Library of existing spoken dialog data for use in generating new natural language spoken dialog systems

• US 9,240,197 B2
• Filed: 07/02/2013
• Issued: 01/19/2016
• Est. Priority Date: 01/05/2005
• Status: Expired due to Fees

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Begeja, Lee, DiFabbrizio, Giuseppe, Gibbon, David Crawford, Hakkani-Tur, Dilek Z., Liu, Zhu, Renger, Bernard S., Shahraray, Behzad, Tur, Gokhan
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): ORTIZ SANCHEZ, MICHAEL
• Application Number:
  US13/933,170
• Publication Number:
  US 20130325443A1
• Time in Patent Office: 931 Days
• Field of Search: 704/257, 704/251
• US Class Current: 1/1
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 25/48 specially adapted for particular use

Abstract

A machine-readable medium may include a group of reusable components for building a spoken dialog system. The reusable components may include a group of previously collected audible utterances. A machine-implemented method to build a library of reusable components for use in building a natural language spoken dialog system may include storing a dataset in a database. The dataset may include a group of reusable components for building a spoken dialog system. The reusable components may further include a group of previously collected audible utterances. A second method may include storing at least one set of data. Each one of the at least one set of data may include ones of the reusable components associated with audible data collected during a different collection phase.

Claims

1. A method comprising:

collecting a plurality of audible utterances, via a plurality of industry specific spoken dialog systems, and during a plurality of collection phases comprising respective defined periods of time, to yield stored audible utterances;

organizing, via a processor, the stored audible utterances into a plurality of datasets having call-type labels, wherein each dataset in the plurality of datasets pertains to a unique industrial sector in a plurality of industrial sectors, and wherein the each dataset for its respective unique industrial sector is independent of other datasets in the plurality of industrial sectors;

generating an annotation guide comprising mnemonic names, textual descriptions, and both a positive example utterance and a negative example utterance for an associated call-type; and

building a natural language spoken dialog system using the plurality of datasets and the annotation guide, wherein datasets having a call-type label associated with the negative example utterance of the annotation guide are not included in the natural language spoken dialog system.

2. The method of claim 1, further comprising, prior to the building of the natural language spoken dialog system, comparing, for each of the call-type labels, each utterance in the plurality of audible utterances to the negative example utterance for the associated call-type.

3. The method of claim 1, wherein the unique industrial sector of each dataset was collected using a corresponding industry specific spoken dialog system of the plurality of industry specific spoken dialog systems.

4. The method of claim 3, wherein the corresponding industry specific spoken dialog system and the unique industrial sector share a common task domain.

5. The method of claim 1, wherein the plurality of datasets are stored in an extensible markup language database.

6. The method of claim 1, wherein the plurality of datasets are stored in a relational database.

7. The method of claim 1, wherein the plurality of audible utterances are each associated with utterance-type category.

8. A system comprising:

a processor; and

a computer-readable storage medium having instructions stored which, when executed by the processor, result in the processor performing operations comprising;

collecting a plurality of audible utterances, via a plurality of industry specific spoken dialog systems, and during a plurality of collection phases comprising respective defined periods of time, to yield stored audible utterances;

organizing the stored audible utterances into a plurality of datasets having call-type labels, wherein each dataset in the plurality of datasets pertains to a unique industrial sector in a plurality of industrial sectors, and wherein the each dataset for its unique industrial sector is independent of other datasets in the plurality of industrial sectors;

generating an annotation guide comprising mnemonic names, textual descriptions, and both a positive example utterance and negative example utterance for an associated call-type; and

building a natural language spoken dialog system using the plurality of datasets and the annotation guide, wherein datasets having a call-type label associated with the negative example utterance of the annotation guide are not included in the natural language spoken dialog system.

9. The system of claim 8, further comprising, prior to the building of the natural language spoken dialog system, comparing, for each of the call-type labels, each utterance in the plurality of audible utterances to the negative example utterance for the associated call-type.

10. The system of claim 8, wherein the unique industrial sector of each dataset was collected using a corresponding industry specific spoken dialog system of the industry specific spoken dialog systems.

11. The system of claim 10, wherein the corresponding industry specific spoken dialog system and the unique industrial sector share a common task domain.

12. The system of claim 8, wherein the plurality of datasets are stored in an extensible markup language database.

13. The system of claim 8, wherein the plurality of datasets are stored in a relational database.

14. The system of claim 8, wherein the plurality of audible utterances are each associated with utterance-type category.

15. A computer-readable storage device having instructions stored which, when executed by a computing device, result in the computing device performing operations comprising:

collecting a plurality of audible utterances, via a plurality of industry specific spoken dialog systems, and during a plurality of collection phases comprising respective defined periods of time, to yield stored audible utterances;

organizing the stored audible utterances into a plurality of datasets having call-type labels, wherein each dataset in the plurality of datasets pertains to a unique industrial sector in a plurality of industrial sectors, and wherein the each dataset for its respective unique industrial sector is independent of other datasets in the plurality of industrial sectors;

generating an annotation guide comprising mnemonic names, textual descriptions, and both a positive example utterance and negative example utterance for an associated call-type; and

building a natural language spoken dialog system using the plurality of datasets and the annotation guide, wherein datasets having a call-type label associated with the negative example utterance of the annotation guide are not included in the natural language spoken dialog system.

16. The computer-readable storage device of claim 15, further comprising, prior to the building of the natural language spoken dialog system, comparing, for each of the call-type labels, each utterance in the plurality of audible utterances to the negative example utterance for the associated call-type.

17. The computer-readable storage device of claim 15, wherein the unique industrial sector of each dataset was collected using a corresponding industry specific spoken dialog system of the industry specific spoken dialog systems.

18. The computer-readable storage device of claim 17, wherein the corresponding industry specific spoken dialog system and the unique industrial sector share a common task domain.

19. The computer-readable storage device of claim 15, wherein the plurality of datasets are stored in an extensible markup language database.

20. The computer-readable storage device of claim 15, wherein the plurality of datasets are stored in a relational database.