Method and system for selecting grammars based on geographic information associated with a caller

• US 7,630,900 B1
• Filed: 12/01/2004
• Issued: 12/08/2009
• Est. Priority Date: 12/01/2004
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Tellme Networks Incorporated ([24]7.ai, Inc.)
• Inventors: Strom, Nikko
• Primary Examiner(s): Dorvil; Richemond
• Assistant Examiner(s): YEN, ERIC L
• Application Number:
  US11/002,003
• Time in Patent Office: 1,833 Days
• Field of Search: 704270-275
• US Class Current: 704/275
• CPC Class Codes: G10L 2015/228 of application context
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/4931 Directory assistance systems
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00

Abstract

A computer implemented method for automatically processing a data request comprising accessing a voice signal from a caller, determining geographic information associated with the caller from telephone network information associated with a call made by the caller, and retrieving a speech recognition grammar customized to the geographic information associated with the caller. The method further includes recognizing the voice signal by matching the voice signal to an entry of the speech recognition grammar and providing a directory listing to the caller based on the entry of the speech recognition grammar. The selected grammar may be customized to bias speech recognition to more frequently recognize cities local to the geographic information.

Claims

1. A method of a directory assistance application using voice recognition to provide information of a requested directory listing, the method, comprising:

identifying caller geographic information associated with a call using telephone network data provided with the call;

wherein, the caller geographic information is associated with a caller geographic locale;

selecting, from a plurality of grammars, a particular grammar that is customized for the caller geographic locale;

wherein the particular grammar comprises a plurality of entries, each of which represents a geographical location;

determining a speech recognition result by matching a voice signal of the call to an entry of the plurality of entries in the particular grammar;

wherein, in the particular grammar customized for the caller geographic locale, a local city entry in the plurality of entries includes the name of the local city without a state name within which the local city resides, the local city being local to the caller geographic locale;

wherein, in the particular grammar, a non-local city entry in the plurality of entries includes the name of the non-local city and another state name within which the non-local city resides, the non-local city being non-local to the caller geographic locale.

2. The method of claim 1, further comprising, assigning weighting values to two or more entries of the plurality of entries of the particular grammar.

3. The method of claim 2, wherein:

the weighting values are assigned according to call statistics.

4. The method of claim 1 wherein the telephone network data is one or more of, an area code and a cellular telephone tower associated with the call.

5. The method of claim 1 wherein the speech recognition result is a city or a city and a state represented by the entry that is matched to the voice signal.

6. The method of claim 1 wherein the plurality of grammars further comprise:

a first grammar comprising a first set of entries associated with a first plurality of weighting values which are customized for a first caller geographic locale; and

a second grammar comprising a second set of entries associated with a second plurality of weighting values which are customized for a second caller geographic locale.

7. The method of claim 6, wherein:

the first caller geographic locale is associated with a first area code; and

the second caller geographic locale is associated with a second area code.

8. The method of claim 2 wherein,

the weighting values are assigned according to a likelihood that the requested directory listing is located within the geographical locations represented by the two or more entries based on the caller geographic locale.

9. The method of claim 3 wherein,

the two or more entries represent a first city and a second city, respectively;

and

wherein, the weighting values are such that a first entry representing the first city that is more frequently requested by callers of the geographical locale is weighted higher in the particular grammar than a second entry representing the second city that is less frequently requested by callers.

10. The method of claim 5 further comprising:

retrieving the information of the requested directory listing based on the city or the city and the state represented by the entry that is matched to the voice signal of the caller; and

providing the information to the caller.

11. A method for performing speech recognition on a voice signal of a caller on a call in a directory assistance application to provide information of a requested directory listing, the method, comprising:

determining caller geographic information associated with the call using telephone network data;

wherein, the caller geographic information is associated with a caller geographic locale;

selecting a speech recognition grammar that is customized for the caller geographic locale, the speech recognition grammar comprising a plurality of entries, each of which represents a geographical location;

determining a speech recognition result by matching said voice signal to an entry of the plurality of entries in said speech recognition grammar;

wherein, the speech recognition result is an identified geographical region within which the requested directory listing is located; and

audibly providing the information of the requested directory listing to said caller;

wherein, in the speech recognition grammar customized for the caller geographic locale, a local city entry in the plurality of entries includes the name of the local city without a state name within which the local city resides, the local city being local to the caller geographic locale;

wherein, in the speech recognition grammar, a non-local city entry in the plurality of entries includes the name of the non-local city and another state name within which the non-local city resides, the non-local city being non-local to the caller geographic locale.

12. The method of claim 11 wherein said telephone network data includes caller identification information associated with said caller.

13. The method of claim 11 wherein said telephone network data comprises a cellular tower identifier.

14. The method of claim 11 further comprising:

retrieving the information of the requested directory listing based on the identified geographical region.

15. The method of claim 11 wherein said caller geographic information is obtained according to caller identification information associated with said call or a cellular tower identifier.

16. The method of claim 11 wherein, two or more entries of the plurality of entries are assigned weighting values, the weighting values assigned based on the caller geographic locale and according to a likelihood the caller will request information of a directory listing located within the geographical locations represented by the two or more entries.

17. The method of claim 11 further comprising:

modifying said speech recognition grammar based on said caller geographic information associated with said call.

18. A system using a speech recognition engine to provide directory assistance to a caller for information of a requested directory listing, the system, comprising:

a geographic determiner for determining caller geographic information using telephone network information associated with a call made by the caller;

wherein, the caller geographic information is associated with a caller geographic locale;

a grammar retriever for selecting a speech recognition grammar based on the caller geographic information;

wherein the speech recognition grammar comprises a plurality of entries, each of which represents a geographical location;

wherein, the speech recognition engine, is biased by the speech recognition grammar to select directory listings local to the caller geographic locale more frequently than directory listings outside the caller geographic locale;

the speech recognition engine recognizes the voice signal by matching the voice signal to an entry of the plurality of entries of the speech recognition grammar;

an output for providing, to the caller, the information of the requested directory listing located within the geographical location represented by the entry;

wherein, in the speech recognition grammar customized for the caller geographic locale, a local city entry in the plurality of entries includes the name of the local city without a state name within which the local city resides, the local city being local to the caller geographic locale;

wherein, in the speech recognition grammar, a non-local city entry in the plurality of entries includes the name of the non-local city and another state name within which the non-local city resides, the non-local city being non-local to the caller geographic locale.

19. The system of claim 18 wherein the telephone network data associated with said caller is determined by one or more of, a cellular tower identifier and an area code.

20. The system of claim 18 wherein the entry in the speech recognition grammar to which the voice signal of the caller is matched, is a city or a city and a state within which the requested directory listing is located.

21. The system of claim 18 wherein:

the speech recognition engine is provided with weighting values of two or more entries of the plurality of entries; and

wherein, the weighting values are assigned according to a likelihood that the requested directory listing is located within the geographical locations represented by the two or more entries determined by the caller geographic locale.

22. The system of claim 21 wherein:

the two or more entries represent a first city and a second city, respectively; and

wherein, the weighting values are assigned according to call statistics such that the first entry representing the first city that is more frequently requested by callers of the geographical locale is weighted higher in the speech recognition grammar than the second entry representing the second city that is less frequently requested by callers.

23. The system of claim 18 wherein the geographical location represented by the entry is a city or a combination of a city and state.

24. A computer readable medium comprising executable instructions which, when executed in a processing unit, causes the processing unit to perform a method of performing speech recognition on a voice signal of a caller on a call in a directory assistance application to provide information of a requested directory listing, the method, comprising:

identifying caller geographic information associated with the call using telephone network data provided with the call made;

wherein, the caller geographic information is associated with a caller geographic locale;

selecting, from a plurality of grammars, a particular grammar which is customized for the caller geographic locale;

wherein the particular grammar comprises a plurality of entries, each of which represents a geographical location;

wherein, two or more entries of the plurality of entries of the particular grammar are assigned weighting values, the weighting values assigned based on the caller geographic locale and according to a likelihood the caller will request information related to a directory listing located within the geographical locations represented by the two or more entries;

determining a speech recognition result by matching the voice signal using a speech recognition engine that is supplied with the particular grammar and using the weighting values to select, an entry of the plurality of entries of the particular grammar as the speech recognition result;

wherein, the speech recognition result is an identified city or city and state within which the requested directory listing is located;

retrieving the information of the requested directory listing based on the identified city or city and state; and

providing the information of the requested directory listing to the caller;

wherein, in the particular grammar customized for the caller geographic locale, a local city entry in the plurality of entries includes the name of the local city without a state name within which the local city resides, the local city being local to the caller geographic locale;

wherein, in the particular grammar, a non-local city entry in the plurality of entries includes the name of the non-local city and another state name within which the non-local city resides, the non-local city being non-local to the caller geographic locale.

25. The computer readable medium of claim 24 wherein the information includes a telephone number for the requested directory listing;

and

wherein, the caller is automatically connected to the requested directory listing.

26. The computer readable medium of claim 24 wherein said telephone network data is an area code associated with said call.

27. The computer readable medium of claim 24 wherein said telephone network data is a cellular telephone tower associated with said call.

28. The computer readable medium of claim 24 wherein the weighting values are assigned according to call statistics.

29. The computer readable medium of claim 28 wherein:

the weighting values are assigned such that entries representing cities that are more frequently requested by callers of a geographical locale identified by the caller geographic information are weighted higher in the particular grammar than entries representing other cities that are less frequently requested.