Abstract: Items within an index may be converted from classic geometry and embedded into a hyperbolic space. The hyperboloids within the hyperbolic space provide higher precision classifications of items within the index relative to their hierarchical structure. A received search query may also be converted to hyperbolic space and mapped as a query hyperboloid against an answer space that includes hyperboloids for items within the index. Distances or overlaps between the hyperboloids may be determined in order to generate a set of search results.

Abstract: Technologies are provided for reformulation of a tail query to a head query with the same purchase intent by mapping the tail query to the head query. In some of the technologies, a reasonable embedding can be learned on historical head queries. The embedding can then be refined by leveraging rewards generated from a persistently noisy oracle that compensates for the lack of historical behavioral signal for tail queries. Further, a contextual sampling technique that uses text-based rewards or oracle-based rewards, or both, can be implemented in order to avoid biases introduced by persistent noise in the oracle. Numerical experiments on large scale e-commerce datasets demonstrate that the provided technologies can outperform several conventional approaches to query reformulation.

Abstract: Technologies are provided for reformulation of a tail query to a head query with the same purchase intent by mapping the tail query to the head query. In some of the technologies, a reasonable embedding can be learned on historical head queries. The embedding can then be refined by leveraging rewards generated from a persistently noisy oracle that compensates for the lack of historical behavioral signal for tail queries. Further, a contextual sampling technique that uses text-based rewards or oracle-based rewards, or both, can be implemented in order to avoid biases introduced by persistent noise in the oracle. Numerical experiments on large scale e-commerce datasets demonstrate that the provided technologies can outperform several conventional approaches to query reformulation.

Abstract: Various embodiments provide for selecting a subset of features to use to train a model for search applications. To select a feature, the candidate features are randomly assigned into two groups. Each of the two groups represents a summation of the respective features that were assigned to it. Then a decision tree building scan is performed on the two groups to determine which of the two groups performs better based a selection criteria. Upon determining which of the two groups is better, the candidate features of the winning group are again randomly assigned into two groups. These two groups are again scanned as described above to determine a winning group. This binary splitting and scanning pattern is continuously performed until the winning group contains one remaining feature. That remaining feature is then designated as a selected feature to be used in the search model.

Abstract: A method, system and computer program product are disclosed for predicting influence in a social network. In one embodiment, the method comprises identifying a set of users of the social network, and identifying a subset of the users as influential users based on defined criteria. A multitude of measures are identified as predictors of which ones of the set of users are the influential users. These measures are aggregated, and a composite predictor model is formed based on this aggregation. This composite predictor model is used to predict which ones of the set of users will have a specified influence in the social network in the future. In one embodiment, the specified influence is based on messages sent from the users, and for example, may be based on the number of the messages sent from each user that are re-sent by other users.

Abstract: In one embodiment, a system may access a graph data structure that includes nodes and connections between the nodes. Each node may be associated with a user; each connection between two nodes may represent a relationship between the associated users; and each node may be either labeled or unlabeled with respect to a label type. For each labeled node, a label of the label type of that labeled node may be propagated to other nodes through the connections. For each node, the system may store a label distribution information associated with the label type based on the propagated labels reaching the node. The system may train a machine-learning model using the labels and the label distribution information of a set of the labeled nodes. A predicted label for each unlabeled node may be generated using the model and the label distribution information of the unlabeled node.

Abstract: Systems, methods, and non-transitory computer readable media can determine a time at which to prefetch one or more content items to be included in a feed associated with a user. The feed can be provided by a social networking system. A number of content items to prefetch at the determined time can be determined. One or more content items to be included in the feed associated with the user can be prefetched based at least in part on the determined time and the determined number of content items. The determined time can be prior to a time at which the user is expected to access an application associated with the social networking system.

Abstract: Certain embodiments described herein relate to peer-to-peer content distribution. In one embodiment, a method includes a first device receiving content and determining a content categorization of the received content. The first device may detect a second computing device and communicate with that it through a direct wireless connection (e.g., Bluetooth). Through the direct wireless connection, the first device may receive information associated with a user of the second computing device from the second device. Based on the information associated with the user and the content categorization of the content, the first device may determine a likelihood of the user being interested in the content. The first device may push the content to the second computing device through the direct wireless connection based on the likelihood of the user being interested in the content.

Abstract: In one embodiment, a method includes identifying a first user node that corresponds to a first user of a social-networking system for whom recommendation candidates are to be generated, where the social-networking system comprises a social graph that comprises nodes and edges representing relationships between the users. The method further includes performing one or more steps of a computation that implements a random walk of the nodes of a social graph, and generates a ranking value for each user node that satisfies one or more constraints, wherein the ranking value represents an importance of the user node to other user nodes in the social graph in accordance with the relationships represented by the edges, and selecting one or more candidate users to be recommended to a particular user based on the ranking values associated with the user nodes.

Abstract: In one embodiment, a method includes receiving, from a client system associated with a user of an online social network, a search query for entities in the online social network, the search query containing one or more n-grams, generating a query embedding corresponding to the search query, where the query embedding represents the search query as a point in a d-dimensional embedding space, retrieving multiple entity embeddings corresponding to a plurality of entities, respectively, where each entity embedding represents the corresponding entity as a point in the d-dimensional embedding space, calculating, for each of the retrieved entity embeddings, a similarity metric between the query embedding and the entity embedding, ranking the entities based on their respective calculated similarity metrics, and sending, to the client system in response to the search query, instructions for presenting a search-results interface.

Abstract: In one embodiment, a method includes determining a current location of a user based on location data received from a client device; and calculating a transition probability of the user transitioning, within a predetermined time window, from the current location to each of a number of candidate geographic locations. The calculating of the transition probability is based at least in part on previously logged location data associated with a number of users who were at the current location. The method also includes determining metadata associated with the user; and calculating an offline probability associated with each of the number of candidate geographic locations using a computer model and the metadata associated with the user. The computer model is generated using machine learning and metadata associated with users who were at the respective candidate geographic location.

Abstract: In one embodiment, a method includes determining a current location of a user based on location data received from a client device; and calculating a transition probability of the user transitioning, within a predetermined time window, from the current location to each of a number of candidate geographic locations. The calculating of the transition probability is based at least in part on previously logged location data associated with a number of users who were at the current location. The method also includes determining metadata associated with the user; and calculating an offline probability associated with each of the number of candidate geographic locations using a computer model and the metadata associated with the user. The computer model is generated using machine learning and metadata associated with users who were at the respective candidate geographic location.

Abstract: In one embodiment, a system may access a graph data structure that includes nodes and connections between the nodes. Each node may be associated with a user; each connection between two nodes may represent a relationship between the associated users; and each node may be either labeled or unlabeled with respect to a label type. For each labeled node, a label of the label type of that labeled node may be propagated to other nodes through the connections. For each node, the system may store a label distribution information associated with the label type based on the propagated labels reaching the node. The system may train a machine-learning model using the labels and the label distribution information of a set of the labeled nodes. A predicted label for each unlabeled node may be generated using the model and the label distribution information of the unlabeled node.

Abstract: Systems, methods, and non-transitory computer-readable media can identify a media content item for which media processing is to be performed. State information associated with the media content item can be acquired. At least some of the media processing can be enabled, based on the state information, to be performed client-side with respect to the media content item. The state information can indicate a next processing step of the at least some of the media processing that is to be performed. The state information can be updated based on the at least some of the media processing performed client-side.

Abstract: Certain embodiments described herein relate to peer-to-peer content distribution. In one embodiment, a method includes a first device receiving content and determining a content categorization of the received content. The first device may detect a second computing device and communicate with that it through a direct wireless connection (e.g., Bluetooth). Through the direct wireless connection, the first device may receive information associated with a user of the second computing device from the second device. Based on the information associated with the user and the content categorization of the content, the first device may determine a likelihood of the user being interested in the content. The first device may push the content to the second computing device through the direct wireless connection based on the likelihood of the user being interested in the content.

Abstract: Systems, methods, and non-transitory computer-readable media can determine one or more chunks for a content item to be captioned. Each chunk can include one or more terms that describe at least a portion of the subject matter captured in the content item. One or more sentiments are determined based on the subject matter captured in the content item. One or more emotions are determined for the content item. At least one emoted caption is generated for the content item based at least in part on the one or more chunks, sentiments, and emotions. The emoted caption can include at least one term that conveys an emotion represented by the subject matter captured in the content item.

Abstract: Search results are received that were generated by a search engine in response to a search query entered by a user. The search results include a first result which contains a first set of identifying data. The first result is matched with a first profile on a first social network which is merged with a second social network. The first profile contains a second set of identifying data which satisfies matching criteria for similarity with the first set of identifying data.

Abstract: A system and method for contextually interpreting image sequences are provided. The method comprises receiving video from one or more video sources, and generating one or more questions associated with one or more portions of the video based on at least one user-defined objective. The method further comprises sending the one or more portions of the video and the one or more questions to one or more assistants, receiving one or more answers to the one or more questions from the one or more assistants, and determining a contextual interpretation of the video based on the one or more answers and the video.

Abstract: Systems, methods, and non-transitory computer readable media can determine a time at which to prefetch one or more content items to be included in a feed associated with a user. The feed can be provided by a social networking system. A number of content items to prefetch at the determined time can be determined. One or more content items to be included in the feed associated with the user can be prefetched based at least in part on the determined time and the determined number of content items. The determined time can be prior to a time at which the user is expected to access an application associated with the social networking system.

Abstract: Systems, methods, and non-transitory computer-readable media can identify a media content item for which media processing is to be performed. State information associated with the media content item can be acquired. At least some of the media processing can be enabled, based on the state information, to be performed client-side with respect to the media content item. The state information can indicate a next processing step of the at least some of the media processing that is to be performed. The state information can be updated based on the at least some of the media processing performed client-side.