System and method for disambiguating multiple intents in a natural language dialog system

• US 9,454,960 B2
• Filed: 04/13/2015
• Issued: 09/27/2016
• Est. Priority Date: 09/27/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: AT&T Intellectual Property I LP (AT&T, Inc.)
• Inventors: Stewart, Osamuyimen Thompson
• Primary Examiner(s): Serrou, Abdelali
• Application Number:
  US14/684,880
• Publication Number:
  US 20150221304A1
• Time in Patent Office: 533 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/18 using natural language modelling
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence

Abstract

The present invention addresses the deficiencies in the prior art by providing an improved dialog for disambiguating a user utterance containing more than one intent. The invention comprises methods, computer-readable media, and systems for engaging in a dialog. The method embodiment of the invention relates to a method of disambiguating a user utterance containing at least two user intents. The method comprises establishing a confidence threshold for spoken language understanding to encourage that multiple intents are returned, determining whether a received utterance comprises a first intent and a second intent and, if the received utterance contains the first intent and the second intent, disambiguating the first intent and the second intent by presenting a disambiguation sub-dialog wherein the user is offered a choice of which intent to process first, wherein the user is first presented with the intent of the first or second intents having the lowest confidence score.

Claims

1. A method comprising:

receiving, via an interactive voice recognition system, a user utterance and converting the user utterance to text;

generating multiple intents based on the text;

establishing, via the interactive voice recognition system, a respective confidence score for each intent in the multiple intents, wherein the respective confidence score for each intent is based on how much training data corresponding to the each intent was used to train a spoken language understanding module, wherein more training data corresponds to a higher confidence;

identifying a first intent and a second intent having confidence scores above a threshold, wherein the first intent and the second intent have a highest two confidence scores in the multiple intents and wherein both the first intent and the second intent are meant to be implemented; and

disambiguating the first intent and the second intent by presenting a disambiguation sub-dialog, via the interactive voice recognition system, wherein a user is offered a choice of which intent to process first, and wherein the disambiguating further comprises concatenating a first prompt associated with the first intent and a second prompt associated with the second intent, wherein the first prompt and the second prompt are predefined prompts from predefined prompts associated with intents.

2. The method of claim 1, wherein the user is first presented with one of the first intent and the second intent having a lowest confidence score between the first intent and the second intent.

3. The method of claim 1, wherein the disambiguation sub-dialog presents one of the first intent and the second intent having a highest confidence score between the first intent and the second intent last.

4. The method of claim 1, further comprising:

receiving a disambiguation utterance from the user clarifying which of the first intent and the second intent should be processed first.

5. The method of claim 1, wherein the user utterance comprises the first intent and the second intent.

6. The method of claim 1, wherein the user utterance comprises a customer service representative request plus an intent.

7. The method of claim 6, wherein when the user utterance comprises a customer service representative request plus the first intent and the second intent, then disambiguating the first intent and the second intent further comprises concatenating prompts from a table, wherein one of the first intent and the second intent having the lowest confidence score between the first intent and the second intent is played first and one of the first intent and the second intent having a highest confidence score between the first intent and the second intent is played last.

8. 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, via an interactive voice recognition system, a user utterance and converting the user utterance to text;

generating multiple intents based on the text;

establishing, via the interactive voice recognition system, a respective confidence score for each intent in the multiple intents, wherein the respective confidence score for each intent is based on how much training data corresponding to the each intent was used to train a spoken language understanding module, wherein more training data corresponds to a higher confidence;

identifying a first intent and a second intent having confidence scores above a threshold, wherein the first intent and the second intent have a highest two confidence scores in the multiple intents and wherein both the first intent and the second intent are meant to be implemented; and

disambiguating the first intent and the second intent by presenting a disambiguation sub-dialog, via the interactive voice recognition system, wherein a user is offered a choice of which intent to process first, and wherein the disambiguating further comprises concatenating a first prompt associated with the first intent and a second prompt associated with the second intent, wherein the first prompt and the second prompt are predefined prompts from predefined prompts associated with intents.

9. The system of claim 8, wherein the user is first presented with one of the first intent and the second intent having a lowest confidence score between the first intent and the second intent.

10. The system of claim 8, wherein the disambiguation sub-dialog presents one of the first intent and the second intent having a highest confidence score between the first intent and the second intent last.

11. The system of claim 8, the computer-readable storage medium having additional instructions stored which, when executed by the processor, result in the processor perform additional operations comprising:

receiving a disambiguation utterance from the user clarifying which of the first intent and the second intent should be processed first.

12. The system of claim 8, wherein the user utterance comprises the first intent and the second intent.

13. The system of claim 8, wherein the user utterance comprises a customer service representative request plus an intent.

14. The system of claim 13, wherein when the received user utterance comprises a customer service representative request plus the first intent and the second intent, then disambiguating the first intent and the second intent further comprises concatenating prompts from a table, wherein one of the first intent and the second intent having the lowest confidence score between the first intent and the second intent is played first and one of the first intent and the second intent having a highest confidence score between the first intent and the second intent is played last.

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

receiving, via an interactive voice recognition system, a user utterance and converting the user utterance to text;

generating multiple intents based on the text;

establishing, via the interactive voice recognition system, a respective confidence score for each intent in the multiple intents, wherein the respective confidence score for each intent is based on how much training data corresponding to the each intent was used to train a spoken language understanding module, wherein more training data corresponds to a higher confidence;

identifying a first intent and a second intent having confidence scores above a threshold, wherein the first intent and the second intent have a highest two confidence scores in the multiple intents and wherein both the first intent and the second intent are meant to be implemented; and

disambiguating the first intent and the second intent by presenting a disambiguation sub-dialog, via the interactive voice recognition system, wherein a user is offered a choice of which intent to process first, and wherein the disambiguating further comprises concatenating a first prompt associated with the first intent and a second prompt associated with the second intent, wherein the first prompt and the second prompt are predefined prompts from predefined prompts associated with intents.

16. The computer-readable storage device of claim 15, wherein the user is first presented with one of the first intent and the second intent having a lowest confidence score between the first intent and the second intent.

17. The computer-readable storage device of claim 15, wherein the disambiguation sub-dialog presents one of the first intent and the second intent having a highest confidence score between the first intent and the second intent last.

18. The computer-readable storage device of claim 15, having additional instructions stored which, when executed by the processor, result in the processor perform additional operations comprising:

receiving a disambiguation utterance from the user clarifying which of the first intent and the second intent should be processed first.

19. The computer-readable storage device of claim 15, wherein the user utterance comprises the first intent and the second intent.

20. The computer-readable storage device of claim 15, wherein the user utterance comprises a customer service representative request plus an intent.