Automating web tasks based on web browsing histories and user actions

• US 9,594,845 B2
• Filed: 09/24/2010
• Issued: 03/14/2017
• Est. Priority Date: 09/24/2010
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bengualid, Mateo Nicolas, Cerruti, Julian Ariel, Lau, Tessa Ann, Manzato, Guillermo, Nichols, Jeffrey William
• Primary Examiner(s): TITCOMB, WILLIAM D
• Application Number:
  US12/890,327
• Publication Number:
  US 20120079395A1
• Time in Patent Office: 2,363 Days
• Field of Search: 715/714, 715/811, 715/762
• US Class Current: 1/1
• CPC Class Codes: G06F 16/957 Browsing optimisation, e.g. caching or content distillation
• CPC Class Codes: G06F 3/0484 for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range
• CPC Class Codes: G06F 40/20 Natural language analysis semantic analysis of natural language G06F40/30

Abstract

Embodiments of the invention relate to the automated processing of web tasks to provide information relevant to user tasks based on past user actions and web histories. An aspect of the invention includes a system for automating the processing of web tasks in response to user requests. The system may include a transport router that receives a user task from a transport such as a short messaging system (SMS), an email system, or a telephone system. The system may further include a web task synthesizer that generates a sequence of actions to accomplish the task, typically in the form of a script for execution by a browser automation server. The script is based on past user actions, logs, and web browsing histories. The system compiles the most relevant information from the visited web pages and returns the results to the user.

Claims

1. A system comprising:

a message transport system to receive a user web task from a user device;

a synthesizer to extract command information from the task, the extracted information including task parameters converted into a command, the command selected from the group consisting of;

syntax and unstructured text in a core command;

the synthesizer to identify a plurality of prior scripts related to the command and rank a relevance of the scripts relative to the task;

the synthesizer to generate a sequence of web actions from a selection of ranked scripts into a combined script to automate and accomplish the task, wherein the sequence of web actions is based on prior user web actions;

and

a browser automation server to execute the synthesized sequence of web actions associated with the combined script, and returning to the user information that is relevant to the user web task, wherein the information results from said execution.

2. The system of claim 1, further comprising the synthesizer to receive an added parameter to one of the scripts.

3. The system of claim 1, further comprising the script execution returning two or more web pages of relevant information and the browser to combine parts of the returned pages into a response.

4. The system of claim 3, further comprising the combined script including a clip command to extract regions of interest from the returned pages.

5. The system of claim 4, further comprising a concatenated output of the extracted regions.

6. The system of claim 1, further comprising the combined script including an auto-clip command to rank document object clusters relative to the user task.

7. A computer program product comprising a computer readable non-transitory storage medium having computer readable program code embodied therewith, the computer readable program code comprising:

computer readable program code configured to receive a user web task from a user device;

computer readable program code to extract command information from the task, the extracted information including task parameters converted into a command, the command selected from the group consisting of;

syntax and unstructured text in a core command;

computer readable program code to identify a plurality of prior scripts related to the command and rank a relevance of the scripts relative to the task;

computer readable program code configured to generate a sequence of web actions from a selection of ranked scripts into a combined script to automate and accomplish the task, wherein the sequence of actions is based on prior user web actions;

computer readable program code configured to execute the synthesized sequence associated with the combined script; and

computer readable program code configured to return to the user information that is relevant to the task, wherein the information results from said execution.

8. The computer program product of claim 7, further comprising computer readable program code configured to allow addition of parameter input to the synthesized sequence.

9. The computer program product of claim 7, further comprising computer readable program code configured to extract said information from a web page downloaded during said execution.

10. The computer program product of claim 9, wherein said information is extracted using an auto-clipping utility.

11. The computer program product of claim 7, wherein the task is specified through a message transport system.

12. The computer program product of claim 11, wherein the message transport system is selected from the group consisting of an email interface, a short message system (SMS) interface, and a portable computer interface.

13. The computer program product of claim 7, wherein the task is specified as a generally short textual message.

14. A method comprising:

receiving a user web task from a user device;

extracting command information from the user task, the extracted information including task parameters converted into a command, the command selected from the group consisting of syntax and unstructured text in a core command;

identifying a plurality of prior scripts related to the command and rank a relevance of the scripts relative to the task;

generating a sequence of web actions from a selection of ranked scripts into a combined script to automate and accomplish the task, wherein the sequence is based on prior web actions;

executing the synthesized sequence associated with the combined script; and

returning to the user, information that is relevant to the task, wherein the information results from said execution.

15. The method of claim 14, wherein the prior actions are derived from a browsing history.

16. The method of claim 14, further comprising adding parameter input to the synthesized sequence.

17. The method of claim 14, further comprising confirming the synthesized sequence of web actions through a user interface.

18. The method of claim 14, wherein said information is extracted from a web page downloaded during said execution.

19. The method of claim 18, wherein said information is extracted based on a clipping instruction.

20. The method of claim 18, wherein said information is extracted using an auto-clipping utility.

21. The method of claim 20, wherein the auto-clipping utility is based on geometric clustering and generates a web page region with a highest relevancy score relative to the task.

22. The method of claim 14, wherein the synthesized sequence is processed by a browser automation server.

23. The method of claim 14, wherein the task is specified through a message transport system.

24. The method of claim 23, wherein the message transport system is selected from the group consisting of an email interface, a short message system (SMS) interface, and a portable computer interface.

25. The method of claim 14, wherein the user web task is specified in a generally short textual format.