical structures are to be included within the speech recognition grammar. The grammatical structures which have been identified in the parsing step and which also have been designated by the grammar generation rules can be included in the speech recognition grammar.

Claims

1. A method of generating a speech recognition grammar for use with a speech recognition engine comprising:

parsing a corpus of data to identify grammatical structures within the corpus of data;

comparing the identified grammatical structures with grammar generation rules to determine particular ones of the identified grammatical structures to include within the speech recognition grammar, wherein the grammar generation rules designate which grammatical structures are to be included within the speech recognition grammar; and

including the grammatical structures within the speech recognition grammar which have been identified in said parsing step and which also are designated by the grammar generation rules.

2. The method of claim 1, said parsing step further comprising:

identifying individual words within the corpus of data as grammatical parts of speech.

3. The method of claim 2, wherein the grammar generation rules further designate particular grammatical parts of speech to be included within the speech recognition grammar, said including step further comprising:

including the grammatical parts of speech within the speech recognition grammar which have been identified in said parsing step and which are designated by the grammar generation rules.

4. The method of claim 1, further comprising:

receiving a spoken query to search the corpus of data;

speech recognizing a portion of the spoken query which is specified in the speech recognition grammar; and

searching the corpus of data for the recognized portion of the spoken query.

5. A system for generating a speech recognition grammar comprising:

a parser configured to identify grammatical structures within a corpus of data having a plurality of well formed sentences;

a set of grammar generation rules designating particular grammatical structures to be included within a speech recognition grammar; and

a grammar processor configured to generate a speech recognition grammar by including within the speech recognition grammar grammatical structures included within the corpus of data and which also are designated by the grammar generation rules.

6. The system of claim 5, wherein said parser is configured to identify individual words within the corpus of data as grammatical parts of speech, wherein said set of grammar generation rules designate particular parts of speech to be included within a speech recognition grammar, and wherein said grammar processor is configured to include within the generated speech recognition grammar grammatical parts of speech included within the corpus of data which also are designated by the grammar generation rules.

7. A machine readable storage, having stored thereon a computer program having a plurality of code sections executable by a machine for causing the machine to perform the steps of:

parsing a corpus of data to identify grammatical structures within the corpus of data;

comparing the identified grammatical structures with grammar generation rules to determine particular ones of the identified grammatical structures to include within the speech recognition grammar, wherein the grammar generation rules designate which grammatical structures are to be included within the speech recognition grammar; and

including the grammatical structures within the speech recognition grammar which have been identified in said parsing step and which also are designated by the grammar generation rules.

8. The machine-readable storage of claim 7, said parsing step further comprising:

identifying individual words within the corpus of data as grammatical parts of speech.

9. The machine-readable storage of claim 8, wherein the grammar generation rules further designate particular grammatical parts of speech to be included within the speech recognition grammar, said including step further comprising:

including the grammatical parts of speech within the speech recognition grammar which have been identified in said parsing step and which are designated by the grammar generation rules.

10. The machine-readable storage of claim 7, further comprising:

receiving a spoken query to search the corpus of data;

speech recognizing a portion of the spoken query which is specified in the speech recognition grammar; and

searching the corpus of data for the recognized portion of the spoken query.