Scoring stream items with models based on user interests
First Claim
Patent Images
1. A method for generating a stream of content for a user, the method comprising:
- receiving candidate content items from heterogeneous data sources including from other users in at least one social network that have a relationship with the user;
applying, with one or more processors, a weight to each of the candidate content items from the other users in the at least one social network that have the relationship with the user, the weight being modified based on a degree of separation between the user and each of the other users;
computing, with the one or more processors, a user-independent global score for each of the candidate content items based on a global popularity of each candidate content item;
generating, with the one or more processors, a ranking for each candidate content item based on the global scores;
generating, with the one or more processors, a model for the user comprising at least one interest of the user and at least one prior action of the user associated with the heterogeneous data sources;
generating, with the one or more processors, a set of the candidate content items that are associated with the at least one interest of the user based on the ranking;
computing, with the one or more processors, an interestingness score for each candidate content item in the set by summing properties of each candidate content item over single-attribute properties using the model and based upon interestingness of each candidate content item to the user and an extent to which the candidate content item'"'"'s popularity has increased within a geographic area associated with the user;
generating, with the one or more processors, the stream of content for the user from selected content items, wherein the interestingness score for each selected content item exceeds a threshold score and wherein at least one of the selected content items is displayed with an explanation identifying an interest from the model that is associated with the selected content item;
receiving a rejection from the user of the interest identified from the model that is associated with the at least one selected content item;
updating the model by removing the interest from the model; and
modifying, with the one or more processors, the stream of content based on updating the model.
2 Assignments
0 Petitions
Accused Products
Abstract
A system and method for generating a stream of content includes a content stream module that generates a model based on user input and/or prior activities using heterogeneous data sources. The heterogeneous data sources include search, entertainment, social activity and activity on third-party sites. The content stream module retrieves candidate content items that have interests that are similar to the user. The candidate content items are compared to the model and scored based upon interestingness of the content item to the user. The content stream module generates the stream of content from the candidate content items.
99 Citations
44 Claims
-
1. A method for generating a stream of content for a user, the method comprising:
-
receiving candidate content items from heterogeneous data sources including from other users in at least one social network that have a relationship with the user; applying, with one or more processors, a weight to each of the candidate content items from the other users in the at least one social network that have the relationship with the user, the weight being modified based on a degree of separation between the user and each of the other users; computing, with the one or more processors, a user-independent global score for each of the candidate content items based on a global popularity of each candidate content item; generating, with the one or more processors, a ranking for each candidate content item based on the global scores; generating, with the one or more processors, a model for the user comprising at least one interest of the user and at least one prior action of the user associated with the heterogeneous data sources; generating, with the one or more processors, a set of the candidate content items that are associated with the at least one interest of the user based on the ranking; computing, with the one or more processors, an interestingness score for each candidate content item in the set by summing properties of each candidate content item over single-attribute properties using the model and based upon interestingness of each candidate content item to the user and an extent to which the candidate content item'"'"'s popularity has increased within a geographic area associated with the user; generating, with the one or more processors, the stream of content for the user from selected content items, wherein the interestingness score for each selected content item exceeds a threshold score and wherein at least one of the selected content items is displayed with an explanation identifying an interest from the model that is associated with the selected content item; receiving a rejection from the user of the interest identified from the model that is associated with the at least one selected content item; updating the model by removing the interest from the model; and modifying, with the one or more processors, the stream of content based on updating the model. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21)
-
-
22. A system for generating a stream of content for a user comprising:
-
a logging unit for receiving at least one interest from the user and at least one prior action of the user associated with heterogeneous data sources, generating a log and transmitting the log to a data storage; a model generation engine coupled to the data storage, the model generation engine for generating a model for the user based on the log, receiving a rejection from the user of an interest identified from the model that is associated with at least one of selected content items being displayed with an explanation and updating the model by removing the interest from the model; and a scoring engine coupled to the model generation engine, the scoring engine for receiving candidate content items from the heterogeneous data sources including from other users in at least one social network that have a relationship with the user, receiving a weight applied to each of the candidate content items from the other users in the at least one social network that have the relationship with the user, the weight being modified based on a degree of separation between the user and each of the other users, computing a user-independent global score for each of the candidate content items based on a global popularity of each candidate content item, generating a ranking for each candidate content item based on the global scores, generating a set of the candidate content items that are associated with the at least one interest of the user based on the ranking, computing an interestingness score for each candidate content item in the set by summing properties of each candidate content item over single-attribute properties using the model and based upon interestingness of each candidate content item to the user and an extent to which the candidate content item'"'"'s popularity has increased within a geographic area associated with the user, generating the stream of content for the user from the selected content items, wherein the interestingness score for each selected content item exceeds a threshold score and wherein the at least one selected content items is displayed with the explanation identifying the interest from the model that is associated with the selected content item, and modifying the stream of content based on updating the model. - View Dependent Claims (23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43)
-
-
44. A computer program product comprising a non-transitory computer useable medium including a computer readable program, wherein the computer readable program when executed on a computer causes the computer to:
-
receive candidate content items from heterogeneous data sources including from other users in at least one social network that have a relationship with a user; apply a weight to each of the candidate content items from the other users in the at least one social network that have the relationship with the user, the weight being modified based on a degree of separation between the user and each of the other users; compute a user-independent global score for each of the candidate content items based on a global popularity of each candidate content item; generate a ranking for each candidate content item based on the global scores; generate a model for the user comprising at least one interest of the user and at least one prior action of the user associated with the heterogeneous data sources; generate a set of the candidate content items that are associated with the at least one interest of the user based on the ranking; compute an interestingness score for each candidate content item in the set by summing properties of each candidate content item over single-attribute properties using the model and based upon interestingness of each candidate content item to the user and an extent to which the candidate content item'"'"'s popularity has increased within a geographic area associated with the user; generate the stream of content for the user from selected content items, wherein the interestingness score for each selected content item exceeds a threshold score and wherein at least one of the selected content items is displayed with an explanation identifying an interest from the model that is associated with the selected content item; receive a rejection from the user of the interest identified from the model that is associated with the at least one selected content item; update the model by removing the interest from the model; and modify the stream of content based on updating the model.
-
Specification