Moderation of user-generated content
A system and method for updating and correcting facts that receives proposed values for facts from users and determines a correctness score which is used to automatically accept or reject the proposed values.
1. A computer-implemented method for moderating user-generated content, the method comprising:
obtaining, by a computing system comprising one or more processors, proposed user-generated content for inclusion in an online database that is accessed from a web-based application;
obtaining, by the computing system, metadata associated with a user providing the proposed user-generated content;
processing, by the computing system, the proposed user-generated content with a reliability engine to generate a value unreliability score, wherein the reliability engine comprises one or more machine-learned models, wherein the value unreliability score is indicative of a probability that unreliable information has been provided;
determining, by the computing system, the value unreliability score exceeds one or more thresholds; and
in response to determining, by the computing system, the value unreliability score meets the one or more thresholds, storing, by the computing system, the proposed user-generated content in the online database.
2. The method of claim 1 , wherein edit sessions of unreliable users are used as a training set for training at least one machine-learned model of the reliability engine.
3. The method of claim 2 , wherein the edit sessions comprise a plurality of different characteristics that are used to differentiate unreliable users from reliable users, the plurality of different characteristics of the edit sessions comprising indications associated with at least one or more of a frequency of the edit sessions, an average time between the edit sessions, a number of edits made per edit session, or a time of day of each edit session.
4. The method of claim 1 , wherein the proposed user-generated content is associated with an online map.
5. The method of claim 1 , wherein the proposed user-generated content comprises information about a map feature.
6. The method of claim 1 , wherein the proposed user-generated content is stored based on the value unreliability score exceeding the one or more thresholds and based on an output of an overclustering engine, wherein an automoderation engine comprises the reliability engine and the overclustering engine.
7. The method of claim 1 , wherein the online database is associated with an online map hosting system, wherein the proposed user-generated content is provided for display via an interface of the online map hosting system after storage.
8. The method of claim 1 , wherein the proposed user-generated content for inclusion in the online database comprises a new feature to be added to the online database.
9. The method of claim 1 , wherein the proposed user-generated content for inclusion in the online database comprises an updated attribute associated with an existing feature already stored in the online database.
10. The method of claim 1 , wherein the online database comprises content from a plurality of sources.
11. The method of claim 1 , wherein the one or more machine-learned models are trained with a training dataset generated based on data from one or more manual moderation sessions.
12. The method of claim 11 , wherein the training dataset comprises a set of records manually identified as overclustered.
13. A computing system for moderating user-generated content, the system comprising:
one or more processors; and
one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
obtaining proposed user-generated content associated with a feature for inclusion in an online database of features;
obtaining metadata associated with a user providing the proposed user-generated content, wherein the metadata comprises information associated with previous user interactions with the online database of features;
processing the proposed user-generated content with a reliability engine to generate a content-based unreliability score, wherein the reliability engine comprises one or more machine-learned models, wherein the content-based unreliability score is indicative of a probability that unreliable information has been provided; and
evaluating the content-based unreliability score based on one or more thresholds to determine an action to implement; and
in response to evaluating the content-based unreliability score based on one or more thresholds, rejecting the proposed user-generated value for inclusion in the online database.
14. The system of claim 13 , wherein the information associated with previous user interactions comprises a frequency of editing sessions.
15. The system of claim 13 , wherein the information associated with previous user interactions comprises a number of edits made per session.
16. The system of claim 13 , wherein the metadata comprises accuracy data associated with one or more previous interactions with the online database, wherein the accuracy data comprises recency weighting that more heavily weights more recently proposed values.
17. The system of claim 13 , wherein the metadata associated with the user providing the proposed user-generated value comprises data from edit logs that contain information about user interactions with the online database of features, each log entry of the edit logs comprising one or more of a timestamp, user ID, and type of interaction.
18. A computing system for moderating user-generated content, the system comprising:
one or more processors; and
one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
obtaining proposed user-generated content for inclusion in an online database that is accessed from a web-based application;
obtaining metadata associated with a user providing the proposed user-generated content;
processing the proposed user-generated content with a reliability engine to generate a value unreliability score, wherein the reliability engine comprises one or more machine-learned models, wherein the value unreliability score is indicative of a probability that unreliable information has been provided;
determining the value unreliability score exceeds one or more thresholds; and
in response to determining the value unreliability score meets the one or more thresholds, storing the proposed user-generated content in the online database.
19. The system of claim 18 , wherein the metadata comprises an edit log associated with one or more previous interactions associated with the user.
20. The system of claim 18 , wherein the value unreliability score is determined based on one or more interactions by a second user.