Rapid development of virtual personal assistant applications

• US 9,081,411 B2
• Filed: 05/10/2013
• Issued: 07/14/2015
• Est. Priority Date: 05/10/2013
• Status: Active Grant

• Current Assignee: SRI International, Inc.
• Original Assignee: SRI International, Inc.
• Inventors: Kalns, Edgar T., Freitag, Dayne B., Mark, William S., Ayan, Necip Fazil, Wolverton, Michael J., Lee, Thomas J.
• Primary Examiner(s): Dao, Thuy
• Assistant Examiner(s): Nguyen, Mongbao
• Application Number:
  US13/891,858
• Publication Number:
  US 20140337814A1
• Time in Patent Office: 795 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 8/00 Arrangements for software engineering testing or debugging G06F11/36; administrative, planning or organisation aspects of software project management G06Q10/06
• CPC Class Codes: G06F 8/20 Software design
• CPC Class Codes: G06F 8/34 Graphical or visual programming
• CPC Class Codes: G06F 8/36 Software reuse

Abstract

A platform for developing a virtual personal assistant (“VPA”) application includes an ontology that defines a computerized structure for representing knowledge relating to one or more domains. A domain may refer to a category of information and/or activities in relation to which the VPA application may engage in a conversational natural language dialog with a computing device user. Re-usable VPA components may be linked to or included in the ontology. An ontology populating agent may at least partially automate the process of populating the ontology with domain-specific information. The re-usable VPA components may be linked with the domain-specific information through the ontology. A VPA application created with the platform may include domain-adapted re-usable VPA components that may be called upon by an executable VPA engine to determine a likely intended meaning of conversational natural language input of the user and/or initiate an appropriate system response to the input.

Claims

1. A method for developing a computerized virtual personal assistant (VPA) application for a domain of interest, the VPA application comprising an executable VPA engine and a re-usable VPA component accessible by the VPA engine to enable the VPA application to determine a likely intended goal of a computing device user based on conversational natural language input of the computing device user relating to the domain of interest, execute a task on behalf of the user, and/or generate a likely appropriate system response to the conversational natural language input, the method comprising, with a computing system:

analyzing content displayed on an Internet web page that supports electronic interactions with computing device users relating to the domain of interest;

identify domain-specific content displayed on the Internet web page;

determining a characteristic that the identified domain-specific content has in common with a computerized ontology that defines a structure for representing knowledge relating to a plurality of domains, wherein each domain refers to a category of information and/or activities in relation to which the VPA application may engage in a conversational natural language dialog with the computing device user;

establishing a data relationship between the identified domain-specific content and the ontology based on the common characteristic; and

adapting the re-usable VPA component to the domain of interest by applying at least a portion of the identified domain-specific content to the re-usable VPA component based on the data relationship between the content and the ontology.

2. The method of claim 1, wherein the re-usable VPA component is linked with at least a portion of the ontology and the method comprises selecting the re-usable VPA component based on the link between the re-usable VPA component and the ontology and the data relationship between the content and the ontology.

3. The method of claim 1, comprising creating a new VPA component by replacing a parameter of the re-usable VPA component with at least a portion of the content, and linking the new VPA component with the ontology.

4. The method of claim 1, comprising obtaining the content from the Internet web page in an automated fashion.

5. The method of claim 1, comprising learning the common characteristic based on one or more interactions between the computing system and a user of the computing system.

6. The method of claim 1, comprising learning the common characteristic in an automated fashion using a machine learning model.

7. The method of claim 1, comprising repeating the method for another domain of interest to create a VPA application for multiple domains of interest.

8. A development platform to at least partially automate the creation of the VPA application, the development platform embodied in one or more non-transitory machine readable storage media and comprising a plurality of instructions stored thereon that in response to being executed result in the computing system performing the method of claim 1.

9. A developer'"'"'s toolkit for developing a virtual personal assistant (VPA) computer application for a domain of interest, the developer'"'"'s toolkit comprising, embodied in one or more non-transitory machine accessible storage media:

a computerized ontology defining a structure for representing knowledge relating to a plurality of domains, each domain referring to a category of information and/or activities in relation to which the VPA computer application may engage in a conversational natural language dialog with a computing device user;

one or more computerized re-usable VPA components, each of the re- usable VPA components being linked with the ontology,

the one or more re-usable VPA components each accessible by an executable VPA engine to enable the VPA computer application to determine a likely intended goal of the computing device user based on conversational natural language input of the computing device user, execute a task on behalf of the computing device user, and/or generate a likely appropriate system response to the conversational natural language input; and

a computerized ontology populating agent to analyze content displayed on one or more Internet web pages that support electronic interactions with computing device users relating to the domain of interest, identify domain-specific content displayed on the Internet web page, determine a characteristic that the identified domain-specific content has in common with at least a portion of the ontology, and establish a data relationship between the identified domain-specific content and the ontology based on the common characteristic.

10. The developer'"'"'s toolkit of claim 9, wherein the ontology comprises a hierarchical structure of ontological concepts each representing a portion of the knowledge relating to the one or more domains, and each of the re-usable VPA components is linked with one or more of the ontological concepts in the hierarchical structure.

11. The developer'"'"'s toolkit of claim 9, wherein the ontology comprises a hierarchy of domain ontologies, and each of the domain ontologies defines a structure for representing knowledge relating to a specific domain.

12. The developer'"'"'s toolkit of claim 11, wherein each of the domain ontologies comprises a hierarchical structure of ontological concepts each representing a portion of the knowledge relating to the specific domain.

13. The developer'"'"'s toolkit of claim 9, wherein the one or more re-usable VPA components comprises one or more of a natural language grammar to interpret the intended goal of conversational natural language input of the computing device user, a user intent derivable from one or more natural language grammars, a task flow associated with one or more user intents, and a natural language response to communicate the result of a task flow to the computing device user in a conversational natural language form.

14. The developer'"'"'s toolkit of claim 9, wherein the ontology populating agent is configured to apply a computerized machine learning model to suggest the common characteristic or the data relationship.

15. The developer'"'"'s toolkit of claim 9, comprising a visualization module configured to present an interactive depiction of the ontology including interactive representations of the one or more re-usable VPA components.

16. A virtual personal assistant (VPA) computer application for a domain of interest, embodied in one or more non-transitory machine-accessible storage media, the VPA computer application comprising:

a user interface through which a conversational natural language dialog may be conducted between a computing device and a user to provide a service or information to the user with the computing device;

a VPA engine to conduct the conversational natural language dialog and initiate the providing of the service or information to the computing device user; and

a re-usable VPA component accessible by the VPA engine to enable the VPA computer application to determine a likely intended goal of the computing device user based on conversational natural language input of the computing device user, execute a task on behalf of the user, and/or generate a likely appropriate system response to the conversational natural language input;

wherein the re-usable VPA component is adapted for use in the domain of interest in an automated fashion by a computerized agent configured to analyze content displayed on an Internet web page, identify domain-specific content displayed on the Internet web page, and establish a data relationship between the identified domain-specific web content and the re-usable VPA component.

17. The VPA computer application of claim 16, wherein the re-usable VPA component is selected for use in the VPA computer application for the domain of interest based on a link between the re-usable VPA component and a computerized ontology, the ontology defines a structure for representing knowledge relating to the domain of interest, and the domain of interest refers to a category of information and/or activities in relation to which the VPA application may conduct the conversational natural language dialog with the computing device user.

18. The VPA computer application of claim 16, wherein the data relationship between the web content and the re-usable VPA component is established through a computerized ontology, the ontology defines a structure for representing knowledge relating to the domain of interest, and the domain of interest refers to a category of information and/or activities in relation to which the VPA application may conduct the conversational natural language dialog with the computing device user.

19. The VPA computer application of claim 16, wherein the domain of interest is a sub-domain of an e-commerce domain, the re-usable VPA component is linked with a computerized ontology for the e-commerce domain, and the re-usable VPA component is adapted for use in the VPA application for the sub-domain of the e-commerce domain based on a data relationship between the web content and the ontology for the e-commerce domain.