Method and system for prompt construction for selection from a list of acoustically confusable items in spoken dialog systems

• US 8,909,528 B2
• Filed: 05/09/2007
• Issued: 12/09/2014
• Est. Priority Date: 05/09/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Eide, Ellen Marie, Goel, Vaibhava, Gopinath, Ramesh, Stewart, Osamuyimen T.
• Primary Examiner(s): Desir, Pierre-Louis
• Assistant Examiner(s): Serrou, Abdelali
• Application Number:
  US11/746,087
• Publication Number:
  US 20080281598A1
• Time in Patent Office: 2,771 Days
• Field of Search: 704/270, 704/270.1, 704/246, 704/257, 704/251, 704/256, 704/255, 704/9, 704/243, 704/E15.021, 704/E15.024
• US Class Current: 704/251
• CPC Class Codes: G10L 15/187 Phonemic context, e.g. pronunciation rules, phonotactical constraints or phoneme n-grams
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A method (and system) of determining confusable list items and resolving this confusion in a spoken dialog system includes receiving user input, processing the user input and determining if a list of items needs to be played back to the user, retrieving the list to be played back to the user, identifying acoustic confusions between items on the list, changing the items on the list as necessary to remove the acoustic confusions, and playing unambiguous list items back to the user.

Claims

1. A method of providing a list of items in a spoken dialog system comprising a plurality of disambiguation strategies, said method comprising:

receiving input speech;

processing said input speech to determine if a clarification of the input speech is desired because the spoken dialog system has returned at least two speech recognition hypotheses having similar confidence values for at least a portion of the input speech;

retrieving, if clarification is desired, a first list of items to be played back to the user;

identifying acoustically confusable items on said first list of items using at least one measure of confusability;

selecting, based, at least in part, on at least one rule in a collection of rules of the spoken dialog system, a disambiguation strategy from the plurality of disambiguation strategies, wherein at least two of the disambiguation strategies in the plurality of disambiguation strategies each includes presenting at least two choices to the user and asking the user to select one of the at least two choices, wherein the plurality of disambiguation strategies includes a first disambiguation strategy comprising spelling at least a portion of each of at least two of the items in the first list of items and a second disambiguation strategy comprising repeating at least two of the items in the first list of items and identifying a first letter of a word in each of the at least two of the items;

generating a disambiguated list of items by modifying at least one of said acoustically confusable items on said first list according to said selected disambiguation strategy; and

playing a prompt comprising the disambiguated list of items back to the user.

2. The method according to claim 1, wherein the at least one measure of confusability is based on phonetic content of said items on the first list of items.

3. The method according to claim 2, wherein said at least one measure of confusability is customized to a playback system to resolve acoustically confusable items specific to the playback system.

4. The method according to claim 1, further comprising:

identifying commonly occurring recognition errors in previous calls; and

wherein the at least one measure of confusability is based, at least in part, on the commonly occurring recognition errors.

5. The method according to claim 1, wherein said disambiguation strategy is selected based on minimal features that distinguish the acoustically confusable items.

6. The method according to claim 1, wherein the disambiguation strategy is selected based on a type of acoustic confusion between the acoustically confusable items.

7. The method according to claim 1, wherein said disambiguation strategy is selected based on a quantity of said acoustically confusable items.

20. The method of claim 1, wherein the at least one measure of confusability is based on orthography of said items on the first list of items.

23. The method of claim 1, further comprising:

selecting one of the at least two of the disambiguation strategies in the plurality of disambiguation strategies that includes presenting at least two choices to the user and asking the user to select one of the at least two choices.

8. A system comprising:

at least one storage medium configured to store a plurality of machine-readable instructions; and

at least one processor programmed to execute the plurality of machine-readable instructions to perform a method comprising;

processing input speech to determine if clarification of the input speech is desired because the spoken dialog system has returned at least two speech recognition hypotheses having similar confidence values for at least a portion of the input speech;

retrieving, if clarification is desired, a first list of items to be played back to the user;

identifying acoustically confusable items on the first list of items using at least one measure of confusability;

selecting, based, at least in part, on at least one rule in a collection of rules, a disambiguation strategy from a plurality of disambiguation strategies, wherein of the plurality of disambiguation strategies includes a first disambiguation strategy comprising spelling at least a portion of each of at least two of the items in the first list of items and a second disambiguation strategy comprising repeating at least two of the items in the first list of items and identifying a first letter of a word in each of the at least two of the items;

generating a disambiguated list of items by modifying at least one of the acoustically confusable items on the first list according to the disambiguation strategy; and

playing a prompt comprising the disambiguated list of items back to the user.

9. The system of claim 8, wherein the at least one measure of confusability is based on phonetic content of said items on the first list of items.

10. The system of claim 8, wherein the at least one measure of confusability is customized to a playback system to resolve acoustically confusable items specific to the playback system.

11. The system of claim 8, wherein the method further comprises:

identifying commonly occurring recognition errors in previous calls; and

wherein the at least one measure of confusability is based, at least in part, on the commonly occurring recognition errors.

12. The system of claim 8, wherein the disambiguation strategy is selected based on minimal features to distinguish the acoustically confusable items.

13. The system of claim 8, wherein the disambiguation strategy is selected based on a type of acoustic confusion between the acoustically confusable items.

14. The system of claim 8, wherein the disambiguation strategy is selected based on a quantity of the acoustically confusable items.

21. The system of claim 8, wherein the at least one measure of confusability is based on orthography of said items on the first list of items.

24. The system of claim 8, wherein the first disambiguation strategy further comprises spelling at least a portion of each of the items in the first list of items and asking the user to select one of the spelled items.

15. At least one non-transitory computer-readable storage medium encoded with a plurality of machine-readable instructions that, when executed by a computer perform a method comprising:

processing input speech to determine if clarification of the input speech is desired because the spoken dialog system has returned at least two speech recognition hypotheses having similar confidence values for at least a portion of the input speech;

retrieving, if clarification is desired, a first list of items to be played back to the user;

identifying acoustically confusable items on the first list of items using at least one measure of confusability;

selecting, based, at least in part, on at least one rule in a collection of rules, a disambiguation strategy from a plurality of disambiguation strategies, wherein the disambiguation strategy is selected based on a type of acoustic confusion between the acoustically confusable items on the first list of items, wherein the plurality of disambiguation strategies includes a first disambiguation strategy comprising spelling at least a portion of each of at least two of the items in the first list of items and a second disambiguation strategy comprising repeating at least two of the items in the first list of items and identifying a first letter of a word in each of the at least two of the items;

generating a disambiguated list of items by modifying at least one of the acoustically confusable items on the first list according to the disambiguation strategy; and

playing a prompt comprising the disambiguated list of items back to the user.

16. The at least one computer-readable storage medium of claim 15, wherein the at least one measure of confusability is based on phonetic contents of said items on the first list of items.

17. The at least one computer-readable storage medium of claim 15, wherein the at least one measure of confusability is customized to a playback system to resolve acoustically confusable items specific to the playback system.

18. The at least one computer-readable storage medium of claim 15, wherein the at least one processor is further programmed to:

identify commonly occurring recognition errors in previous calls; and

wherein the at least one measure of confusability is based, at least in part, on the commonly occurring recognition errors.

19. The at least one computer-readable storage medium of claim 15, wherein the disambiguation strategy is selected based on minimal features to distinguish the acoustically confusable items.

22. The at least one computer-readable storage medium of claim 15, wherein the at least one measure of confusability is based on orthography of said items on the first list of items.