ELECTRONIC ASSISTANT

• US 20100179961A1
• Filed: 08/10/2009
• Published: 07/15/2010
• Est. Priority Date: 01/12/2009
• Status: Active Grant

• Current Assignee: SRI International, Inc.
• Original Assignee: SRI International, Inc.
• Inventors: Donneau-Golencer, Thierry, Pound, Leslie, Peintner, Bart, Cheyer, Adam, Yorke-Smith, Neil, Berry, Pauline M, Gervasio, Melinda, Hardt, Stephen, Nitz, Kenneth
• Granted Patent: US 9,443,226 B2
• Time in Patent Office: Days
• US Class Current: 707/769
• CPC Class Codes: G06F 16/80 of semi-structured data, e.g. markup language structured data such as SGML, XML or HTML content-based retrieval of web data G06F16/95
• CPC Class Codes: G06F 40/284 Lexical analysis, e.g. tokenisation or collocates
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]

Abstract

The present invention relates to an electronic assistant. In one embodiment, a system for processing data on a computing device includes at least one extractor for extracting data from an email to produce extracted data, a user model for generating a high-level model of an email environment on the computing device, based on the extracted data, and at least one predictor for generating a prediction related to the email, based on the extracted data and on the high-level model. In another embodiment, a system for scheduling an event includes an extractor for extracting constraints from a request for the event, a user model for receiving and storing scheduling preferences from a user, and a predictor coupled to the extractor and to the user model, for generating a candidate schedule for the event in accordance with the constraints and the scheduling preferences.

Claims

1. A system for processing data on a computing device, the system comprising:

at least one extractor for extracting data from an email to produce extracted data;

a user model coupled to the at least one extractor, for generating a high-level model of an email environment on the computing device, based on the extracted data; and

at least one predictor coupled to the at least one extractor and to the user model, for generating a prediction related to the email, based on the extracted data and on the high-level model.

2. The system of claim 1, wherein the at least one extractor comprises:

a semantic extractor for extracting semantic data from the email.

3. The system of claim 2, wherein the semantic data is at least one of:

at least one relation feature or at least one bag of words feature.

4. The system of claim 1, wherein the at least one extractor comprises:

a metadata extractor for extracting metadata from the email.

5. The system of claim 1, wherein the high-level model identifies one or more patterns of behavior exhibited by a user of the email environment.

6. The system of claim 1, wherein the at least one predictor comprises:

an importance predictor for predicting an importance of the email to a user of the email environment, where the email is an incoming email.

7. The system of claim 1, wherein the at least one predictor comprises:

an attachment predictor for predicting whether a file attachment should be included in the email, where the email is an outgoing email.

8. The system of claim 1, wherein the at least one predictor comprises:

a carbon copy predictor for predicting whether a carbon copy address should be included in the email, where the email is an outgoing email.

9. The system of claim 1, further comprising:

a rules database coupled to the at least one predictor, for storing rules defined by a user of the email environment, where the rules are used to generate the prediction.

10. The system of claim 1, further comprising:

a database coupled to the at least one extractor, for building a set based on the extracted data, where the set comprises at least one of;

at least one contact, at least one task, or at least one event.

11. A method for processing data on a computing device, the method comprising:

extracting data from an email to produce extracted data;

generating a high-level model of an email environment on the computing device, based on the extracted data; and

generating a prediction related to the email, based on the extracted data and on the high-level model.

12. The method of claim 11, wherein the email is an incoming email received by the computing device from an external source.

13. The method of claim 12, wherein the extracted data comprises one or more relational features, the one or more relational features being based on a relation of a user of the email environment to at least one of:

a sender of the email, one or more main recipients of the email, one or more carbon copy recipients of the email, one or more Internet domains contained in the email, or one or more file attachments included in the email.

14. The method of claim 12, wherein the extracted data comprises one or more bag of word features extracted from at least one of:

a subject of the email or a body of the email.

15. The method of claim 12, wherein the extracted data comprises metadata.

16. The method of claim 11, wherein the high-level model contains at least one of:

a total number of emails received from a particular correspondent, a total number of emails sent to a particular correspondent, a group of one or more correspondents to whom a user of the email environment is most responsive, or a frequency with which the user of the email environment replies to email sent by correspondents who are in a contact list.

17. The method of claim 11, wherein the prediction predicts an importance of the email to a user of the email environment.

18. The method of claim 17, further comprising:

receiving one or more user-defined rules that identify one or more criteria to be considered in predicting the importance of the email;

storing the one or more user-defined rules in a database; and

applying the one or more user defined rules to the email in accordance with generating the prediction.

19. The method of claim 12, wherein the extracted data comprises at least one of:

one or more semantic types contained in the email, one or more contacts contained in the email, one or more tasks contained in the email, or one more events contained in the email.

20. The method of claim 19, further comprising:

identifying one or more patterns in the incoming email, in accordance with the one or more semantic types.

21. The method of claim 20, wherein the identifying comprises:

generalizing a search over text of the incoming email; and

extracting data from the incoming email that matches the search.

22. The method of claim 20, wherein the identifying comprises:

generating an array of one or more words appearing in the incoming email;

vectorizing one or more words in the array into a vector comprising a plurality of digits, each digit corresponding to a characteristic of an associated one of the one or more words.

23. The method of claim 19, wherein the prediction predicts at least one contact, task, or event to be added to an application on the computing device.

24. The method of claim 23, wherein generating the prediction comprises:

receiving one or more scheduling preferences from the user;

extracting a request for an event from the email;

extracting one or more constraints on the event from the email; and

identifying at least one candidate schedule based on the one or more scheduling preferences, the request, and the one or more constraints.

25. The method of claim 24, wherein the one or more constraints include at least one of:

one or more constraints on a time of the event, one or more required participants for the event, one or more optional participants for the event, one or more proposed locations for the event, one or more existing schedules for the one or more required participants, one or more existing schedules for the one or more optional participants, a preference model that indicates a relative importance of each type of the one or more constraints, one or more preference models relating to date and time preferences for the one or more required participants, or one or more preference models relating to date and time preferences for the one or more optional participants.

26. The method of claim 24, wherein the identifying comprises:

applying a variable-ordering heuristic to the one or more scheduling preferences, the request, and the one or more constraints.

27. The method of claim 26, wherein the applying comprises:

tagging at least one of;

at least one of the one or more constraints or at least one variable associated with the request with a semantic tag to produce tagged data; and

applying one or more temporal soft constraints and one or more flexible preference models to the tagged data.

28. The method of claim 11, wherein the prediction comprises a tag for the email.

29. The method of claim 28, wherein the generating comprises:

receiving a query from a user of the computing device;

determining that the email satisfies the query; and

tagging the email with a tag associated with the query.

30. The method of claim 29, wherein the tagging comprises:

storing the tag as metadata for the email.

31. The method of claim 28, wherein the generating comprises:

observing a user of the computing device tagging at least one item of the data on the computing device to produce at least one tagged item;

computing a similarity between the email and the at least one tagged item; and

tagging the email with a tag assigned to the at least one tagged item if the similarity meets or exceeds a predefined threshold.

32. The method of claim 11, wherein the email is an outgoing email about to be sent by the computing device to an external source.

33. The method of claim 32, wherein the prediction predicts whether the email should include a file attachment.

34. The method of claim 32, wherein the prediction predicts whether the email should include a carbon copy address.

35. The method of claim 11, further comprising:

displaying the prediction to a user of the email environment; and

receiving feedback from the user in response to the prediction.

36. The method of claim 35, further comprising:

updating the high-level model in accordance with the feedback.

37. A computer readable storage medium containing an executable program for processing data on a computing device, where the program performs the steps of:

extracting data from an email to produce extracted data;

generating a high-level model of an email environment on the computing device, based on the extracted data; and

generating a prediction related to the email, based on the extracted data and on the high-level model.

38. A system for scheduling an event, comprising:

at least one extractor for extracting one or more constraints from a request for the event;

a user model for receiving and storing one or more scheduling preferences from at least one user; and

at least one predictor coupled to the at least one extractor and to the user model, for generating at least one candidate schedule for the event in accordance with the one or more constraints and the one or more scheduling preferences.

39. A method for scheduling an event by a computing device, the method comprising:

extracting one or more constraints from a request for the event;

storing one or more scheduling preferences received from at least one user; and

generating at least one candidate schedule for the event in accordance with the one or more constraints and the scheduling preferences.

40. The method of claim 39, wherein the one or more constraints include at least one of:

one or more constraints on a time of the event, one or more required participants for the event, one or more optional participants for the event, one or more proposed locations for the event, one or more existing schedules for the one or more required participants, one or more existing schedules for the one or more optional participants, a preference model that indicates a relative importance of each type of the one or more constraints, one or more preference models relating to date and time preferences for the one or more required participants, or one or more preference models relating to date and time preferences for the one or more optional participants.

41. The method of claim 40, wherein the one or more constraints are extracted from at least one email or are supplied directly by the at least one user.

42. The method of claim 39, wherein the generating comprises:

applying a variable-ordering heuristic to the one or more scheduling preferences, the request, and the one or more constraints.

43. The method of claim 42, wherein the applying comprises:

tagging at least one of;

at least one of the one or more constraints or at least one variable associated with the request with a semantic tag to produce tagged data; and

applying one or more temporal soft constraints and one or more flexible preference models to the tagged data.

44. The method of claim 39, further comprising:

receiving a selection of one of the at least one candidate schedule from at least one of the at least one user; and

updating the scheduling preferences associated with the at least one of the at least one user in accordance with the selection.

45. The method of claim 39, wherein at least one of:

the request and the one or more scheduling preferences is received in an email.

46. The method of claim 45, wherein the email is an outgoing email sent by the computing device to an external location.

47. The method of claim 45, wherein the email is an incoming email received by the computing device from an external source.