Trust based moderation
View Patent ↗A network device, system, and method are directed towards detecting trusted reporters and/or abusive users in an online community using reputation event inputs, such as abuse reports. When an abuse report is received for a content item, the combined trust (reputation) of previous reporters on the reported content item and the trust (reputation) of the content author are compared to determine whether to trust the content item. If the content item is un-trusted, the content item may be hidden from public view. In one embodiment, the content item might still be visible to the content author, and/or members in the author's contact list, or the like, while being hidden from another user in the community. In one embodiment, the author may appeal the determined trust, and results of the appeal may be used to modify a trust of at least one reporter.
1. A network device, comprising:
a transceiver that is operative to send and receive data over a communications network; and
a processor that is operative to perform actions on at least some of the data, the actions including:
receiving from one or more abuse reporters one or more abuse reports about a message within a content system;
obtaining reputation information to determine reputation of the one or more abuse reporters;
determining abuse reputation of the message based on the reputation of the one or more abuse reporters;
determining, from a plurality of content types, a content type of the message;
determining, for the message, a content creation reputation of a presenter of the message corresponding to the determined content type, wherein the presenter is assigned a separate content creation reputation for each of the plurality of content types;
comparing the abuse reputation of the message and the content creation reputation of the presenter of the message corresponding to the determined content type; and
hiding the message from at least one user of the content system based on the comparison between the abuse reputation and the content creation reputation.
2. The network device of claim 1 , wherein determining the abuse reputation of the message based on the reputation of the one or more abuse reporters comprises determining a reputation of each of the one or more abuse reporters based on a combination of a given reporter's activity in a community, the given reporter's membership value, and the given reporter's contribution value.
3. The network device of claim 1 , wherein determining the content creation reputation of the presenter of the message further comprises:
determining the content creation reputation of the presenter based on a comparison of an aggregation of the presenter's confirmed abusive reputation to an aggregation of the presenter's positive reputation based on positive input from at least one user of the content system; and
when the presenter is determined to be a new user to the content system, modifying the determined content creation reputation of the presenter to induce an initial positive content creation reputation of the presenter for the message.
4. The network device of claim 1 , wherein the content system includes at least one of a web system or messaging system.
5. The network device of claim 1 , wherein the message includes at least one advertisement.
6. The network device of claim 1 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a confidence level that an abuse report is associated with a single abuse reporter; and
modifying the determined reputation of a respective abuse reporter based on the determined confidence level.
7. The network device of claim 1 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a given reporter's abuse suspicion reputation from a combination of abuse profiles from third parties about the given reporter, community reports about the given reporter having been reported by other users, and reports about abuse confidence reputation for the given reporter; and
employing the abuse suspicion reputation to further determine the given reporter's reputation.
8. A non-transitory machine-readable medium having stored thereon computer-executable instructions, when executed, causing a network device to perform actions, the actions including:
receiving from one or more abuse reporters one or more abuse reports about a message within a content system;
obtaining reputation information to determine reputation of the one or more abuse reporters;
determining abuse reputation of the message based on the reputation of the one or more abuse reporters;
determining, from a plurality of content types, a content type of the message;
determining, for the message, a content creation reputation of a presenter of the message corresponding to the determined content type, wherein the presenter is assigned a separate content creation reputation for each of the plurality of content types;
comparing the abuse reputation of the message and the content creation reputation of the presenter of the message corresponding to the determined content type; and
hiding the message from at least one user of the content system based on the comparison between the abuse reputation and the content creation reputation.
9. The non-transitory machine-readable medium of claim 8 , wherein determining the abuse reputation of the message based on reputation of the one or more abuse reporters comprises determining a reputation of each of the one or more abuse reporters based on a combination of a given reporter's activity in a community, the given reporter's membership value, and the given reporter's contribution value.
10. The non-transitory machine-readable medium of claim 8 , wherein determining the content creation reputation of the presenter of the message further comprises:
determining the content creation reputation of the presenter based on a comparison of an aggregation of the presenter's confirmed abusive reputation to an aggregation of the presenter's positive reputation based on positive input from at least one user of the content system; and
when the presenter is determined to be a new user to the content system, modifying the determined content creation reputation of the presenter to induce an initial positive content creation reputation of the presenter for the message.
11. The non-transitory machine-readable medium of claim 8 , wherein the content system includes at least one of a web system or messaging system.
12. The non-transitory machine-readable medium of claim 8 , wherein the message includes at least one advertisement.
13. The non-transitory machine-readable medium of claim 8 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a confidence level that an abuse report is associated with a single abuse reporter; and
modifying the determined reputation of a respective abuse reporter based on the determined confidence level.
14. The non-transitory machine-readable medium claim 8 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a given reporter's abuse suspicion reputation from a combination of abuse profiles from third parties about the given reporter, community reports about the given reporter having been reported by other users, and reports about abuse confidence reputation for the given reporter; and
employing the abuse suspicion reputation to further determine the given reporter's reputation.
15. A method for selectively displaying a message to a user, the method being implemented in one or more processors configured to execute computer programs, the method comprising:
receiving from one or more abuse reporters one or more abuse reports about a message within a content system;
obtaining reputation information to determine reputation of the one or more abuse reporters;
determining abuse reputation of the message based on the reputation of the one or more abuse reporters;
determining, from a plurality of content types, a content type of the message;
determining, for the message, a content creation reputation of a presenter of the message corresponding to the determined content type, wherein the presenter is assigned a separate content creation reputation for each of the plurality of content types;
comparing the abuse reputation of the message and the content creation reputation of the presenter of the message corresponding to the determined content type; and
hiding the message from at least one user of the content system based on the comparison between the abuse reputation and the content creation reputation.
16. The method of claim 15 , wherein determining the content creation reputation of the presenter of the message further comprises:
determining the content creation reputation of the presenter based on a comparison of an aggregation of the presenter's confirmed abusive reputation to an aggregation of the presenter's positive reputation based on positive input from at least one user of the content system; and
when the presenter is determined to be a new user to the content system, modifying the determined content creation reputation of the presenter to induce an initial positive content creation reputation of the presenter for the message.
17. The method of claim 15 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a confidence level that an abuse report is associated with a single abuse reporter; and
modifying the determined reputation of a respective abuse reporter based on the determined confidence level.
18. The method of claim 15 , wherein determining the reputation of the one or more abuse reporters further comprises:
determining a given reporter's abuse suspicion reputation from a combination of abuse profiles from third parties about the given reporter, community reports about the given reporter having been reported by other users, and reports about abuse confidence reputation for the given reporter; and
employing the abuse suspicion reputation to further determine the given reporter's reputation.
19. The method of claim 15 , wherein hiding the message is further based on additional inputs about the message from at least one machine classification mechanism.
20. The method of claim 15 , wherein determining a given reporter's reputation further comprising combining a confirmed reputation value of the given reporter with a bootstrap reputation to generate the given reporter's reputation.