Sharing Intents to Provide Virtual Assistance in a Multi-Person Dialog

• US 20150149182A1
• Filed: 11/27/2013
• Published: 05/28/2015
• Est. Priority Date: 11/27/2013
• Status: Active Grant

• Current Assignee: SRI International, Inc.
• Original Assignee: SRI International, Inc.
• Inventors: Bercow, Douglas A., Carpenter, James F., Kalns, Edgar A.
• Granted Patent: US 10,096,316 B2
• Time in Patent Office: Days
• US Class Current: 704/275
• CPC Class Codes: G10L 15/18 using natural language modelling

Abstract

A computing system is operable as virtual personal assistant (VPA) to understand relationships between different instances of natural language dialog expressed by different people in a multi-person conversational dialog session. The VPA can develop a common resource, a shared intent, which represents the VPA'"'"'s semantic understanding of at least a portion of the multi-person dialog experience. The VPA can store and manipulate multiple shared intents, and can alternate between different shared intents as the multi-person conversation unfolds. With the shared intents, the computing system can generate useful action items and present the action items to one or more of the participants in the dialog session.

Claims

1. A machine-accessible storage medium of a computing system, the machine-accessible storage medium comprising computer code to cause the computing system to operate as a virtual personal assistant (“

VPA”

), the computer code comprising instructions executable by a processor of the computing system to;

detect a natural language dialog session involving a plurality of human participants;

determine semantic content of a first instance of conversational natural language input, the first instance of conversational natural language input supplied by one of the human participants in the dialog session;

develop a first intent, the first intent comprising a structured representation of the semantic content of the first instance of conversational natural language input;

access a shared dialog context, the shared dialog context storing a plurality of intents developed by the VPA from conversational natural language inputs supplied by different human participants in the dialog session;

select a second intent from the dialog context, the second intent comprising a structured representation of semantic content of a second instance of conversational natural language input, the second instance of conversational natural language input supplied by another of the human participants in the dialog session;

determine an intent mapping, the intent mapping defining a relationship between a portion of the first intent and a portion of the second intent;

create a shared intent by combining the portion of the first intent and the portion of the second intent based on the intent mapping; and

execute an action in response to the shared intent.

2. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to determine, with the intent mapping, a parameter of the first intent that maps to a parameter of the second intent, determine a data value associated with the parameter of the second intent, and include the data value associated with the parameter of the second intent in the shared intent.

3. The machine-accessible storage medium of claim 2, wherein the instructions are executable by the processor to determine a data value associated with the parameter of the first intent, and include both the data value associated with the parameter of the second intent and the data value associated with the parameter of the first intent in the shared intent.

4. The machine-accessible storage medium of claim 2, wherein the instructions are executable by the processor to determine a data value associated with the parameter of the first intent, and replace the data value associated with the parameter of the second intent with the data value associated with the parameter of the first intent in the shared intent.

5. The machine-accessible storage medium of claim 2, wherein the instructions are executable by the processor to determine a data value associated with the parameter of the first intent, compare the data value associated with the parameter of the first intent and the data value associated with the parameter of the second intent, and determine whether the human participants are in agreement as to the shared intent based on the comparison.

6. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to determine a mappable component of the first intent, the mappable component comprising one or more of:

a name of the first intent, a parameter of the first intent, and a data value associated with the parameter of the first intent, and to select the second intent based on the mappable component of the first intent.

7. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to create the shared intent as an updated version of the first intent that includes the portion of the second intent.

8. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to create the shared intent as an updated version of the second intent that includes the portion of the first intent.

9. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to update the shared intent after a plurality of instances of conversational natural language input supplied by the human participants in the dialog session.

10. The machine-accessible storage medium of claim 9, wherein the instructions are executable by the processor to execute the action after the shared intent is updated based on the plurality of instances of conversational natural language input.

11. The machine-accessible storage medium of claim 1, wherein the first and second intents are different types of intents that have at least one common parameter.

12. The machine-accessible storage medium of claim 1, wherein the first and second intents are different types of intents and the first and second intents have different parameters, and the instructions are executable by the processor to determine, with the intent mapping, a parameter of the first intent that maps to a parameter of the second intent.

13. The machine-accessible storage medium of claim 1, comprising instructions executable by the processor to determine, based on the shared intent, whether the dialog session involves a negotiation between two or more of the human participants.

14. The machine-accessible storage medium of claim 1, comprising instructions executable by the processor to create an electronic action item comprising semantic information extracted from the shared intent, and to present the electronic action item to one or more of the human participants.

15. The machine-accessible storage medium of claim 1, wherein at least one of the first intent and the second intent is a partial intent and the shared intent is a complete intent.

16. The machine-accessible storage medium of claim 1, wherein the instructions are executable by the processor to, in response to the shared intent, generate an electronic message and present the electronic message at a computing device of the computing system.

17. A method for developing a semantic understanding of a natural language dialog session involving at least two human participants, the method comprising, with a computing system:

determining semantic content of a first instance of natural language dialog input supplied by a first human participant in the dialog session,

determining that the semantic content of the first instance of natural language dialog input comprises first incomplete information relating to a common intention of the first and second human participants;

determining semantic content of a second instance of natural language dialog input supplied by a second human participant in the dialog session,

determining that the semantic content of the second instance of natural language dialog input comprises second incomplete information relating to the common intention of the first and second human participants; and

formulating a structured representation of the common intention of the first and second human participants, the structured representation comprising a portion of the semantic content of the first instance of natural language dialog input and a portion of the semantic content of the second instance of natural language dialog input.

18. The method of claim 17, comprising determining that the second incomplete information of the semantic content of the second instance of natural language dialog input relates to the first incomplete information of the semantic content of the first instance of natural language dialog input.

19. The method of claim 17, comprising, in response to formulating the structured representation of the common intention, executing an action on a computing device of the computing system.

20. The method of claim 17, comprising in response to formulating the structured representation of the common intention, generating an electronic document and presenting the electronic document at a computing device of the computing system.