Generating and applying event data extraction templates

• US 9,652,530 B1
• Filed: 08/27/2014
• Issued: 05/16/2017
• Est. Priority Date: 08/27/2014
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Josifovski, Vanja, Bendersky, Mike, Heymans, Maureen, Saikia, Amitabh, Cartright, Marc-Allen, Pueyo, Luis Garcia, Lou, Jinan, Yang, Jie, Yang, MyLinh, Tan, Hui
• Primary Examiner(s): Pham, Hung Q
• Application Number:
  US14/470,416
• Time in Patent Office: 993 Days
• Field of Search: None
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/83 Querying
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 10/1095 Meeting or appointment
• CPC Class Codes: H04W 4/029 Location-based management or tracking services

Abstract

Methods and apparatus are described herein for generating and applying event data extraction templates. In various implementations, a set of structural paths may be identified from a corpus of communications. A first structural path of the set of structural paths, associated with a first segment of text, may be classified as transient in response to a determination that a frequency of occurrences of the first segment of text across the corpus satisfies a criterion. Event heuristics may be applied to the communications of the corpus. A determination may be made, based on the applying, that the communications of the corpus are event-related. An event data type may be assigned to the transient structural path based on the applying. An event data extraction template may be generated to extract, from one or more subsequent communications, one or more event-related segments of text associated with the transient structural path.

Claims

1. A computer-implemented method, comprising:

identifying, from a corpus of structured communications, a set of markup language paths, wherein each markup language path specifies a node that underlies a segment of text within a body of a structured communication from the corpus of structured communications;

classifying at least one markup language path of the set of markup language paths, associated with a first segment of text, as a transient markup language path in response to a determination that a frequency of occurrences of the first segment of text across the corpus of structured communications fails to satisfy a threshold;

applying one or more event heuristics to the structured communications of the corpus;

determining, based on the applying, an event data type for the transient markup language path;

assigning the event data type to the transient markup language path based on the determining;

generating an event data extraction template to extract, from one or more subsequent structured communications, one or more event-related segments of text associated with the transient markup language path associated with the assigned event data type;

extracting at least one event-related segment of text from a subsequent structured communication of the one or more subsequent structured communications by applying the event data extraction template to the subsequent structured communication, wherein the subsequent structured communication is addressed to a user; and

providing the extracted at least one event-related segment of text to the user using one or more computing devices operated by the user.

2. The computer-implemented method of claim 1, wherein the applying comprises extracting, from one or more structured communications of the corpus, one or more dates or times pursuant to one or more rules.

3. The computer-implemented method of claim 2, wherein the applying further comprises traversing the one or more extracted dates or times in document object model traversal order to determine that the event data type of the transient markup language path is an event date, event start time, or event end time.

4. The computer-implemented method of claim 3, wherein determining that the event data type of the first segment of text is the event date, event start time, or event end time comprises comparing, to a first date or time extracted from a structured communication of the corpus, a second date or time extracted from elsewhere in the one or more structured communications.

5. The computer-implemented method of claim 1, wherein the determining comprises determining that the event data type of the transient markup language path is an event address based on comparison of the first segment of text to one or more text patterns.

6. The computer-implemented method of claim 1, wherein the determining comprises determining that the event data type of the transient markup language path is an event establishment based on comparison of the first segment of text to one or more location entities in a database.

7. The computer-implemented method of claim 1, wherein the assigning further comprises assigning, to the transient markup language path, a strength of association between the event data type and the transient markup language path.

8. The computer-implemented method of claim 7, further comprising altering the strength of association between the event data type and the transient markup language path based on one or more signals.

9. The computer-implemented method of claim 8, wherein the one or more signals include user feedback concerning data extracted from the subsequent structured communication using the event data extraction template.

10. The computer-implemented method of claim 9, wherein the user feedback comprises user acceptance or rejection of a calendar entry created based on the extracted data.

11. The computer-implemented method of claim 1, wherein the determining comprises determining that the event data type of the transient markup language path is an event title based on one or more signals.

12. The computer-implemented method of claim 11, wherein the one or more signals include a determination that the transient markup language path is at a same level as another transient markup language path underlying another segment of text that is indicative of an event date, event time, or an event location.

13. The computer-implemented method of claim 1, wherein generating the event data extraction template comprises generating the event data extraction template to ignore, in one or more structured communications, one or more segments of text associated with the transient markup language path in response to a determination, based on one or more signals related to structured communications of the corpus, that another semantic data type of the transient markup language path is confidential.

14. The computer-implemented method of claim 1, further comprising discarding communications that are not structured.

15. At least one non-transitory computer-readable medium comprising instructions that, in response to execution of the instructions by a computing device, cause the computing device to perform the following operations:

grouping a corpus of communications into a plurality of clusters based on one or more similarities between communications of the corpus, the plurality of clusters including at least one cluster of structured communications;

identifying, from the at least one cluster of structured communications, a set of markup language paths, wherein each markup language path specifies a node that underlies a segment of text within a body of a structured communication from the at least one cluster of structured communications;

classifying at least one markup language structural path of the set of markup language paths, associated with a first segment of text, as a transient markup language path in response to a determination that a count of occurrences of the first segment of text across the at least one cluster of structured communications fails to satisfy a threshold;

applying one or more event heuristics to the structured communications of the at least one cluster of structured communications;

determining, based on the applying, an event data type for the transient markup language path;

assigning the event data type to the transient markup language path based on the determining;

generating a data extraction template to extract, from one or more subsequent structured communications, one or more segments of text associated with the transient markup language path associated with the assigned event data toe;

extracting at least one event-related segment of text from a subsequent structured communication of the one or more subsequent structured communications by applying the event data extraction template to the subsequent structured communication, wherein the subsequent structured communication is addressed to a user; and

providing the extracted at least one event-related segment of text to the user using one or more computing devices operated by the user.

16. The non-transitory computer-readable medium of claim 15, further comprising instructions that, in response to execution of the instructions by the computing device, cause the computing device to perform the following operations:

identifying, from one or more structured communications of the at least one cluster pursuant to one or more rules, one or more markup language paths associated with dates or times; and

traversing the identified one or more markup language paths in document object model traversal order to determine that the event data type of the transient markup language path is an event date, event start time, or event end time.