Method and apparatus for providing speech recognition resolution on a database

• US 7,752,048 B2
• Filed: 05/27/2005
• Issued: 07/06/2010
• Est. Priority Date: 05/27/2005
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Sprague, Karen Lynn, Vora, Ashish
• Primary Examiner(s): Chawan; Vijay B
• Application Number:
  US11/140,647
• Publication Number:
  US 20060271361A1
• Time in Patent Office: 1,866 Days
• Field of Search: 704/207, 704/255, 704/256, 704/247, 704/9, 704/251, 704/236, 704/239, 704/240, 704/201, 379/88.04, 379/88.02, 709/218, 705/7, 705/35
• US Class Current: 704/255
• CPC Class Codes: G06F 16/2448 for particular applications; for extensibility, e.g. user defined types
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes

Abstract

A method of providing speech recognition resolution in a database includes receiving an utterance from an end-user or application, and dispatching it to the database which is coupled to or has access to a speech recognition technology. The method further includes converting the utterance to an intermediate form suitable for searching, and transmitting the intermediate form from the database to the speech recognition technology. An item-matching is then performed via the speech recognition technology which returns items matching the utterance to the database along with a confidence metric, and the utterance matches are then provided from the database in the form of an ordered result set.

Claims

1. A method for providing speech recognition in a database, comprising:

receiving, with a first processor, an utterance from an end-user or an input application;

converting the utterance to an intermediate form suitable for searching using speech recognition technology;

receiving, with a database, intermediate form of the utterance that is suitable for searching using speech recognition technology;

transmitting the intermediate form from the database to a speech recognition technology, where the speech recognition technology comprises a second processor;

causing the speech recognition technology to perform item-matching as between the intermediate form and voice searchable representations of data stored in the database via the speech recognition technology, and to return items matching the intermediate form to the database; and

providing data associated with the one or more of the returned items in the form of an ordered result set.

2. The method as in claim 1 wherein the database is initialized by generating at least one shadow table that stores the voice searchable representations associated with data in the database.

3. The method as in claim 2 wherein the dispatching step further comprises utilizing an SQL function having at least a representation of the utterance as an argument.

4. The method as in claim 2 wherein the items matching the intermediate forms are returned to the database as a graded shadow table.

5. The method as in claim 1 wherein the speech recognition technology is large scale speech recognition technology.

6. The method as in claim 1 comprising returning a confidence metric for each returned item and where the providing one or more of the returned items is performed by selecting one or more items matching the intermediate form for return based, at least in part, on the confidence metric associated with the matching item.

7. A method for providing speech resolution in a database, the method comprising:

causing a processor to determine a voice searchable form for data in a database;

electronically storing, in the database, for use by speech recognition technology, a respective determined voice searchable form for respective data in a manner that associates the voice searchable form with the data;

providing access to the voice searchable form to speech recognition technology for item matching as between an utterance received as part of a query on the database and the voice searchable form.

8. The method as in claim 7 where the associating of a voice searchable form of data is performed by constructing a shadow table that includes a transcription of an utterance corresponding to the data, the method further comprising maintaining the database by automatically updating the shadow tables in response to changes to the database.

9. The method as in claim 7 wherein the speech recognition technology is large scale speech recognition technology.

10. The method as in claim 7 further comprising providing a parameter in the database to flag data which is to be transcribed for a shadow table and linked thereto.

11. The method as in claim 7 wherein the returned match results are converted into an ordered result set by the database.

12. A system for providing speech recognition resolution within a database comprising:

means for receiving an utterance from an end-user or an input application;

means for converting the utterance to an intermediate form suitable for searching using speech recognition technology;

means associated with the database for performing item-matching as between the intermediate form and voice searchable representations of data associated with items in the database via a speech recognition search algorithm, and returning items matching the intermediate form to the data base; and

means associated with the database for converting the returned items into an ordered result set.

13. The system as in claim 12 wherein the database further comprises means for initializing by generating at least one shadow table that stores the voice searchable representations of data associated with items in the database.

14. A computer-readable medium having computer-executable instructions stored thereon that if executed by a computer cause the computer to perform a method for providing speech recognition in a database, the method comprising:

receiving an intermediate form of an utterance suitable for searching using speech recognition technology from an end-user or an input application with a database coupled to a speech recognition technology;

with the database, providing the intermediate form to speech recognition technology, where the speech recognition technology performs item-matching using, and receiving items matching the intermediate form from the speech recognition technology; and

returning one or more matching items the form of an ordered result set for communication to the user.

15. The computer-readable medium having computer-executable instructions for performing the method as in claim 14 wherein the speech recognition technology is a large scale speech recognition technology.

16. The computer-readable medium having computer-executable instructions for performing the method as in claim 15 wherein the database is initialized by adding to the database at least one shadow table including a transcription of data in the database to a searchable form and a link to the corresponding data in the database.

17. The computer-readable medium as in claim 16 wherein the item matching is performed by the search recognition technology by matching the shadow table data to the intermediate form.

18. The computer-readable medium of claim 17 wherein the receiving step comprises receiving an SQL function having at least a representation of the utterance as an argument.

19. The computer-readable medium of claim 17 further comprising updating the shadow table automatically in response to changes in the database.

20. The computer-readable medium of claim 17 wherein the item matching compares the shadow table to the intermediate form of the utterance.