System and method for representing and resolving ambiguity in spoken dialogue systems

• US 20030233230A1
• Filed: 06/12/2002
• Published: 12/18/2003
• Est. Priority Date: 06/12/2002
• Status: Abandoned Application

• Current Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Ammicht, Egbert, Potamianos, Alexandros, Fosler-Lussier, J. Eric
• Application Number:
  US10/170,510
• Publication Number:
  US 20030233230A1
• Time in Patent Office: Days
• US Class Current: 704/235
• CPC Class Codes: G10L 15/183 using context dependencies, e.g. language models
• CPC Class Codes: G10L 15/19 Grammatical context, e.g. disambiguation of the recognition hypotheses based on word sequence rules
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A system for, and method of, representing and resolving ambiguity in natural language text and a spoken dialogue system incorporating the system for representing and resolving ambiguity or the method. In one embodiment, the system for representing and resolving ambiguity includes: (1) a context tracker that places the natural language text in context to yield candidate attribute-value (AV) pairs and (2) a candidate scorer, associated with the context tracker, that adjusts a confidence associated with each candidate AV pair based on system intent.

Claims

1. A system for representing and resolving ambiguity in natural language text, comprising:

a context tracker that places said natural language text in context to yield candidate attribute-value (AV) pairs; and

a candidate scorer, associated with said context tracker, that adjusts a confidence associated with each candidate AV pair based on system intent.

2. The system as recited in claim 1 wherein said natural language text is selected from the group consisting of:

recognized spoken language, and

typed text.

3. The system as recited in claim 1 wherein said context tracker models value ambiguities and position ambiguities with respect to said natural language text.

4. The system as recited in claim 1 wherein said candidate scorer analyzes raw data to adjust said confidence.

5. The system as recited in claim 1 wherein said candidate scorer comprises a pragmatic analyzer that conducts a pragmatic analysis to adjust said confidence.

6. The system as recited in claim 1 wherein said candidate scorer matches at least one hypothesis to a current context to adjust said confidence.

7. The system as recited in claim 1 further comprising an override subsystem that allows a user to provide explicit error correction to said system.

8. A method of representing and resolving ambiguity in natural language text, comprising:

placing said natural language text in context to yield candidate attribute-value (AV) pairs; and

adjusting a confidence associated with each candidate AV pair based on system intent.

9. The method as recited in claim 8 wherein said natural language text is selected from the group consisting of:

recognized spoken language, and

typed text.

10. The method as recited in claim 8 wherein said placing comprises modeling value ambiguities and position ambiguities with respect to said natural language text.

11. The method as recited in claim 8 wherein said adjusting comprises analyzing raw data.

12. The method as recited in claim 8 wherein said adjusting comprises conducting a pragmatic analysis.

13. The method as recited in claim 8 wherein said adjusting comprises matching at least one hypothesis to a current context.

14. The method as recited in claim 8 further comprising allowing a user to provide explicit error correction to said system.

15. A spoken dialogue system, comprising:

a speech recognizer that recognizes spoken language received from a user;

a parser, coupled to said recognizer, that parses said recognized spoken language;

an interpreter that further processes said recognized spoken language to yield natural language text;

a context tracker that places said natural language text in context to yield candidate attribute-value (AV) pairs;

a candidate scorer, associated with said context tracker, that adjusts a confidence associated with each candidate AV pair based on system intent; and

a voice responder that generates spoken language back to said user.

16. The system as recited in claim 15 wherein said context tracker models value ambiguities and position ambiguities with respect to said natural language text.

17. The system as recited in claim 15 wherein said candidate scorer analyzes raw data to adjust said confidence.

18. The system as recited in claim 15 wherein said candidate scorer comprises a pragmatic analyzer that conducts a pragmatic analysis to adjust said confidence.

19. The system as recited in claim 15 wherein said candidate scorer matches at least one hypothesis to a current context to adjust said confidence.

20. The system as recited in claim 15 further comprising an override subsystem that allows a user to provide explicit error correction to said system.