Concise dynamic grammars using N-best selection

• US 7,630,899 B1
• Filed: 09/07/2006
• Issued: 12/08/2009
• Est. Priority Date: 06/15/1998
• Status: Expired due to Fees

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Marcus, Stephen Michael, Rosinski, Richard R., Brown, Deborah W., Goldberg, Randy G.
• Primary Examiner(s): Han, Qi
• Application Number:
  US11/530,002
• Time in Patent Office: 1,188 Days
• Field of Search: 704/1, 704/231, 704/270, 704/271, 704/235, 704/243, 704/270.1
• US Class Current: 704/270.1
• CPC Class Codes: G10L 15/02 Feature extraction for speech recognition; Selection of recognition unit
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/197 Probabilistic grammars, e.g. word n-grams
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: G10L 25/27 characterised by the analysis technique

Abstract

A method and apparatus derive a dynamic grammar composed of a subset of a plurality of data elements that are each associated with one of a plurality of reference identifiers. The present invention generates a set of selection identifiers on the basis of a user-provided first input identifier and determines which of these selection identifiers are present in a set of pre-stored reference identifiers. The present invention creates a dynamic grammar that includes those data elements that are associated with those reference identifiers that are matched to any of the selection identifiers. Based on a user-provided second identifier and on the data elements of the dynamic grammar, the present invention selects one of the reference identifiers in the dynamic grammar.

Claims

1. A computer-implemented method for selecting one of a plurality of reference identifiers stored prior to receiving user speech input, each reference identifier being associated with at least one data element, the method comprising:

a) generating, via processor, a plurality of selection identifiers associated with a first user speech input, wherein the first user speech input comprises at least one non-letter, non-number typographical character;

b) determining reference identifiers that match any one of the plurality of selection identifiers;

c) creating a data element subset comprising at least one data element associated with matching reference identifiers;

d) selecting at least one of the reference identifiers that matches any of the selection identifiers on the basis of the at least one data element; and

e) if more than one reference identifier matches any of the selection identifiers, then generating a dynamic speech recognition grammar based on the more than one reference identifier and including the data element subset associated with the more than one matching reference identifier and processing a second user input with the dynamic speech recognition grammar.

2. The method according to claim 1, wherein the step a) comprises generating the plurality of selection identifiers in accordance with a criterion.

3. The method according to claim 2, wherein the step a) comprises applying a predetermined algorithm to the criterion.

4. The method according to claim 3, wherein the predetermined algorithm comprises a Hidden Markov Model algorithm and wherein the criterion comprises a first input identifier.

5. The method according to claim 3, wherein the predetermined algorithm comprises a predetermined criterion based on one of a confusion matrix and a plurality of confusion sets, and wherein the predetermined criterion comprises a first input identifier.

6. The method according to claim 2, wherein the step d) comprises:

i) providing a criterion; and

ii) determining which reference identifier is associated with a data element that has a predetermined relationship with respect to the received criterion.

7. The method according to claim 6, wherein the predetermined relationship is a matching relationship.

8. The method of claim 1, further comprising prompting the user for the second user input that is to be processed by the dynamic grammar.

9. An apparatus for selecting one of a plurality of reference identifiers stored prior to receiving user speech input, each reference identifier being associated with at least one data element, the apparatus comprising:

a) means for generating a plurality of selection identifiers associated with user speech input, wherein the user speech input comprises at least one non-letter, non-number typographical character;

b) means for determining reference identifiers that match any one of the selection identifiers;

c) means for creating a data element subset comprising at least one data element associated with those reference identifiers that match any of the selection identifiers;

d) means for selecting one of the reference identifiers that matches any of the selection identifiers on the basis of the at least one data element associated with the selected reference identifiers; and

e) means for, if more than one reference identifier matches any of the selection identifiers, generating a dynamic grammar based on the more than one reference identifier and including the data element subset associated with the more than one matching reference identifier and processing a second user input with the dynamic grammar.

10. The apparatus according to claim 9, wherein the means for generating comprises means for generating the plurality of selection identifiers in accordance with a criterion.

11. The apparatus according to claim 10, wherein the means for generating further comprises means for applying a predetermined algorithm to the criterion.

12. The apparatus according to claim 11, wherein the predetermined algorithm comprises a Hidden Markov Model algorithm and wherein the criterion comprises a first input identifier.

13. The apparatus according to claim 11, wherein the predetermined algorithm comprises one based on one of a confusion matrix and a plurality of confusion sets.

14. The apparatus according to claim 9, wherein the means for selecting comprises:

i) means for providing a criterion; and

ii) second means for determining which reference identifier is associated with a data element that has a predetermined relationship with respect to the received criterion.

15. The apparatus according to claim 14, wherein the predetermined relationship is a matching relationship.

16. An apparatus for responding to at least one vocal input from a user, comprising:

a speech recognizing device for producing a predetermined identifier in response to user vocal input, wherein the user vocal input comprises at least one non-letter, non-number typographical character;

a processing device that receives the predetermined identifier from the speech recognizer;

at least one memory coupled to the processing device and including at least one reference identifier stored prior to receiving the user vocal input having a predetermined relationship to the predetermined identifier;

means for providing the user with access to the reference identifier having the predetermined relationship to the predetermined identifier;

means for generating a dynamic speech recognition grammar from the at least one reference identifier and including the data element subset associated with the more than one matching reference identifier; and

means for prompting the user to receive second user vocal input that is to be processed by the dynamic speech recognition grammar.

17. The apparatus according to claim 16, wherein each reference identifier includes at least one data element associated therewith, and wherein the predetermined relationship exists between at least a portion of the predetermined identifier and the data element of the reference identifier provided by the means for providing.