NEWS FEED RANKING MODEL BASED ON SOCIAL INFORMATION OF VIEWER

• US 20130031489A1
• Filed: 07/29/2011
• Published: 01/31/2013
• Est. Priority Date: 07/29/2011
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Kao, Wayne, Maykov, Alexey, Gubin, Max, Vickrey, David
• Granted Patent: US 9,582,786 B2
• Time in Patent Office: Days
• US Class Current: 715/753
• CPC Class Codes: G06F 11/34 Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Machine learning models are used for ranking news feed stories presented to users of a social networking system. The social networking system divides its users into different sets, for example, based on demographic characteristics of the users and generates one model for each set of users. The models are periodically retrained. The news feed ranking model may rank news feeds for a user based on information describing other users connected to the user in the social networking system. Information describing other users connected to the user includes interactions of the other users with objects associated with news feed stories. These interactions include commenting on a news feed story, liking a news feed story, or retrieving information, for example, images, videos associated with a news feed story.

Claims

1. A computer-implemented method comprising:

receiving information describing past interactions of users of a social networking system with news feed stories, the information represented as a plurality of tuples, each tuple comprising information identifying a news feed story presented to a viewer, information identifying the viewer, and an interaction of the viewer with the news feed story;

for each tuple, identifying features describing interactions of other users connected to the viewer in the social networking system, the interactions performed with objects associated with the news feed story;

training a news feed ranking model using the identified features, the news feed ranking model configured to rank candidate news feed stories selected for presentation to users of the social networking system; and

storing the news feed ranking model.

2. The computer-implemented method of claim 1, wherein the identified features further describe interactions of the other users with the news feed story.

3. The computer-implemented method of claim 1, wherein the identified features describe interactions of the other users with news feed stories similar to the news feed story.

4. The computer-implemented method of claim 1, wherein the identified features comprise an aggregate measure of scores describing interactions of the other users with objects associated with the news feed story.

5. The computer-implemented method of claim 4, wherein the aggregate measure weighs interactions by other users connected to the viewer based on their affinity with the viewer.

6. The computer-implemented method of claim 1, wherein the identified features further describe interactions of the viewer with objects associated with the news feed story.

7. The computer-implemented method of claim 6, wherein the identified features describing interactions of other users with the news feed story are weighted higher than the interactions of the viewer if the number of interactions of the viewer is below a threshold value.

8. The computer-implemented method of claim 1, wherein the identified features comprise demographic information of other users connected to the viewer in the social networking system.

9. The computer-implemented method of claim 1, wherein the plurality of tuples comprise tuples with positive interactions of viewers with news feed stories and tuples with negative interactions of viewers with news feed stories, the positive interactions indicative of viewer interest in the news feed story and the negative interactions indicative of lack of viewer interest in the news feed story.

10. The computer-implemented method of claim 9, wherein a positive interaction comprises one or more of, retrieving additional information describing the news feed story, recommending the news feed story, liking the news feed story, commenting on the story, or selecting a link from the news feed story.

11. The computer-implemented method of claim 9, wherein a negative interaction comprises one or more of, deleting the story from the user'"'"'s profile page, hiding the story, or ignoring the story.

19. The computer-implemented method of claim 1, wherein the identified features comprise demographic information of other users connected to the viewer in the social networking system.

20. The computer-implemented method of claim 1, wherein interactions of other users connected to the viewer comprise retrieving additional information describing the news feed story, recommending the news feed story, liking the news feed story, commenting on the story, or selecting a link from the news feed story, deleting the story from the user'"'"'s profile page, hiding the story, or ignoring the story.

12. A computer-implemented method comprising:

selecting a set of candidate news feed stories in a social networking system for presentation to a viewer;

for each candidate story, identifying features describing interactions of other users connected to the viewer in the social networking system with objects associated with the candidate news feed story;

providing the identified features as input to a news feed story ranking model trained to rank news feed stories for a user, the training based on interactions of other users connected to the user with objects associated with the news feed stories;

ranking the candidate news feed stories using the news feed ranking model; and

sending one or more candidate news feed stories selected for presentation to the viewer based on their ranks

13. The computer-implemented method of claim 12, wherein the identified features further describe interactions of the other users with the news feed story.

14. The computer-implemented method of claim 12, wherein the identified features further describe interactions of the other users with news feed stories similar to the news feed story.

15. The computer-implemented method of claim 12, wherein the identified features comprise an aggregate measure of scores describing interactions of the other users with objects associated with the news feed story.

16. The computer-implemented method of claim 15, wherein the aggregate measure weighs interactions by other users connected to the viewer based on their affinity with the viewer.

17. The computer-implemented method of claim 12, wherein the identified features further describe interactions of the viewer with objects associated with the news feed story.

18. The computer-implemented method of claim 17, wherein the identified features describing interactions of other users with the news feed story are weighted higher than the interactions of the viewer if the number of interactions of the viewer is below a threshold value.