Modeling actions for entity-centric search

• US 9,767,201 B2
• Filed: 12/06/2011
• Issued: 09/19/2017
• Est. Priority Date: 12/06/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Pantel, Patrick, Gamon, Michael, Kannan, Anitha, Fuxman, Ariel, Lin, Thomas
• Primary Examiner(s): TRUONG, CAM Y T
• Application Number:
  US13/311,581
• Publication Number:
  US 20130144854A1
• Time in Patent Office: 2,114 Days
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering

Abstract

In one embodiment, a web service engine server 104 may predict a successive action by a user based on an entity reference 302. The web service engine server 104 identifies an entity reference 302 in a data transmission caused by a user. The web service engine server 104 determines from the data transmission a user intention towards the entity reference 302 using an intention model based on a transmission log. The web service engine server 104 predicts a related successive web action option 522 for the entity reference 302 based on the user intention.

Claims

1. A machine-implemented method, comprising:

observing, by a computer system, directly in a transmission log, a context word distribution and a click through host distribution, in a network;

identifying, by the computer system, an entity reference in a data transmission transmitted by a user via the network;

determining, by the computer system, from the data transmission a user intention towards the entity reference using an intention model having a context multinomial set of parameters describing the context word distribution, an action multinomial parameter set describing an action prediction distribution related to the context multinomial set of parameters and a host multinomial parameter set describing the click through the host distribution from the transmission log;

predicting, by the computer system, a probability of at least one related successive web action option for the entity reference taken after the data transmission in furtherance of the user intention towards the entity reference according to the intention model for determining presentation to the user;

selecting, by the computer system, a related successive web action option set based on the probability;

generating automatically, by the computer system, an action phrasing set for the related successive web action option set;

sending the related successive web action option set to a user device for presentation to the user on a graphical user interface of the user device; and

adjusting, by the computer system, the intention model based on the action phrasing set and a user selection of a related successive web action option from the related successive web action option set presented to user on the user device.

2. The method of claim 1, further comprising:

identifying the entity reference in at least one of a search query, an e-mail, and a social network post.

3. The method of claim 1, further comprising:

compiling the transmission log of user transmissions for multiple users; and

developing the intention model using the transmission log.

4. The method of claim 1, further comprising:

ranking the related successive web action option in a related successive web action option set.

5. The method of claim 4, further comprising:

incorporating a monetization factor into ranking the related successive web action option set.

6. The method of claim 1, further comprising:

associating at least one of a website link and an application programming interface with the related successive web action option.

7. The method of claim 1, further comprising:

identifying at least one of a website link and an application programming interface having the related successive web action option in a metadata tag.

8. The method of claim 1, further comprising:

determining the entity type for the entity reference.

9. The method of claim 1, further comprising:

factoring an entity type into the intention model.

10. The method of claim 1, further comprising:

tailoring the intention model to the user.

11. A tangible machine-readable medium having a set of instructions detailing a method stored thereon that when executed by one or more processors cause the one or more processors to perform the method, the method comprising:

observing, by the processor, directly in a transmission log, a context word distribution and a click through host distribution, in a network;

developing, by the processor, an intention model to model a user intention, the intention model having a context multinomial set of parameters describing a context word distribution, an action multinomial parameter set describing an action prediction distribution related to the context multinomial set of parameters and a host multinomial parameter set describing the click through the host distribution from the transmission log;

identifying, by the processor, an entity reference in a data transmission transmitted by a user via the network;

determining, by the processor, from the data transmission a user intention towards the entity reference using the intention model;

predicting, by the processor, a probability of at least one related successive web action option for the entity reference taken after the data transmission in furtherance of a user intention towards the entity reference according to the intention model for determining presentation to the user;

selecting, by the processor, a related successive web action option set based on the probability;

generating automatically, by the processor, an action phrasing set for the related successive web action option set;

sending the related successive web action option set to a user device for presentation to the user; and

adjusting, by the processor, the intention model based on the action phrasing set and a user selection of a related successive web action option from the related successive web action option set presented to the user on the user device.

12. The tangible machine-readable medium of claim 11, wherein the method further comprises:

factoring an entity type into the intention model.

13. The tangible machine-readable medium of claim 11, wherein the method further comprises:

incorporating a monetization factor into ranking the related successive web action option set.

14. The tangible machine-readable medium of claim 11, wherein the method further comprises:

ranking the related successive web action option in a related successive web action option set.

15. The tangible machine-readable medium of claim 11, wherein the method further comprises:

associating at least one of a website link and an application programming interface with the related successive web action option.

16. The tangible machine-readable medium of claim 11, wherein the method further comprises:

tailoring the intention model to the user.

17. A search engine server, comprising:

a communication interface configured to receive a search query from a user via a network;

a processor configured to;

observe, directly in a transmission log, a context word distribution and a click through host distribution, in the network;

develop an intention model to model a user intention, the intention model having a context multinomial set of parameters describing a context word distribution, an action multinomial parameter set describing an action prediction distribution related to the context multinomial set of parameters and a host multinomial parameter set describing the click through the host distribution from the transmission log;

identify an entity reference in the search query;

determine a user intention towards the entity reference from the search query using the intention model;

predict a probability of at least one related successive web action option set for the entity reference taken after the search query in furtherance of the user intention towards the entity reference according to the intention model for determining presentation to the user;

select a related successive web action option set based on the probability;

generate automatically an action phrasing set for the related successive web action option set;

send the related successive web action option set to a user device for presentation to the user; and

adjust the intention model based on the action phrasing set and a user selection of a related successive web action option from the related successive web action option set presented to the user on the user device; and

a memory configured to store the intention model.

18. The search engine server of claim 17, wherein the processor is configured to rank the related successive web action option set.

19. The search engine server of claim 17, wherein the processor is configured to label an action cluster of the intention model for generation of the action phrasing set.

20. The search engine server of claim 17, wherein the processor is configured to associate at least one of a website link and an application programing interface with the related successive web action option.