Facebook, Inc.

Facbook, Inc.

Kihyuk Hong

Sriram Sankar

Rajat Raina

Kittipat Virochsiri

Michael Curtiss

Chaitanya Mishra

Kedar Dhamdhere

Gintaras Andrius Woss

Zhongxian Chen

In one embodiment, a method includes accessing a first set of scored results received from a client system of a first user, the first set of scored results including: one or more results, each having a rank with respect to the other results, generated by a first search algorithm, and one or more scores inputted by the first user corresponding to the results. The method includes calculating a discounted cumulative gain for each scored result based on the rank and score of the result. The method includes generating a second search algorithm based on the calculated gains and the first search algorithm, the second search algorithm including a modification of the first search algorithm to improve the ranking of results, and, if a quality metric of the second search algorithm is higher than a quality metric of the first, substituting the second search algorithm for the first.

Ranking test framework for search results on an online social network

In one embodiment, a method includes receiving a query comprising one or more n-grams, generating a plurality of query commands based on a parsing of the query input, wherein each query command comprises an inner query constraint and an outer query constraint, scoring the query commands based at least in part on a number of objects matching the inner query constraint of the respective query command, selecting a first query command based at least in part on the respective scores of the query commands, and executing the first query command to identify a plurality of objects based on: (1) a first set of objects matching the first inner query constraint and at least in part matching the first outer query constraint, and (2) a second set of objects matching the first outer query constraint.

Search query interactions

In one embodiment, a method includes receiving a query comprising one or more n-grams, generating a plurality of query commands based on a parsing of the query input, wherein each query command comprises a plurality of query constraints, scoring the plurality of query commands based at least in part on a number of objects matching each of the query constraints of the respective query command, selecting a first query command based at least in part on the respective scores of the query commands, wherein the first query command comprises an inverse constraint corresponding to a first query constraint that has previously been flagged as identifying greater than a threshold number of objects and one or more second query constraints, and executing the first query command to identify a plurality of objects matching the inverse constraint and the one or more second query constraints.

Using inverse operators for queries

In one embodiment, a method includes receiving, from a client system of a first user, a request to post content to the online social network, identifying a plurality of second users, wherein each second user is associated with a privacy setting, and wherein a percentage of the identified second users have a privacy setting likely to make the content posted by the first user accessible to the second user, determining, for each identified second user, whether the content posted by the first user is accessible to the second user, and sending, to one or more client systems of one or more of the second users, respectively, instructions for presenting the content posted by the first user, wherein the content is sent to second users determined to be able to access the content and is not sent to second users determined to be not able to access the content.

Approximate privacy indexing for search queries on online social networks

In one embodiment, a method includes retrieving, from one or more data stores, one or more objects requested by an application accessed by a first user of an online social network, wherein each object is associated with a privacy setting, and wherein a percentage of the retrieved objects have a privacy setting likely to make the object visible to the first user, determining, for each retrieved object, a visibility of the object with respect to the first user, and sending, to the application accessed by the first user, one or more of the retrieved objects, wherein each sent object has a visibility that is visible to the first user, and wherein objects having a visibility that is not visible to the first user are not sent.

In one embodiment, a method includes receiving, from a client system of a first user, a query comprising one or more n-grams, parsing the one or more n-grams of the query to generate a query command, wherein the query command comprises an inner query constraint and an outer query constraint, identifying from a plurality of objects based on the query command (1) a first set of objects matching the inner query constraint and at least in part matching the outer query constraint, and (2) a second set of objects matching the outer query constraint, and generating one or more search results based on the first and second sets of objects, wherein each search result corresponds to an object of the plurality of objects.

In one embodiment, a method includes receiving, from a client system of a first user, a query comprising one or more n-grams, parsing the one or more n-grams of the query to identify a first query constraint and one or more second query constraints, wherein the first query constraint has been previously flagged as identifying greater than a threshold number of objects, generating a query command based on the query, wherein the query command comprises an inverse constraint associated with the first query constraint and the one or more second query constraints, and generating one or more search results based on the query command, each search result corresponding to an object of a plurality of objects.

In one embodiment, a method includes accessing a first set of scored results from a first user, the scored results comprising one or more results generated by a first search algorithm in response to a query from the first user, wherein the one or more results correspond to one or more content objects, respectively, the one or more results being personalized for the first user; and one or more scores inputted by the first user corresponding to the one or more results, respectively, calculating a discounted cumulative gain for each result in the first set of scored results based on the score inputted by the first user corresponding to the result, and modifying the first search algorithm based on the calculated gain for each result, wherein the first search algorithm is modified to improve the ranking of results personalized for the first user.

In one embodiment, a method includes receiving, from a client device of a first user of an online social network, a structured query comprising references to one or more selected objects associated with the online social network, parsing the structured query to identify a first query constraint and one or more second query constraints, identifying an inverse constraint associated with the first query constraint, wherein the first query constraint has been previously flagged as identifying greater than a threshold number of objects, and generating a query command based on the structured query, wherein the query command comprises the inverse constraint and the one or more second query constraints.

Using inverse operators for queries on online social networks

In one embodiment, a method includes receiving, from a client system of a first user of an online social network, a structured query comprising references to one or more selected objects associated with the online social network, generating a query command based on the structured query, wherein the query command comprises an inner query constraint and an outer query constraint, identifying a first set of objects matching the inner query constraint and at least in part matching the outer query constraint, identifying a second set of objects matching the outer query constraint, and generating one or more search results based on the first and second sets of objects, wherein each search result corresponds to an object of the plurality of objects.

Search query interactions on online social networks

In one embodiment, a method includes accessing a social graph comprising a plurality of nodes and edges, receiving a set of scored results from a user that include results generated by a search algorithm in response to a query from the user and a score for each result, where each result corresponds to a node of the social graph, calculating a gain for each result based on the score of the result, and modifying the search algorithm based on the calculated gain.

In one embodiment, a method includes accessing a social graph that includes a plurality of nodes and edges, receiving from a user a structured query comprising references to selected nodes and selected edges, generating a query command based on the structured query based on the structured query comprising a first query constraint and a second query constraint, identifying a first set of nodes matching the first query constraint and at least in part matching the second query constraint, identifying a second set of nodes matching the second query constraint, and generating search results based on the first and second set of nodes.

In one embodiment, a method includes accessing a social graph that includes a plurality of nodes and edges, receiving from a user a structured query comprising references to selected nodes and selected edges, parsing the structure query to identify a first query constraint and one or more second query constraints, identifying a inverse constraint associated with the first query constraint, and generating a query command based on the structured query, where the query command includes the inverse constraint and the one or more second query constraints.

In one embodiment, a method includes receiving from a first user of an online social network a search query, searching data stores to identify objects that match the search query, where identifying the objects is also based in part on the privacy setting for each object, determining for each identified object a visibility of the object with respect to the first user, and generating one or more search results corresponding to the search query.

Growing community of inventors

Seoul, South Korea

Kihyuk Hong