Automatic target audience suggestions when sharing in a social network

• US 10,346,411 B1
• Filed: 03/14/2013
• Issued: 07/09/2019
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Deselaers, Thomas, Keysers, Daniel Martin
• Primary Examiner(s): Vital, Pierre M
• Assistant Examiner(s): Ho, Andrew N
• Application Number:
  US13/831,481
• Time in Patent Office: 2,308 Days
• Field of Search: 707749
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A method includes receiving from a user a message for sharing with others, identifying message attributes of the received message, the message attributes comprising message content, identifying a group of recipients with whom the user has previously interacted, computing a relevance score for each recipient in the group of recipients, ranking each recipient in the group of recipients according to the relevance score and determining one or more target recipients for the received message from the group of target recipients ranked according to the relevance score, the determined target recipients having a ranking within a predetermined threshold of highest relevance scores.

Claims

1. A computer-implemented method comprising:

providing a computer learning model, the computer learning model being trained by using a set of previously composed messages by a user, each message from the set of previously composed messages including a group of recipients with whom the user shared each message;

computing a relevance score for each recipient in the group of recipients of each message from the set of previously composed messages using the computer learning model, wherein the relevance score for each recipient in the group of recipients is based at least in part on whether each recipient has previously further interacted with the user based on previous correspondence sent to the user, the previous correspondence including message content related to message attributes of the set of previously composed messages by the user, the message attributes including an attribute indicating a location corresponding to where the user previously composed a particular message from the set of previously composed messages, and the relevance score of each recipient is further based on the attribute indicating the location;

ranking each recipient in the group of recipients according to the relevance score;

receiving, from the user, a message for sharing with others;

filtering the group of recipients to determine a subset of the group of recipients having a ranking within a predetermined threshold of highest relevance scores, the subset including one or more target recipients for the received message, wherein the one or more target recipients are determined based at least in part on an intersection of recipients from the group of recipients of each message from the set of previously composed messages; and

providing, for transmission, a target recipient recommendation to the user in response to the received message, the target recipient recommendation comprising a list identifying the one or more target recipients,

wherein computing the relevance score for each recipient in the group of recipients of each message from the set of previously composed messages further comprises;

determining respective weights for the message attributes related to an indication of importance, wherein a first message attribute corresponding to a particular location and a second message attribute corresponding to content are weighted higher than a third message attribute corresponding to timestamp information and a fourth message attribute corresponding to an attachment; and

generating the relevance score based on a first respective message attribute based at least in part on the determined respective weights.

2. The computer-implemented method of claim 1, further comprising:

identifying the message attributes from message content of the set of previously composed messages;

identifying the group of recipients with whom the user has previously interacted; and

processing the message attributes to identify a signal for identifying the group of recipients with whom the user has previously interacted.

3. The computer-implemented method of claim 1, wherein the relevance score corresponds to how often each recipient in the group of recipients has interacted with the user.

4. The computer-implemented method of claim 1, further comprising:

generating, in response to the filtering, the target recipient recommendation.

5. The computer-implemented method of claim 4, further comprising:

receiving from the user an indication to generate a new target recipient recommendation comprising at least one of, more target recipients or fewer target recipients than indicated in a list of indicated one or more target recipients.

6. The computer-implemented method of claim 5, further comprising:

adjusting the list of indicated one or more target recipients for the received message based on the indication received from the user; and

generating the new target recipient recommendation comprising the adjusted list of indicated one or more target recipients from the received message.

7. The computer-implemented method of claim 1, wherein the message is a social network post.

8. The computer-implemented method of claim 1, wherein the message is an email message.

21. The computer-implemented method of claim 1, wherein identifying the message attributes comprises:

determining a prioritized listing of the message attributes, wherein the message attributes are prioritized based on a weight assigned to each of the message attributes, and wherein the relevance score is computed using the prioritized listing of the message attributes.

22. The computer-implemented method of claim 21, further comprising:

selecting a predetermined number of message attributes from the prioritized listing of the message attributes, the predetermined number of message attributes being a subset of the prioritized listing of the message attributes, wherein the relevance score is computed using the selected predetermined number of message attributes.

23. The computer-implemented method of claim 2, wherein identifying the message attributes comprises:

identifying one or more keywords from the message content of the received message;

associating the one or more keywords to a message attribute for the received message; and

comparing the message attribute of the received message to a message attribute of at least one message previously shared with the user,

wherein the relevance score corresponds to an amount of correlation between keywords of the at least one message previously shared with the user and the one or more keywords of the received message.

24. The computer-implemented method of claim 1, wherein the relevance score indicates that the user has corresponded with a particular recipient when the user is at work based on the attribute indicating the location.

9. A non-transitory machine-readable medium comprising instructions stored therein, which when executed by one or more processors, cause the one or more processors to perform operations comprising:

identifying a group of recipients of messages previously composed and shared by a user;

providing a computer learning model, the computer learning model being trained by using a set of previously composed messages by the user, each message from the set of previously composed messages including the group of recipients with whom the user shared each message;

computing a relevance score for each recipient of the identified group of recipients using the trained computer learning model, wherein the relevance score corresponds to how often each recipient has been included in a list of recipients in each of the messages previously composed and shared by the user, the relevance score further indicating how often the user has selected each recipient as a particular recipient in the identified group of recipients of the messages previously composed and shared by the user, and wherein the relevance score for each recipient in the identified group of recipients is based at least in part on whether each recipient has previously further interacted with the user based on previous correspondence sent to the user, the previous correspondence including message content related to message attributes of the set of previously composed messages by the user, the message attributes including an attribute indicating a location corresponding to where the user previously composed a particular message from the set of previously composed messages, and the relevance score of each recipient is further based on the attribute indicating the location;

ranking each recipient of the identified group of recipients according to the computed relevance score;

receiving, from the user, a message for sharing with others at a social network;

filtering the identified group of recipients to determine a subset of the identified group of recipients having a ranking within a predetermined threshold of highest relevance scores, the subset including one or more target recipients for the received message, wherein the one or more target recipients are determined based at least in part on an intersection of recipients from the identified group of recipients of the messages previously composed and shared by the user; and

providing, for transmission, a target recipient recommendation to the user in response to the received message, the target recipient recommendation comprising a list identifying the one or more target recipients,

wherein computing the relevance score for each recipient in the group of recipients of each message from the set of previously composed messages further comprises;

determining respective weights for the message attributes related to an indication of importance, wherein a first message attribute corresponding to a particular location and a second message attribute corresponding to content are weighted higher than a third message attribute corresponding to timestamp information and a fourth message attribute corresponding to an attachment; and

generating the relevance score based on a first respective message attribute based at least in part on the determined respective weights.

10. The non-transitory machine-readable medium of claim 9, the operations further comprising:

generating, in response to the filtering, the target recipient recommendation.

11. The non-transitory machine-readable medium of claim 9, wherein the received message is a social network post.

12. The non-transitory machine-readable medium of claim 9, wherein the message is an email message.

13. The non-transitory machine-readable medium of claim 9, wherein the operations further comprise:

identifying message attributes from message content of the set of previously composed messages; and

identifying the group of recipients with whom the user has previously interacted,

wherein identifying the messages previously composed and shared by the user comprises comparing message attributes of the at least one message previously shared by each recipient of the group of recipients to message attributes of the messages previously composed and shared by the user.

14. The non-transitory machine-readable medium of claim 13, wherein the messages previously composed and shared by the user are shared by the same user who composed the received message.

15. The non-transitory machine-readable medium of claim 9, wherein identifying the messages previously composed and shared by the user comprises selecting a group of most relevant messages previously composed and shared by the user.

16. The non-transitory machine-readable medium of claim 15, the operations further comprising:

identifying recipients of the selected group of most relevant previously shared messages; and

determining one or more target recipients for the received message based on the identified recipients of the selected group of the most relevant previously shared messages.

17. A system comprising:

one or more processors; and

a machine-readable medium comprising instructions stored therein, which when executed by the one or more processors, cause the one or more processors to perform operations comprising;

providing a computer learning model computer learning model being trained by using a set of previously composed messages by a user, each message from the set of previously composed messages including a group of recipients with whom the user shared each message;

computing a relevance score for each recipient in the group of recipients of each message from the set of previously composed messages using the trained computer learning model, wherein the relevance score for each recipient in the group of recipients is based at least in part on whether each recipient has previously further interacted with the user based on previous correspondence sent to the user, the previous correspondence including message content related to message attributes of messages from the set of previously composed messages by the user, the message attributes including an attribute indicating a location corresponding to where the user previously composed a particular message from the set of previously composed messages, and the relevance score of each recipient is further based on the attribute indicating the location;

ranking each recipient in the group of recipients according to the relevance score;

receiving, from the user, a message for sharing with others;

filtering the group of recipients to determine a subset of the group of recipients having a ranking within a predetermined threshold of highest relevance scores, the subset including one or more target recipients for the received message, wherein the one or more target recipients are determined based at least in part on an intersection of recipients from the group of recipients with whom the user shared each message from the set of previously composed messages; and

providing, for transmission, a target recipient recommendation to the user in response to the received message, the target recipient recommendation comprising a list identifying the one or more target recipients,

wherein computing the relevance score for each recipient in the group of recipients of each message from the set of previously composed messages further comprises;

determining respective weights for the message attributes related to an indication of importance, wherein a first message attribute corresponding to a particular location and a second message attribute corresponding to content are weighted higher than a third message attribute corresponding to timestamp information and a fourth message attribute corresponding to an attachment; and

generating the relevance score based on a first respective message attribute based at least in part on the determined respective weights.

18. The system of claim 17, wherein the operations comprise generating, in response to the filtering, the target recipient recommendation.

19. The system of claim 18, wherein the operations comprise:

receiving from the user an indication to generate a new target recipient recommendation comprising at least one of, more target recipients or fewer target recipients;

adjusting a list of indicated one or more target recipients for the received message based on the indication received from the user; and

generating the new target recipient recommendation comprising the adjusted list of indicated one or more target recipients for the received message.

20. The system of claim 17, wherein the operations comprise:

identifying the message attributes from message content of the set of previously composed messages;

identifying the group of recipients with whom the user has previously interacted; and

processing the message attributes to identify a signal for identifying the group of recipients with whom the user has previously interacted.