System and method for supporting natural language queries and requests against a user's personal data cloud

• US 9,471,666 B2
• Filed: 01/18/2012
• Issued: 10/18/2016
• Est. Priority Date: 11/02/2011
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Singh, Rajan, Donneau-Golencer, Thierry, Hulen, Corey, Mark, William Scott
• Primary Examiner(s): Uddin, Md. I
• Application Number:
  US13/353,237
• Publication Number:
  US 20130185336A1
• Time in Patent Office: 1,735 Days
• Field of Search: 707/706, 707/769, 707/790, 707/794, 707/609, 707/737, 707/741
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries
• CPC Class Codes: G06F 16/3329 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/337 Profile generation, learning or modification
• CPC Class Codes: G06F 16/338 Presentation of query results
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A machine-implemented method for supporting a natural language user request against a user'"'"'s personal data cloud can include a machine receiving the natural language user request from the user, determining a semantic interpretation of the natural language user request, querying a semantically-indexed, integrated knowledge store based on the semantic interpretation, and responding to the natural language user request by displaying results of the querying, wherein the results correspond to an item within the user'"'"'s personal data cloud.

Claims

1. A machine-implemented method for supporting a natural language user request against a user'"'"'s personal data cloud, the method comprising:

a machine extracting personal textual data identifying names, locations, and/or contact information for persons and places from electronic mail (email) messages, contact data, or calendar data stored on different network sources in the user'"'"'s personal data cloud;

the machine constructing a semantically-indexed integrated knowledge store for storage and future retrieval of the personal textual data;

the machine extracting additional information from other network sources;

the machine correlating the additional information from the other network sources with the previously stored personal textual data and generating additional personal textual data for the semantically-indexed integrated knowledge store identifying additional names, locations, and/or contact information for the persons and places;

the machine receiving the natural language user request from the user;

the machine determining a semantic interpretation of the natural language user request, wherein determining the semantic interpretation comprises referencing a stored ontology that defines a semantic relationship among a set of personal data terminology;

the machine querying the semantically-indexed integrated knowledge store based at least in part on the semantic interpretation; and

the machine responding to the natural language user request by displaying one or more results of the querying, wherein the one or more results include at least some of the names, locations, and/or contact information identified by the semantically-indexed integrated knowledge store;

the machine displaying a suggestions panel based on the semantically-indexed integrated knowledge store offering refinements for types of personal textual data available responsive to the natural language user request; and

the machine querying the semantically-indexed integrated knowledge store based on selected ones of the refinements.

2. The machine-implemented method of claim 1, wherein the personal textual data includes user data from calendar events, shared documents.

3. The machine-implemented method of claim 1, wherein the network sources comprise network-based user accounts associated with, social networks, file sharing, and customer relationship management CRM.

4. The machine-implemented method of claim 1, wherein constructing the semantically-indexed integrated knowledge store comprises an automated semantic analysis of the personal textual data.

5. The machine-implemented method of claim 4, wherein the automated semantic analysis of the personal textual data comprises one or both of automatic clustering and tagging operations.

6. The machine-implemented method of claim 1, wherein the natural language user request includes a natural language query, a natural language command, or both.

7. The machine-implemented method of claim 1, further comprising:

the machine displaying a list of calendar items in response to the natural language user request;

the machine querying the semantically-indexed integrated knowledge store for attendees, documents, and messages associated with a selected one of the calendar items; and

the machine displaying the attendees, documents, and messages.

8. The machine-implemented method of claim 1, further comprising receiving and responding to a non-natural language user request received from the user.

9. The machine-implemented method of claim 1, further comprising refining the stored ontology over time based at least in part on machine learning.

10. The machine-implemented method of claim 1, further comprising refining the stored ontology over time based at least in part on interactive user feedback, wherein the interactive user feedback comprises at least one from a group consisting of:

a star rating mechanism, a thumbs-up or thumbs-down mechanism, and a numbered scale mechanism.

11. The machine-implemented method of claim 1, wherein the network sources in the user'"'"'s personal data cloud correspond to more than one user account, and wherein at least one of the more than one user account corresponds to a different cloud service than a different account of the more than one user account.

12. A system, comprising:

a storage device; and

a processor configured to operate a machine-implemented data extractor and correlator configured to;

extract personal data from a first heterogeneous group of network sources, wherein at least one of the network sources of the first heterogeneous group corresponds to a different user account than another one of the network sources of the first heterogeneous group, the personal data including names, locations, and/or contact information associated persons and places;

construct a semantically-indexed knowledge store in the storage device for future retrieval of the personal data associated with the persons and places;

extract additional information from documents in a second group of network sources that is different than the first heterogeneous group of network sources, wherein the second group of network sources includes at least one network source and each one of said at least one network source is different than each network source of the first heterogeneous group of network sources, the additional information extracted from one(s) of the network sources of the second group that are associated with the persons and places;

correlate the additional information with the semantically-indexed knowledge store to identify additional names, locations, and/or contact information associated with the persons and places;

store the additional information as part of the personal data associated with the persons and places;

receive a natural language user request from a user;

determine a semantic interpretation of the natural language user request by reference to a stored ontology that defines a semantic relationship among a set of personal data terminology;

query the semantically-indexed integrated knowledge store based on the semantic interpretation of the natural language user request to identify results including at least some of the names, locations, and/or contact information associated with the persons or places;

display a suggestions panel based on the semantically-indexed integrated knowledge store to offer refinements for types of personal textual data available responsive to the natural language request; and

if a selection from the refinements is received from a user, query the semantically-indexed integrated knowledge store based on the selection.

13. The system of claim 12, wherein the data extractor and correlator is further configured to present a ranked list of alternative potential responses to the natural language user request.

14. The system of claim 13, the data extractor and correlator is further configured to refine the ranked list based on one or more additional user inputs solicited from the user.

15. The system of claim 12, wherein the data extractor and correlator is further configured to determine a ranked list of online documents containing a probable answer responsive to the natural language user request.

16. The system of claim 12, wherein the data extractor and correlator is further configured to determine at least one probable answer responsive to the natural language user request, and present to the user the at least one probable answer.

17. The system of claim 12, wherein the data extractor and correlator is further configured to invoke one or more of a plurality of network services to perform a desired action responsive to the natural language user request.

18. The system of claim 17, wherein the plurality of network services comprises an alert creation mechanism configured to create an alert, and wherein the desired action comprises creation of the alert.

19. The system of claim 17, wherein the plurality of network services comprises a reminder creation mechanism configured to create a reminder, and wherein the desired action comprises creation of the reminder.

20. The system of claim 12, wherein the data extractor and correlator is further configured to integrate the results with one or more additional results obtained by querying one or more network information sources that are external to the knowledge store, each of said one or more network information sources different than each network source of the first heterogeneous group.

21. The system of claim 20, wherein the one or more external network information sources comprise a geographic positioning system (GPS) source, and wherein the one or more additional results comprises a geographic location corresponding to at least one of the results.

22. The system of claim 12, wherein the natural language user request is received by way of a voice command from the user.

23. The system of claim 12, wherein the data extractor and correlator is further configured to present suggested actions to the user responsive to the user activating a help feature.

24. The system of claim 12, wherein the data extractor and correlator is further configured to sort the results responsive to a command from the user.

25. The system of claim 12, wherein the extractor and correlator is further configured to filter the results responsive to a command from the user.

26. The system of claim 12, wherein the extractor and correlator is further configured to disambiguate the natural language user request.

27. The system of claim 26, wherein the disambiguating comprises presenting at least one element of the natural language user request in a manner distinct from the remainder of the natural language user request.

28. The system of claim 12, wherein the extractor and correlator is to receive a user edit of the natural language user request.

29. The system of claim 12, wherein the results include a link to an external website.

30. The system of claim 12, wherein the data extractor and correlator is further configured to receive a follow-up query from the user.

31. The system of claim 12, wherein the extractor and correlator is further configured to receive a phone number responsive to the user selecting the results.

32. The system of claim 12, wherein the results include at least one image corresponding to at least one other user.

33. The system of claim 12, wherein the second group of network sources includes at least one public network source, and wherein the data extractor and correlator is further configured to store information derived algorithmically from structured and/or unstructured data from the at least one public network source of the second group of network sources.

34. The system of claim 33, wherein the results correspond to at least one item within the at least one public network source.