Updating a profile of an online system user to include an affinity for an item based on an image of the item included in content received from the user and/or content with which the user interacted

Abstract

An online system receives a content item including an image from a content-providing user and/or receives an interaction with the content item from a viewing user. The online system accesses a machine-learning model that is trained based on a set of images of items associated with an entity and attributes of each image. The online system applies the model to predict a probability that the content item includes an image of an item associated with the entity based on attributes of the image included in the content item. Based on the predicted probability, the online system updates a profile of the user (i.e., the content-providing user and/or the viewing user) to include an affinity for the item. Upon determining an opportunity to present content to the user, the online system selects content for presentation to the user based on the profile and sends the content for presentation to the user.

What is claimed is: 
     
         1 . A method comprising:
 receiving a content item comprising an image from a content-providing user of an online system, wherein the content item is to be presented to one or more viewing users of the online system;   accessing a trained machine-learning model, the trained machine-learning model trained based at least in part on a set of images of one or more items associated with an entity having a presence on the online system and a set of attributes of each of the set of images;   applying the trained machine-learning model to predict a probability that the content item comprises an image of an item associated with the entity, the probability predicted based at least in part on one or more attributes of the image comprising the content item;   accessing an additional trained machine-learning model, the additional trained machine-learning model trained based at least in part on an additional set of images of the content-providing user maintained in the online system and a set of attributes of each of the additional set of images; and   applying the additional trained machine-learning model to predict an additional probability that the content item comprises an image of the content-providing user, the probability predicted based at least in part on the one or more attributes of the image comprising the content item;   updating a profile of the content-providing user to include an affinity of the content-providing user for the item associated with the entity based at least in part on the predicted probability, wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity is further based at least in part on the additional probability;   determining an opportunity to present content to the content-providing user;   selecting one or more content items for presentation to the content-providing user based at least in part on the profile of the content-providing user; and   sending the selected one or more content items for presentation to the content-providing user.   
     
     
         2 . The method of  claim 1 , wherein selecting the one or more content items for presentation to the content-providing user comprises:
 identifying a set of candidate content items eligible for presentation to the content-providing user,   predicting an affinity of the content-providing user for each of the set of candidate content items based at least in part on the profile of the content-providing user, and selecting the one or more content items from the set of candidate content items for presentation to the content-providing user based at least in part on the predicted affinity of the content-providing user for each of the set of candidate content items   
     
     
         3 . The method of  claim 1 , wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity is further based at least in part on the additional probability. 
     
     
         4 . The method of  claim 3 , wherein the affinity of the content-providing user for the item associated with the entity describes one or more of: a taste of the content-providing user in clothing worn by the content-providing user and a taste of the content-providing user in clothing worn by individuals other than the content-providing user. 
     
     
         5 . The method of  claim 1 , wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity comprises:
 determining whether the predicted probability is at least a threshold probability;   responsive to determining that the predicted probability is at least the threshold probability, generating an embedding describing the affinity of the content-providing user for the item associated with the entity; and   storing the embedding in association with the profile of the content-providing user.   
     
     
         6 . The method of  claim 1 , wherein updating a profile of the content-providing user comprises:
 comparing the predicted probability to one or more threshold probabilities, wherein each of the one or more threshold probabilities is associated with a topic and the topic is associated with one or more items associated with one or more entities having a presence on the online system;   determining a topic associated with the content item based at least in part on the comparing; and   updating the profile of the content-providing user to include an affinity of the content-providing user for the topic.   
     
     
         7 . The method of  claim 1 , further comprising:
 receiving an interaction with the content item from a viewing user of the online system;   updating an additional profile of the viewing user to include an affinity of the viewing user for the item associated with the entity based at least in part on the predicted probability;   determining an opportunity to present content to the viewing user;   selecting one or more additional content items for presentation to the viewing user based at least in part on the additional profile of the viewing user; and   sending the selected one or more additional content items for presentation to the viewing user.   
     
     
         8 . The method of  claim 7 , wherein updating the additional profile of the viewing user to include the affinity of the viewing user for the item associated with the entity comprises:
 comparing the predicted probability to one or more threshold probabilities, wherein each of the one or more threshold probabilities is associated with a topic and the topic is associated with one or more items associated with one or more entities having a presence on the online system;   determining a topic associated with the content item based at least in part on the comparing; and   updating the additional profile of the viewing user to include an affinity of the viewing user for the topic.   
     
     
         9 . The method of  claim 7 , wherein selecting the one or more additional content items for presentation to the viewing user comprises:
 identifying a set of candidate content items eligible for presentation to the viewing user;   predicting an affinity of the viewing user for each of the set of candidate content items based at least in part on the additional profile of the viewing user; and   selecting the one or more additional content items from the set of candidate content items for presentation to the viewing user based at least in part on the predicted affinity of the viewing user for each of the set of candidate content items.   
     
     
         10 . A non-transitory computer-readable storage medium storing instructions executable by one or more processors for performing steps including:
 receiving a content item comprising an image from a content-providing user of an online system, wherein the content item is to be presented to one or more viewing users of the online system;   accessing a trained machine-learning model, the trained machine-learning model trained based at least in part on a set of images of one or more items associated with an entity having a presence on the online system and a set of attributes of each of the set of images;   applying the trained machine-learning model to predict a probability that the content item comprises an image of an item associated with the entity, the probability predicted based at least in part on one or more attributes of the image comprising the content item;   accessing an additional trained machine-learning model, the additional trained machine-learning model trained based at least in part on an additional set of images of the content-providing user maintained in the online system and a set of attributes of each of the additional set of images; and   applying the additional trained machine-learning model to predict an additional probability that the content item comprises an image of the content-providing user, the probability predicted based at least in part on the one or more attributes of the image comprising the content item;   updating a profile of the content-providing user to include an affinity of the content-providing user for the item associated with the entity based at least in part on the predicted probability, wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity is further based at least in part on the additional probability;   determining an opportunity to present content to the content-providing user;   selecting one or more content items for presentation to the content-providing user based at least in part on the profile of the content-providing user; and   sending the selected one or more content items for presentation to the content-providing user.   
     
     
         11 . The non-transitory computer-readable storage medium of  claim 10 , wherein selecting the one or more content items for presentation to the content-providing user comprises:
 identifying a set of candidate content items eligible for presentation to the content-providing user,   predicting an affinity of the content-providing user for each of the set of candidate content items based at least in part on the profile of the content-providing user, and   selecting the one or more content items from the set of candidate content items for presentation to the content-providing user based at least in part on the predicted affinity of the content-providing user for each of the set of candidate content items   
     
     
         12 . The non-transitory computer-readable storage medium of  claim 10 , wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity is further based at least in part on the additional probability. 
     
     
         13 . The non-transitory computer-readable storage medium of  claim 12 , wherein the affinity of the content-providing user for the item associated with the entity describes one or more of: a taste of the content-providing user in clothing worn by the content-providing user and a taste of the content-providing user in clothing worn by individuals other than the content-providing user. 
     
     
         14 . The non-transitory computer-readable storage medium of  claim 10 , wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity comprises:
 determining whether the predicted probability is at least a threshold probability;   responsive to determining that the predicted probability is at least the threshold probability, generating an embedding describing the affinity of the content-providing user for the item associated with the entity; and   storing the embedding in association with the profile of the content-providing user.   
     
     
         15 . The non-transitory computer-readable storage medium of  claim 10 , wherein updating a profile of the content-providing user comprises:
 comparing the predicted probability to one or more threshold probabilities, wherein each of the one or more threshold probabilities is associated with a topic and the topic is associated with one or more items associated with one or more entities having a presence on the online system;   determining a topic associated with the content item based at least in part on the comparing; and   updating the profile of the content-providing user to include an affinity of the content-providing user for the topic.   
     
     
         16 . The non-transitory computer-readable storage medium of  claim 10 , further storing instructions executable by one or more processors for performing steps including:
 receiving an interaction with the content item from a viewing user of the online system;   updating an additional profile of the viewing user to include an affinity of the viewing user for the item associated with the entity based at least in part on the predicted probability;   determining an opportunity to present content to the viewing user;   selecting one or more additional content items for presentation to the viewing user based at least in part on the additional profile of the viewing user; and   sending the selected one or more additional content items for presentation to the viewing user.   
     
     
         17 . The non-transitory computer-readable storage medium of  claim 16 , wherein updating the additional profile of the viewing user to include the affinity of the viewing user for the item associated with the entity comprises:
 comparing the predicted probability to one or more threshold probabilities, wherein each of the one or more threshold probabilities is associated with a topic and the topic is associated with one or more items associated with one or more entities having a presence on the online system;   determining a topic associated with the content item based at least in part on the comparing; and   updating the additional profile of the viewing user to include an affinity of the viewing user for the topic.   
     
     
         18 . The non-transitory computer-readable storage medium of  claim 16 , wherein selecting the one or more additional content items for presentation to the viewing user comprises:
 identifying a set of candidate content items eligible for presentation to the viewing user;   predicting an affinity of the viewing user for each of the set of candidate content items based at least in part on the additional profile of the viewing user; and   selecting the one or more additional content items from the set of candidate content items for presentation to the viewing user based at least in part on the predicted affinity of the viewing user for each of the set of candidate content items.   
     
     
         19 . A computer system comprising:
 one or more processors; and   a non-transitory computer-readable storage medium storing instructions executable by the one or more processors for performing steps including:
 receiving a content item comprising an image from a content-providing user of an online system, wherein the content item is to be presented to one or more viewing users of the online system; 
 accessing a trained machine-learning model, the trained machine-learning model trained based at least in part on a set of images of one or more items associated with an entity having a presence on the online system and a set of attributes of each of the set of images; 
 applying the trained machine-learning model to predict a probability that the content item comprises an image of an item associated with the entity, the probability predicted based at least in part on one or more attributes of the image comprising the content item; 
 accessing an additional trained machine-learning model, the additional trained machine-learning model trained based at least in part on an additional set of images of the content-providing user maintained in the online system and a set of attributes of each of the additional set of images; and 
 applying the additional trained machine-learning model to predict an additional probability that the content item comprises an image of the content-providing user, the probability predicted based at least in part on the one or more attributes of the image comprising the content item; 
 updating a profile of the content-providing user to include an affinity of the content-providing user for the item associated with the entity based at least in part on the predicted probability, wherein updating the profile of the content-providing user to include the affinity of the content-providing user for the item associated with the entity is further based at least in part on the additional probability; 
 determining an opportunity to present content to the content-providing user; 
 selecting one or more content items for presentation to the content-providing user based at least in part on the profile of the content-providing user; and 
 sending the selected one or more content items for presentation to the content-providing user. 
   
     
     
         20 . The system of  claim 19 , wherein selecting the one or more content items for presentation to the content-providing user comprises:
 identifying a set of candidate content items eligible for presentation to the content-providing user,   predicting an affinity of the content-providing user for each of the set of candidate content items based at least in part on the profile of the content-providing user, and   selecting the one or more content items from the set of candidate content items for presentation to the content-providing user based at least in part on the predicted affinity of the content-providing user for each of the set of candidate content items.