Abstract: Described are systems and methods for determining a diversity function in connection with potential advertisements in determining one or more advertisements to present to a user. This can facilitate presenting diverse advertisements to users. The diversity function can yield a diversity penalty for each potential advertisement to be able to determine a rank order of the potential advertisements. The diversity function can include a repetition score and a multiplier for each potential advertisement.

Abstract: A method includes accessing a first set of content objects of an online social network posted within a predetermined timeframe, and identifying trending topics. Each object may be associated with one or more topics. The method further includes accessing a second set of content objects, and identifying a plurality of co-occurring topics associated with the second set. Each content object of the second set is associated with at least one of the identified trending topics. Then, the method includes accessing a third set of content objects, and analyzing the plurality of trending topics and co-occurring topics to detect key topics. Each content object of the third set of content objects may be associated with at least one of the identified co-occurring topics or trending topics. The key topics may be detected based on determining the co-occurring topics or trending topics that are associated with a threshold number of content objects.

Abstract: A method includes accessing a first set of content objects of an online social network posted within a predetermined timeframe, and identifying trending topics. Each object may be associated with one or more topics. The method further includes accessing a second set of content objects, and identifying a plurality of co-occurring topics associated with the second set. Each content object of the second set is associated with at least one of the identified trending topics. Then, the method includes accessing a third set of content objects, and analyzing the plurality of trending topics and co-occurring topics to detect key topics. Each content object of the third set of content objects may be associated with at least one of the identified co-occurring topics or trending topics. The key topics may be detected based on determining the co-occurring topics or trending topics that are associated with a threshold number of content objects.