System and method for combining geographic metadata in automatic speech recognition language and acoustic models

• US 8,892,443 B2
• Filed: 12/15/2009
• Issued: 11/18/2014
• Est. Priority Date: 12/15/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Bocchieri, Enrico, Caseiro, Diamantino Antonio
• Primary Examiner(s): Opsasnick, Michael N
• Application Number:
  US12/638,667
• Publication Number:
  US 20110144973A1
• Time in Patent Office: 1,799 Days
• Field of Search: 704/270
• US Class Current: 704/270
• CPC Class Codes: G06F 40/58 Use of machine translation, e.g. for multi-lingual retrieval, for server-side translation for client devices or for real-time translation
• CPC Class Codes: G10L 15/08 Speech classification or search
• 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 2015/081 Search algorithms, e.g. Baum-Welch or Viterbi
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: G10L 25/54 for retrieval

Abstract

Disclosed herein are systems, methods, and computer-readable storage media for a speech recognition application for directory assistance that is based on a user'"'"'s spoken search query. The spoken search query is received by a portable device and portable device then determines its present location. Upon determining the location of the portable device, that information is incorporated into a local language model that is used to process the search query. Finally, the portable device outputs the results of the search query based on the local language model.

Claims

1. A method comprising:

receiving a spoken user search query at a portable device;

determining a present location based on the portable device;

incorporating a granularity description of the present location into a local language model used to process the spoken user search query, the granularity description using weights for topologically concentric locations to determine probabilities; and

outputting results associated with the spoken user search query based on the present location and a term in the spoken user search query.

2. The method of claim 1 further comprising:

incorporating the present location into a local acoustic model used to process the spoken user query.

3. The method of claim 1 wherein the present location of the portable device is determined using one of a global positioning system, triangulation, and automatic number identification.

4. The method of claim 1 wherein the model is a query model described by a probability distribution function of the form p(x|l).

5. The method of claim 1, wherein incorporating of the present location into the language model further comprises:

partitioning training data into geographical areas;

estimating models for each partition; and

selecting a portioned model based on the location of the portable device.

6. The method of claim 5 further comprising generating a user model by combining two models using interpolation.

7. The method of claim 1, wherein entities in the local language model are related at one of a city, an area code, and a state level.

8. The method of claim 1 further comprising:

collecting queries at the portable device, wherein the queries are both textual and spoken;

analyzing the queries for a query location;

determining the query location from the analysis and the present location;

dividing the collection of queries into geographical areas, wherein the geographical areas are based on the determined query locations;

estimating language and acoustic models for each geographical area; and

training a speech recognition application using the estimated language and acoustic models.

9. A system comprising:

a processor; and

a computer-readable storage medium having instructions stored which, when executed by the processor, cause the processor to perform operations comprising;

receiving a spoken user search query at a portable device;

determining a present location based on the portable device;

incorporating a granularity description of the present location into a local language model used to process the spoken user search query, the granularity description using weights for topologically concentric locations to determine probabilities; and

outputting results associated with the spoken user search query based on the present location and a term in the spoken user search query.

10. The system of claim 9, wherein the location of the portable device is determined using one of a global positioning system, triangulation, and automatic number identification.

11. The system of claim 9, the computer-readable storage medium having additional instructions stored which result in the method further comprising:

partitioning training data into geographical areas;

estimating models for each partition; and

selecting a partitioned model based on the location of the portable device.

12. The system of claim 9, the computer-readable storage medium having instructions stored which result in the method further comprising generating a user model by combining two models using interpolation.

13. The system of claim 9, the computer-readable storage medium having additional instructions stored which result in the operations further comprising:

collecting queries at the portable device, wherein the queries are both textual and spoken;

analyzing the queries for a query location;

determining the query location from the analysis and the present location;

dividing the collection of queries into geographical areas, wherein the geographical areas are based on the determined query locations;

estimating language and acoustic models for each geographical area; and

training a speech recognition application using the estimated language and acoustic models.

14. A computer-readable storage device having additional instructions stored which, when executed by a computing device, cause the computing device to perform operations comprising:

receiving a spoken user search query at a portable device;

determining a present location based on the portable device;

incorporating a granularity description of the present location into a local language model used to process the spoken user search query, the granularity description using weights for topologically concentric locations to determine probabilities; and

outputting results associated with the spoken user search query based on the present location and a term in the spoken user search query.

15. The computer-readable storage device of claim 14 wherein the present location of the portable device is determined using one of a global positioning system, triangulation, and automatic number identification.

16. The computer-readable storage device of claim 14 wherein the model is a query model described by probability distribution function of the form p(x|l).

17. The computer-readable storage device of claim 14, the computer-readable storage device having additional instructions stored which result in the operations further comprising:

partitioning training data into geographical areas;

estimating models for each partition; and

selecting a partitioned model based on the location of the portable device.

18. The computer-readable storage device of claim 14, the computer-readable storage device having additional instructions stored which result in the operations further comprising generating a user model by combining two models using interpolation.

19. The computer-readable storage device of claim 14 wherein entities in the local language model are related at one of a city, an area code, and a state level.

20. The computer-readable storage device of claim 14, the computer-readable storage device having additional instruction stored which result in the operations further comprising:

collecting queries at the portable device, wherein the queries are both textual and spoken;

analyzing the queries for a query location;

determining the query location from the analysis and the present location;

dividing the collection of queries into geographical areas, wherein the geographical areas are based on the determined query locations;

estimating language and acoustic models for each geographical area; and

training a speech recognition application using the estimated language and acoustic models.