Arranging information describing items within a page maintained in an online system based on an interaction with a link to the page

Abstract

An online system receives information describing (an) item(s) associated with an entity and a content item including an image. The online system accesses and applies a trained machine-learning model to predict a probability that the content item includes an image of an item associated with the entity. If the probability is at least a threshold probability, a link to a page associated with the item that includes a set of the information describing the item(s) is added to the content item by the online system. Responsive to receiving an interaction with the link from a user presented with the content item, the online system determines a measure of similarity between the item and each additional item based on the information describing the item(s), arranges the set of the information describing the item(s) within the page based on the measure(s) of similarity, and sends the page for display to the user.

What is claimed is: 
     
         1 . A method comprising:
 receiving information describing one or more items associated with an entity having a presence on an online system, the information comprising one or more images of each of the one or more items;   receiving a content item to be presented to one or more viewing users of the online system, the content item comprising an image;   accessing a trained machine-learning model, the trained machine-learning model trained based at least in part on the information describing the one or more items;   for at least one target item of the plurality of items associated with the entity, applying the trained machine-learning model to a set of attributes of the content item to predict a probability that the image of the content item contains the target item;   determining that the image contains the target item based on the predicted probability;   adding, to the content item, a link to a page associated with the target item; and   sending the content item comprising the link for display to the one or more viewing users of the online system.   
     
     
         2 . The method of  claim 1 , wherein the information describing the one or more items comprises one or more of: a color associated with an item of the one or more items, a style associated with an item of the one or more items, a genre associated with an item of the one or more items, a material associated with an item of the one or more items, a type associated with an item of the one or more items, a brand associated with an item of the one or more items, an artist associated with an item of the one or more items, a version associated with an item of the one or more items, a model associated with an item of the one or more items, a set of specifications associated with an item of the one or more items, and a price associated with an item of the one or more items. 
     
     
         3 . The method of  claim 1 , further comprising:
 receiving an interaction with the link from a viewing user of the one or more viewing users;   responsive to receiving the interaction with the link, determining a measure of similarity between the item and each additional item of the one or more items, the measure of similarity determined based at least in part on the information describing the one or more items; and   rearranging the set of the information describing the one or more items within the page based at least in part on the measure of similarity between the item and each additional item of the one or more items.   
     
     
         4 . The method of  claim 3 , wherein determining the measure of similarity between the item and each additional item of the one or more items comprises:
 for each of the one or more items, generating an embedding in an embedding space based at least in part on a set of the information describing a corresponding item;   identifying an embedding corresponding to the item; and   for each additional item of the one or more items:
 identifying an additional embedding corresponding to the additional item, 
 determining a distance between the embedding corresponding to the item and the additional embedding corresponding to the additional item, and 
 determining the measure of similarity between the item and the additional item based at least in part on the distance. 
   
     
     
         5 . The method of  claim 3 , further comprising:
 responsive to receiving the interaction with the link, identifying a set of the one or more items having at least a threshold measure of similarity to the item based at least in part on the measure of similarity between the item and each additional item of the one or more items; and   including information describing the item and the set of the one or more items having at least the threshold measure of similarity to the item in the page.   
     
     
         6 . The method of  claim 5 , wherein identifying the set of the one or more items having at least the threshold measure of similarity to the item comprises:
 for each of the one or more items, generating an embedding in an embedding space based at least in part on a set of the information describing a corresponding item;   identifying an embedding corresponding to the item;   identifying a set of embeddings within a threshold distance of the embedding corresponding to the item; and   identifying the set of the one or more items having at least the threshold measure of similarity to the item, wherein the set of the one or more items correspond to the set of embeddings.   
     
     
         7 . The method of  claim 5 , wherein identifying the set of the one or more items having at least the threshold measure of similarity to the item comprises:
 determining a set of metadata associated with each of the one or more items based at least in part on the information describing the one or more items;   generating a hierarchy of categories associated with the one or more items based at least in part on the set of metadata associated with each of the one or more items, the hierarchy of categories comprising a plurality of nodes and one or more edges, wherein each of the plurality of nodes represents a category and each of the one or more edges connecting a pair of the plurality of nodes represents a relationship between a plurality of categories represented by the pair of nodes;   identifying a node corresponding to a category of the hierarchy of categories associated with the item based at least in part on a set of the information describing the item;   for each additional item of the one or more items, identifying an additional node corresponding to an additional category of the hierarchy of categories associated with the additional item based at least in part on a set of the information describing a corresponding item;   identifying a set of nodes within a threshold distance of the node corresponding to the category of the hierarchy of categories associated with the item; and   identifying the set of the one or more items having at least the threshold measure of similarity to the item, wherein the set of the one or more items correspond to the set of nodes.   
     
     
         8 . The method of  claim 1 , wherein the set of attributes of the content item comprises one or more selected from the group consisting of: a set of pixel values associated with the image comprising the content item, a tag comprising the content item, a link comprising the content item, a caption comprising the content item, a location associated with the content item, a comment on the content item, an audience associated with the content item, an inferred relationship between the content item and a topic, and any combination thereof. 
     
     
         9 . The method of  claim 1 , wherein the content item is received from the entity. 
     
     
         10 . The method of  claim 1 , wherein the content item is received from a content-providing user of the online system other than the entity. 
     
     
         11 . A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to:
 receive information describing one or more items associated with an entity having a presence on an online system, the information comprising one or more images of each of the one or more items;   receive a content item to be presented to one or more viewing users of the online system, the content item comprising an image;   access a trained machine-learning model, the trained machine-learning model trained based at least in part on the information describing the one or more items;   for at least one target item of the plurality of items associated with the entity, apply the trained machine-learning model to a set of attributes of the content item to predict a probability that the image of the content item contains the target item;   determine that the image contains the target item based on the predicted probability;   add, to the content item, a link to a page associated with the target item; and   send the content item comprising the link for display to the one or more viewing users of the online system.   
     
     
         12 . The computer program product of  claim 11 , wherein the information describing the one or more items comprises one or more of: a color associated with an item of the one or more items, a style associated with an item of the one or more items, a genre associated with an item of the one or more items, a material associated with an item of the one or more items, a type associated with an item of the one or more items, a brand associated with an item of the one or more items, an artist associated with an item of the one or more items, a version associated with an item of the one or more items, a model associated with an item of the one or more items, a set of specifications associated with an item of the one or more items, and a price associated with an item of the one or more items. 
     
     
         13 . The computer program product of  claim 11 , wherein the non-transitory computer-readable storage medium further has instructions encoded thereon that, when executed by the processor, cause the processor to:
 receive an interaction with the link from a viewing user of the one or more viewing users;   responsive to receiving the interaction with the link, determine a measure of similarity between the item and each additional item of the one or more items, the measure of similarity determined based at least in part on the information describing the one or more items; and   rearrange the set of the information describing the one or more items within the page based at least in part on the measure of similarity between the item and each additional item of the one or more items.   
     
     
         14 . The computer program product of  claim 13 , wherein determine the measure of similarity between the item and each additional item of the one or more items comprises:
 for each of the one or more items, generate an embedding in an embedding space based at least in part on a set of the information describing a corresponding item;   identify an embedding corresponding to the item; and   for each additional item of the one or more items:
 identify an additional embedding corresponding to the additional item, 
 determine a distance between the embedding corresponding to the item and the additional embedding corresponding to the additional item, and 
 determine the measure of similarity between the item and the additional item based at least in part on the distance. 
   
     
     
         15 . The computer program product of  claim 13 , wherein the non-transitory computer-readable storage medium further has instructions encoded thereon that, when executed by the processor, cause the processor to:
 responsive to receiving the interaction with the link, identify a set of the one or more items having at least a threshold measure of similarity to the item based at least in part on the measure of similarity between the item and each additional item of the one or more items; and   including information describe the item and the set of the one or more items having at least the threshold measure of similarity to the item in the page.   
     
     
         16 . The computer program product of  claim 15 , wherein identify the set of the one or more items having at least the threshold measure of similarity to the item comprises:
 for each of the one or more items, generate an embedding in an embedding space based at least in part on a set of the information describing a corresponding item;   identify an embedding corresponding to the item;   identify a set of embeddings within a threshold distance of the embedding corresponding to the item; and   identify the set of the one or more items having at least the threshold measure of similarity to the item, wherein the set of the one or more items correspond to the set of embeddings.   
     
     
         17 . The computer program product of  claim 15 , wherein identify the set of the one or more items having at least the threshold measure of similarity to the item comprises:
 determine a set of metadata associated with each of the one or more items based at least in part on the information describing the one or more items;   generate a hierarchy of categories associated with the one or more items based at least in part on the set of metadata associated with each of the one or more items, the hierarchy of categories comprising a plurality of nodes and one or more edges, wherein each of the plurality of nodes represents a category and each of the one or more edges connecting a pair of the plurality of nodes represents a relationship between a plurality of categories represented by the pair of nodes;   identify a node corresponding to a category of the hierarchy of categories associated with the item based at least in part on a set of the information describing the item;   for each additional item of the one or more items, identify an additional node corresponding to an additional category of the hierarchy of categories associated with the additional item based at least in part on a set of the information describing a corresponding item;   identify a set of nodes within a threshold distance of the node corresponding to the category of the hierarchy of categories associated with the item; and   identify the set of the one or more items having at least the threshold measure of similarity to the item, wherein the set of the one or more items correspond to the set of nodes.   
     
     
         18 . The computer program product of  claim 11 , wherein the set of attributes of the content item comprises one or more selected from the group consisting of: a set of pixel values associated with the image comprising the content item, a tag comprising the content item, a link comprising the content item, a caption comprising the content item, a location associated with the content item, a comment on the content item, an audience associated with the content item, an inferred relationship between the content item and a topic, and any combination thereof. 
     
     
         19 . The computer program product of  claim 11 , wherein the content item is received from the entity. 
     
     
         20 . The computer program product of  claim 11 , wherein the content item is received from a content-providing user of the online system other than the entity.