Development system for a dialog system

• US 8,046,227 B2
• Filed: 09/05/2003
• Issued: 10/25/2011
• Est. Priority Date: 09/06/2002
• Status: Expired due to Fees

• Current Assignee: Telstra Corporation Limited (Telstra Group Ltd.)
• Original Assignee: Telestra Corporation Limited
• Inventors: Starkie, Bradford Craig
• Primary Examiner(s): CHAWAN, VIJAY B
• Application Number:
  US10/526,552
• Publication Number:
  US 20060203980A1
• Time in Patent Office: 2,972 Days
• Field of Search: 704/270.1, 704/270, 704/275, 704/257, 704/251, 704/231, 704/9, 706/11, 706/45, 706/61, 715/978, 715/841, 715/707, 705/26, 379/88.18, 379/265.05, 707/999.104, 717/104, 717/170, 719/328, 370/352
• US Class Current: 704/270.1
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 2209/545 Gui
• CPC Class Codes: G06F 8/30 Creation or generation of source code
• CPC Class Codes: G06F 8/38 for implementing user interfaces
• CPC Class Codes: G06F 8/65 Updates security arrangements therefor G06F21/57
• CPC Class Codes: H04M 2203/355 Interactive dialogue design tools, features or methods
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00
• CPC Class Codes: Y10S 715/978 Audio interaction as part of an operator interface

Abstract

A development system, including a scenario generator for generating a plurality of sample interactions representative of interactions between a dialog system and a user of the dialog system on the basis of definition data for the dialog system. The definition data includes state machine code, prompt data, and recognition data. A simulator generates prompts of sample interactions on the basis of the prompt data, and the scenario generator generates responses to the prompts on the basis of the recognition data. A scenario editor allows a user to modify the sample interactions. A grammar learner then updates the prompt data and the recognition data on the basis of the modified sample interactions. An application builder generates the definition data from specification data provided by a developer, an application template, and/or an existing dialog application.

Claims

1. A process for developing a dialog application for an interactive voice response system and executed by a computer system, comprising:

accessing application specification data for the dialog application, the application specification data defining tasks performed by the dialog application, items of information needed for the tasks, and information created by executing the tasks;

processing the application specification data to generate application definition data defining a dialog state machine, recognition grammars, and prompt grammars for the dialog application;

processing the application definition data to generate a plurality of example interactions representative of interactions between said dialog application and a user of said dialog application, wherein the plurality of the example interactions includes at least one example of a possible prompt corresponding to each rule of a plurality of rules of the prompt grammars, and at least one example of a possible user response corresponding to each rule of a plurality of rules of the recognition grammars;

providing the generated example interactions to a user for modification;

receiving the modified example interactions from the user;

processing the modified example interactions to update the recognition grammar, the prompt grammar, and the state machine;

generating the dialog application from the updated state machine and the updated recognition grammars and the updated prompt grammars; and

storing the generated dialog application in a memory device.

2. The process of claim 1, wherein said processing the application definition data to generate includes automatically generating a plurality of prompts on the basis of said application definition data, providing said prompts to a user, and receiving responses to said prompts from said user.

3. The process of claim 1, wherein said processing the application definition data to generate includes automatically generating a plurality of prompts on the basis of said application definition data, and automatically generating responses to said prompts on the basis of said application definition data.

4. The process of claim 3, wherein said processing the application definition data to generate a plurality of example interactions includes generating first example interactions, each of said first example interactions including interaction text and one item of application data;

and subsequently generating second sample interactions, each of said second sample interactions including interaction text and two or more items of application data.

5. The process of claim 3, wherein said processing the application definition data to generate a plurality of example interactions includes generating a plurality of prompts and respective lists of possible responses to said prompts, and selecting responses from said lists.

6. The process of claim 5, wherein said selecting includes randomly selecting unselected responses from said lists.

7. The process of claim 5, including identifying a dialog state with at least one unselected response, and determining a sequence of events for entering said dialog state.

8. The process of claim 7, wherein said sequence of events is determined on the basis of example interactions.

9. The process of claim 7, wherein said determining includes determining a first sequence of events for reaching an end dialog state and a second sequence of events for reaching said dialog state from an initial dialog state.

10. The process of claim 7, including determining two or more sequences of events for entering said dialog state and selecting from said two or more sequences to determine an event sequence with the least number of events.

11. The process of claim 7, wherein said determining a sequence of events for entering said dialog state includes providing one or more prompts to a user and receiving one or more responses from said user.

12. The process of claim 7, including generating events on the basis of said sequence to enter said dialog state.

13. The process of claim 3, wherein said automatically generating a plurality of prompts includes selecting grammar rules of said prompt grammars to generate prompts that are likely to have responses having the largest numbers of items of information.

14. The process of claim 1, wherein said example interactions include interaction text and application data.

15. The process of claim 1, wherein said recognition grammars and prompt grammars are updated using grammar learning.

16. The process of claim 1, including selecting prompts of said example interactions for updating only if said prompts have not been modified by a developer.

17. The process of claim 1, wherein said application definition data is generated on the basis of said application specification data and and at least one dialog application template.

18. A computer readable non-transitory storage medium having stored thereon program instructions for executing the process of claim 1.

19. A development system for developing a dialog application for an interactive voice response system, and comprising:

an application builder configured to process application specification data defining tasks performed by the dialog application, items of information needed for the tasks, and information created by executing the tasks to generate application definition data defining a dialog state machine, recognition grammars, and prompt grammars for the dialog application;

a scenario generator configured to process the application definition data to a plurality of example interactions representative of interactions between said dialog application and a user of said dialog application, wherein the plurality of the example interactions includes at least one example of a possible prompt corresponding to each rule of a plurality of rules of the prompt grammars, and at least one example of a possible user response corresponding to each rule of a plurality of rules of the recognition grammars;

a scenario editor including graphical user interface components for modifying said example interactions;

one or more learning components configured to process the modified example interactions to update the recognition grammars, the prompt grammars, and the state machine;

a code generator configured to generate the dialog application from the updated state machine, the updated recognition grammars, and the updated prompt grammars.

20. The development system of claim 19, including a simulator for generating prompts of said example interactions on the basis of said prompt grammars, and wherein said scenario generator is configured to generate responses to said prompts on the basis of said recognition grammars.

21. The development system of claim 19, wherein said application builder is configured to selectably generate said application definition data from said application specification data and, an application template, or from an existing application.

22. A dialogue application generated by the development system of claim 19.

23. The development system of claim 19, wherein said graphical user interface components include a component for deleting a selected example interaction.

24. The development system of claim 23, wherein said graphical user interface components include a component for a developer to select a predetermined reason why said selected example interaction was deleted.

25. The development system of claim 19, wherein said graphical user interface components include a component for defining a question and an answer to said question for display to a user of said dialog system.

26. The development system of claim 25, wherein said graphical user interface components include a question text component for defining question text and an answer text component for defining answer text.

27. The development system of claim 25, wherein said graphical user interface components include a component for selecting from a list of predetermined answers to said question.

28. The development system of claim 25, wherein said graphical user interface components include a component for specifying whether said question can be asked at any time.

29. The development system of claim 25, wherein said graphical user interface components include a component for specifying that said question can be asked during a selected example interaction.