Granted Patent B2

US 8,521,528 · App. 13/526,451 · Granted Aug 27, 2013

Abstract

Claims (51)

1. A system for distributed speech recognition, comprising:

audio data from a caller participating in a call with an agent; and

a processor to execute the following:

secondary recognizers, each secondary recognizer to receive the audio data and a reference that identifies a secondary grammar, wherein each secondary grammar is a non-overlapping section of a main grammar template, to perform speech recognition comprising identifying speech recognition results by applying the secondary grammar to the audio data and by selecting an n number of most likely speech recognition results; and

a main recognizer to receive the main grammar template and the audio data, to construct a new grammar using the speech recognition results from each of the secondary recognizers as a new vocabulary based on the main grammar template, and to identify further speech recognition results by applying the new grammar to the audio data.

2. A system according to claim 1 , further comprising:

a display to display the further speech recognition results to the agent.

3. A system according to claim 1 , further comprising:

a response to the audio data based on the further speech recognition results.

4. A system according to claim 1 , further comprising:

a prompt selected by the agent in response to the audio data from the caller, wherein the prompt comprises one of a file and a script.

5. A system according to claim 1 , further comprising:

a confidence level assigned to at least one of the further speech recognition results;

a predetermined threshold applied to the confidence levels; and

a placeholder assigned to the speech recognition results with confidence levels below the threshold, wherein the placeholder provides notification that further results are being generated.

6. A system according to claim 5 , wherein the further results are queried by at least one of the agent and automatically.

7. A system according to claim 1 , further comprising:

further grammar recognizers, each grammar recognizer to receive further grammar references from one of the secondary recognizers, to perform speech recognition on each of the further grammar references, and to identify a further n number of most likely speech recognition results; and

the secondary recognizers, each secondary recognizer to receive a secondary grammar template and to construct a new secondary grammar using the further n number of most likely speech recognition results from the associated further grammar recognizers, based on the secondary grammar templates.

8. A system according to claim 7 , wherein the new secondary grammar is used to perform speech recognition.

9. A system according to claim 1 , wherein each of the secondary recognizers is directly interfaced to the main recognizer as a second level of recognition hierarchy.

10. A system according to claim 1 , wherein the main recognizer and each of the secondary recognizers simultaneously receive the audio data.

11. A method for distributed speech recognition, comprising:

obtaining audio data from a caller participating in a call with an agent;

receiving on a main recognizer, a main grammar template and the audio data;

receiving on each of a plurality of secondary recognizers, the audio data and a reference that identifies a secondary grammar, wherein each secondary grammar is a non-overlapping section of the main grammar template;

performing speech recognition on each of the secondary recognizers, comprising:

identifying speech recognition results by applying the secondary grammar to the audio data; and

selecting an n number of most likely speech recognition results;

constructing by the main recognizer, a new grammar using the speech recognition results from each of the secondary recognizers as a new vocabulary based on the main grammar template; and

identifying further speech recognition results by applying the new grammar to the audio data.

12. A method according to claim 11 , further comprising:

displaying the further speech recognition results to the agent.

13. A method according to claim 11 , further comprising:

generating a response to the audio data based on the further speech recognition results.

14. A method according to claim 11 , further comprising:

playing a prompt selected by the agent in response to the audio data from the caller, wherein the prompt comprises one of a file and a script.

15. A method according to claim 11 , further comprising:

assigning a confidence level to at least one of the further speech recognition results;

applying a predetermined threshold to the confidence levels; and

assigning a placeholder to the speech recognition results with confidence levels below the threshold, wherein the placeholder provides notification that further results are being generated.

16. A method according to claim 15 , wherein the further results are queried by at least one of the agent and automatically.

17. A method according to claim 11 , further comprising:

receiving on one or more of the secondary recognizers, a secondary grammar template;

receiving on each of a plurality of further grammar recognizers, further grammar references to further grammars, wherein each further grammar is a non-overlapping section of the secondary grammar template received on the secondary recognizer;

performing speech recognition on each of the further grammars and identifying a further n number of most likely speech recognition results; and

constructing on the secondary recognizer, a new secondary grammar using the further speech recognition results based on the secondary grammar templates.

18. A method according to claim 17 , further comprising:

performing speech recognition using the new secondary grammar.

19. A method according to claim 11 , wherein each of the secondary recognizers is directly interfaced to the main recognizer as a second level of recognition hierarchy.

20. A method according to claim 11 , wherein the main recognizer and each of the secondary recognizers simultaneously receive the audio data.

Assignments (15)