Back to results
US 10,769,222 B2Granted
Search result ranking based on post classifiers on online social networks
AssigneeFacebook, Inc.
InventorSai Nishanth Parepally
Life of this patent
Estimating the term…
Abstract
In one embodiment, a method includes generating a reconstructed embedding of a query based on one or more term embeddings associated with the one or more query terms, respectively, on receiving a query with the one or more query terms, formulating an evaluation model based at least on the reconstructed embedding of the query, where the evaluation model calculates a relevance score for posts with respect to the search query based at least on the classifier vectors of the posts, and calculating, for each of the retrieved posts, a relevance score for the post by applying the associated classifier vector to the formulated evaluation model.
Claims19
A method comprising, by a computing device:
- receiving, from a first client system of a first user of an online social network, a search query for posts of the online social network, the search query comprising one or more query terms;
- retrieving, from one or more verticals of the online social network, a plurality of posts of the online social network, each retrieved post containing in the text of the post at least one of the one or more query terms, wherein each post is associated with a pre-calculated classifier vector comprising a plurality of classifier scores corresponding to a plurality of classifications, respectively, wherein each classifier score in a classifier vector associated with a post represents a degree of relevance of the post to the corresponding classifier;
- generating a reconstructed embedding of the query based on one or more term embeddings associated with the one or more query terms, respectively, wherein each of the one or more term embeddings corresponds to a point in a d-dimensional embedding space, and wherein the reconstructed embedding of the query is generated by pooling the one or more term embeddings associated with the one or more of the query terms, respectively;
- formulating an evaluation model based at least on the reconstructed embedding of the query, wherein the evaluation model calculates a relevance score for posts with respect to the search query based at least on the classifier vectors of the posts;
- calculating, for each of the retrieved posts, a relevance score for the post by applying the associated classifier vector to the formulated evaluation model;
- ranking the retrieved posts based on their respective relevance scores;
- and
- sending, to the first client system in response to the search query, instructions for generating a search-results interface for presentation to the first user, the search-results interface comprising references to one or more of the posts presented in ranked order.
5 more elements
A system comprising:
- one or more processors;
- and a non-transitory memory coupled to the processors comprising instructions executable by the processors, the processors operable when executing the instructions to:
- receive, from a first client system of a first user of an online social network, a search query for posts of the online social network, the search query comprising one or more query terms;
- retrieve, from one or more verticals of the online social network, a plurality of posts of the online social network, each retrieved post containing in the text of the post at least one of the one or more query terms, wherein each post is associated with a pre-calculated classifier vector comprising a plurality of classifier scores corresponding to a plurality of classifications, respectively, wherein each classifier score in a classifier vector associated with a post represents a degree of relevance of the post to the corresponding classifier;
- generate a reconstructed embedding of the query based on one or more term embeddings associated with the one or more query terms, respectively, wherein each of the one or more term embeddings corresponds to a point in a d-dimensional embedding space, and wherein the reconstructed embedding of the query is generated by pooling the one or more term embeddings associated with the one or more of the query terms, respectively;
- formulate an evaluation model based at least on the reconstructed embedding of the query, wherein the evaluation model calculates a relevance score for posts with respect to the search query based at least on the classifier vectors of the posts;
- calculate, for each of the retrieved posts, a relevance score for the post by applying the associated classifier vector to the formulated evaluation model;
- rank the retrieved posts based on their respective relevance scores;
- and
- send, to the first client system in response to the search query, instructions for generating a search-results interface for presentation to the first user, the search-results interface comprising references to one or more of the posts presented in ranked order.
7 more elements
One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
- receive, from a first client system of a first user of an online social network, a search query for posts of the online social network, the search query comprising one or more query terms;
- retrieve, from one or more verticals of the online social network, a plurality of posts of the online social network, each retrieved post containing in the text of the post at least one of the one or more query terms, wherein each post is associated with a pre-calculated classifier vector comprising a plurality of classifier scores corresponding to a plurality of classifications, respectively, wherein each classifier score in a classifier vector associated with a post represents a degree of relevance of the post to the corresponding classifier;
- generate a reconstructed embedding of the query based on one or more term embeddings associated with the one or more query terms, respectively, wherein each of the one or more term embeddings corresponds to a point in a d-dimensional embedding space, and wherein the reconstructed embedding of the query is generated by pooling the one or more term embeddings associated with the one or more of the query terms, respectively;
- formulate an evaluation model based at least on the reconstructed embedding of the query, wherein the evaluation model calculates a relevance score for posts with respect to the search query based at least on the classifier vectors of the posts;
- calculate, for each of the retrieved posts, a relevance score for the post by applying the associated classifier vector to the formulated evaluation model;
- rank the retrieved posts based on their respective relevance scores;
- and
- send, to the first client system in response to the search query, instructions for generating a search-results interface for presentation to the first user, the search-results interface comprising references to one or more of the posts presented in ranked order.
5 more elements
Dependencies are inferred from the claim text.