PROMPT RANKING

• US 20180063062A1
• Filed: 08/30/2016
• Published: 03/01/2018
• Est. Priority Date: 08/30/2016
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Burdakov, Daniil, Liang, Xiaocong, Duan, Sunyu, Giginiak, Volodymyr, Silva, Pedro Veras Bezerra da
• Granted Patent: US 10,419,381 B2
• Time in Patent Office: Days
• CPC Class Codes: H04L 51/224 providing notification on incoming messages, e.g. pushed notifications of received messages
• CPC Class Codes: H04L 51/226 Delivery according to priorities
• CPC Class Codes: H04L 51/52 for supporting social networking services
• CPC Class Codes: H04L 67/306 User profiles

Abstract

In one embodiment, a method includes accessing a number of candidate notifications stored on the social-networking system. Each of the candidate notifications includes a prompt to perform an action on the social-networking system. The method also includes generating a user feature vector quantifying features of the user; and assessing, using a machine-learning model, one or more feature vectors in order to calculate an interaction score for each of the candidate notifications. The one or more feature vectors includes the user feature vector. The method also includes ranking each of the candidate notifications based at least in part on the respective calculated interaction score; and providing, based at least in part on the ranking, one or more of the candidate notifications to a client device of the user. Each of the provided candidate notifications satisfies a pre-determined threshold value.

Claims

1. A method comprising:

by a computing device and for a user of a social-networking system, accessing a plurality of candidate notifications stored on the social-networking system, wherein each of the candidate notifications comprises a prompt to perform an action on the social-networking system;

by the computing device, generating a user feature vector quantifying features of the user;

by the computing device, assessing, using a machine-learning model, one or more feature vectors in order to calculate an interaction score for each of the candidate notifications, wherein the one or more feature vectors comprises the user feature vector;

by the computing device, ranking each of the candidate notifications based at least in part on the respective calculated interaction score; and

by the computing device, providing, based at least in part on the ranking, one or more of the candidate notifications to a client device of the user, wherein each of the provided candidate notifications satisfies a pre-determined threshold value.

2. The method of claim 1, wherein the one or more feature vectors further comprises, for each of the candidate notifications, a content feature vector quantifying features of the candidate notification.

3. The method of claim 2, wherein one or more of the features of the content feature vector is time dependent.

4. The method of claim 1, wherein the assessing comprises:

for each of the for the candidate notifications, determining a posting probability corresponding to a probability that the user will perform the action associated with the respective candidate notification; and

wherein the interaction score is calculated based on the posting probability.

5. The method of claim 4, wherein the assessing further comprises for each of the for the candidate notifications, determining a dismissal probability corresponding to a probability that the user dismisses the respective candidate notification, and wherein the score calculated further based on the dismissal probability.

6. The method of claim 1, wherein the machine-learning model applies one or more weights to one or more of the features of at least one of the one or more feature vectors.

7. The method of claim 6, further comprising receiving data corresponding to a response by the user to the provided candidate notifications.

8. The method of claim 7, further comprising updating a value of one or more of the weights based on the received data.

9. The method of claim 1, wherein one or more of the features of the user feature vector corresponds to activity measured during a pre-determined amount of time.

10. The method of claim 1, wherein the user feature vector comprises data corresponding to demographics of the user.

11. The method of claim 1, wherein the user feature vector comprises data corresponding to a current location of the user.

12. The method of claim 1, wherein the user feature vector quantifies features comprising historical information regarding past interactions by the user with prompts to perform actions on the social-networking system, or one or more interests of the user, past activity of the user on the social-networking system.

13. The method of claim 12, wherein the features comprising historical information regarding past interactions classify the past interactions based on a type of the past interaction, a type of the candidate notification with which the user interacted, a determination as to whether or not the user performed the action, past interactions with other users on the social-networking system, or an amount of time spent interacting with the social-networking system.

14. The method of claim 1, wherein the action comprises posting content related to content stored on the client device, posting content related to a current location, posting content related to a cultural event, or posting content related to an user occasion.

15. One or more computer-readable non-transitory storage media embodying software configured when executed to:

for a user of a social-networking system, access a plurality of candidate notifications stored on the social-networking system, wherein each of the candidate notifications comprises a prompt to perform an action on the social-networking system;

generate a user feature vector quantifying features of the user;

assess, using a machine-learning model, one or more feature vectors in order to calculate an interaction score for each of the candidate notifications, wherein the one or more feature vectors comprises the user feature vector;

rank each of the candidate notifications based at least in part on the respective calculated interaction score; and

provide, based at least in part on the ranking, one or more of the candidate notifications to a client device of the user, wherein each of the provided candidate notifications satisfies a pre-determined threshold value.

16. The media of claim 15, wherein the one or more feature vectors further comprises, for each of the candidate notifications, a content feature vector quantifying features of the candidate notification.

17. The media of claim 15, wherein the software is further configured to:

for each of the for the candidate notifications, determine a posting probability corresponding to a probability that the user will perform the action associated with the respective candidate notification; and

wherein the interaction score is calculated based on the posting probability.

18. A device comprising:

one or more processors; and

one or more computer-readable non-transitory storage media coupled to the processors and embodying software configured when executed to;

for a user of a social-networking system, access a plurality of candidate notifications stored on the social-networking system, wherein each of the candidate notifications comprises a prompt to perform an action on the social-networking system;

generate a user feature vector quantifying features of the user;

assess, using a machine-learning model, one or more feature vectors in order to calculate an interaction score for each of the candidate notifications, wherein the one or more feature vectors comprises the user feature vector;

rank each of the candidate notifications based at least in part on the respective calculated interaction scores; and

provide, based at least in part on the ranking, one or more of the candidate notifications to a client device of the user, wherein each of the provided candidate notifications satisfies a pre-determined threshold value.

19. The device of claim 18, wherein the one or more feature vectors further comprises, for each of the candidate notifications, a content feature vector quantifying features of the candidate notification.

20. The device of claim 18, wherein the software is further configured to:

for each of the for the candidate notifications, determine a posting probability corresponding to a probability that the user will perform the action associated with the respective candidate notification; and

wherein the interaction score is calculated based on the posting probability.