Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing large datasets using a computationally-efficient representation are disclosed. A request to apply a coverage algorithm to a large input dataset is received. The large dataset includes sets of elements. A computationally-efficient representation of the large dataset is generated by generating a reduced set of elements that contains fewer elements based on a defined probability. For each element in the reduced set, a determination is made regarding whether the element appears in more than a threshold number of sets. When the element appears in more than the threshold number, the element is removed from sets until the element appears in only the threshold number. The coverage algorithm is then applied to the computationally-efficient representation to identify a subset of the sets. The system provides data identifying the subset of the sets in response to the received request.

Abstract: This specification describes technologies relating to displaying online content. In general, one aspect of the subject matter described in this specification can be embodied in methods that include determining a timing penalty for a content item, the timing penalty based in part on weights of previous impressions the content item has been allocated during a plurality of intervals within a delivery period. The methods may further include determining a weight associated with a pairing of the content item and an impression, the weight based in part on characteristics of the impression. The methods may further include determining a score for the content item, based in part on the weight and the timing penalty. The methods may further include selecting one of a set of matching content items, based in part on the score for the content item, and allocating the selected content item in response to a request.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing large datasets using a computationally-efficient representation are disclosed. A request to apply a coverage algorithm to a large input dataset is received. The large dataset includes sets of elements. A computationally-efficient representation of the large dataset is generated by generating a reduced set of elements that contains fewer elements based on a defined probability. For each element in the reduced set, a determination is made regarding whether the element appears in more than a threshold number of sets. When the element appears in more than the threshold number, the element is removed from sets until the element appears in only the threshold number. The coverage algorithm is then applied to the computationally-efficient representation to identify a subset of the sets. The system provides data identifying the subset of the sets in response to the received request.

Abstract: Systems and methods offer an efficient approach to computing similarity rankings in bipartite graphs. An example system includes at least one processor and memory storing a bipartite graph having a first set and a second set of nodes, with nodes in the first set being connected to nodes in the second set by edges. The memory also stores instructions that, when executed by the at least one processor, cause the system to assign each node in the second set to one of a plurality of categories and, for each of the plurality of categories, generate a subgraph. The subgraph comprises of a subset of nodes in the first set and edges linking the nodes in the subset, where the nodes in the subset are selected based on connection to a node in the second set that is assigned to the category. The system uses the subgraph to respond to queries.

Abstract: Systems and methods for sending asynchronous messages include receiving, using at least one processor, at a node in a distributed graph, a message with a first value and determining, at the node, that the first value replaces a current value for the node. In response to determining that the first value replaces the current value, the method also includes setting a status of the node to active and sending messages including the first value to neighboring nodes. The method may also include receiving the messages to the neighboring nodes at a priority queue. The priority queue propagates messages in an intelligently asynchronous manner, and the priority queue propagates the messages to the neighboring nodes, the status of the node is set to inactive. The first value may be a cluster identifier or a shortest path identifier.

Abstract: Systems and methods for improving the time and cost to calculate connected components in a distributed graph are disclosed. One method includes reducing a quantity of map-reduce rounds used to determine a cluster assignment for a node in a large distributed graph by alternating between two hashing functions in the map stage of a map-reduce round and storing the cluster assignment for the node in a memory. Another method includes reducing a quantity of messages sent during map-reduce rounds by performing a predetermined quantity of rounds to generate, for each node, a set of potential cluster assignments, generating a data structure in memory to store a mapping between each node and its potential cluster assignment, and using the data structure during remaining map-reduce rounds, wherein the remaining map-reduce rounds do not send messages between nodes. The method can also include storing the cluster assignment for the node in a memory.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for analyzing user profiles. One of the methods includes identifying websites with which users of a social networking service have established an affiliation. The method includes filtering the identified websites by removing websites in which a count of users who have established an affiliation with the website exceeds a threshold. The method includes producing a list of valid affiliations for each filtered website, including generating a validity score for each of the established affiliations with the respective filtered website, determining whether each of the established affiliations with the respective filtered website is valid based on the corresponding validity score, and ranking the users who established valid affiliations with the respective website based on the corresponding validity scores. The method also includes providing the list of the valid affiliations ordered by the ranking.

Abstract: Systems and methods offer an efficient approach to computing similarity rankings in bipartite graphs. An example system includes at least one processor and memory storing a bipartite graph having a first set and a second set of nodes, with nodes in the first set being connected to nodes in the second set by edges. The memory also stores instructions that, when executed by the at least one processor, cause the system to assign each node in the second set to one of a plurality of categories and, for each of the plurality of categories, generate a subgraph. The subgraph comprises of a subset of nodes in the first set and edges linking the nodes in the subset, where the nodes in the subset are selected based on connection to a node in the second set that is assigned to the category. The system uses the subgraph to respond to queries.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for analyzing user profiles. One of the methods includes identifying websites with which users of a social networking service have established an affiliation. The method includes filtering the identified websites by removing websites in which a count of users who have established an affiliation with the website exceeds a threshold. The method includes producing a list of valid affiliations for each filtered website, including generating a validity score for each of the established affiliations with the respective filtered website, determining whether each of the established affiliations with the respective filtered website is valid based on the corresponding validity score, and ranking the users who established valid affiliations with the respective website based on the corresponding validity scores. The method also includes providing the list of the valid affiliations ordered by the ranking.

Abstract: Systems and methods for improving the time and cost to calculate connected components in a distributed graph are disclosed. One method includes reducing a quantity of map-reduce rounds used to determine a cluster assignment for a node in a large distributed graph by alternating between two hashing functions in the map stage of a map-reduce round and storing the cluster assignment for the node in a memory. Another method includes reducing a quantity of messages sent during map-reduce rounds by performing a predetermined quantity of rounds to generate, for each node, a set of potential cluster assignments, generating a data structure in memory to store a mapping between each node and its potential cluster assignment, and using the data structure during remaining map-reduce rounds, wherein the remaining map-reduce rounds do not send messages between nodes. The method can also include storing the cluster assignment for the node in a memory.

Abstract: Systems and methods for sending asynchronous messages include receiving, using at least one processor, at a node in a distributed graph, a message with a first value and determining, at the node, that the first value replaces a current value for the node. In response to determining that the first value replaces the current value, the method also includes setting a status of the node to active and sending messages including the first value to neighboring nodes. The method may also include receiving the messages to the neighboring nodes at a priority queue. The priority queue propagates messages in an intelligently asynchronous manner, and the priority queue propagates the messages to the neighboring nodes, the status of the node is set to inactive. The first value may be a cluster identifier or a shortest path identifier.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for identifying trends via representative micro-posts. One of the methods includes identifying query terms; selecting candidate documents that satisfy one or more of the terms of the query terms; selecting, from the candidate documents, candidate documents based on a weighted-coverage function for the query terms; pairing documents from the candidate documents based on a distance between the paired documents; generating one or more clusters from the paired documents, a cluster being associated with one or more topics of the query terms; selecting from the candidate documents, one or more particular documents, for one or more clusters, based on a diversity function; and providing for presentation the particular documents as representative documents that cover one or more topics associated with the one or more clusters.

Abstract: This specification describes technologies relating to selection and delivery of online content. One aspect of the subject matter described in this specification can be embodied in methods that include determining a share fraction based on a received reserve price and based in part on a distribution of past bids for content inventory units in one or more content slots provided by a publisher. The methods may further include determining a second reserve price based in part on the received reserve price and based in part on a distribution of past bids for content inventory units in one or more content slots provided by the publisher. The methods may further include receiving one or more bids for the content inventory unit and allocating the content inventory unit to a buyer based in part on the one or more bids and the second reserve price.

Abstract: Methods, systems, and apparatus for allocating advertising budgets. In one aspect, a method includes receiving a set of candidate keywords specifying targeting keywords for content items. Each targeting keyword can be a phrase to be matched for a content item to be eligible to be provided in response to a content item request. For each of the targeting keywords, a probability distribution specifying likelihoods of specified user actions occurring in response to a content item is received. Based on the probability distributions, a set of final keywords with which distribution of the content items will be controlled is selected, the set of final keywords being one or more keywords for which a likelihood of a specified outcome occurring exceeds a threshold likelihood. The set of final keywords is a set which a total cost of distributing the content items is below a specified value.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for content distribution. In one aspect, a method includes obtaining reservation data specifying a reservation period and a number of impression reservations for the reservation period; receiving, during the reservation period, content item request data specifying that a resource was requested; identifying candidate content item providers; determining an expected revenue that will be obtained by selecting one of the candidate content item providers to provide a content item; providing, to a content item exchange, auction data including a reserve price; receiving positive response data specifying that the content item exchange will provide an exchange content item for a price that meets the reserve price; and in response to receiving the positive response data, selecting the exchange content item as the content item to be provided in response to receiving the content item request data.

Abstract: Methods, systems, and apparatus for content distribution. In one aspect, a method includes obtaining reservation data for a resource, the reservation data specifying a reservation period and a number of impression reservations for the reservation period; receiving, during the reservation period, content item request data specifying that the resource was requested; identifying candidate content item providers; determining an expected revenue that will be obtained by selecting one of the candidate content item providers to provide a content item providing, to a content item exchange, auction data including a reserve price; receiving positive response data specifying that the content item exchange will provide an exchange content item for a price that meets the reserve price; and in response to receiving the positive response data, selecting the exchange content item as the content item to be provided in response to receiving the content item request data.

Abstract: Systems and methods for allocating and providing advertisements in response to impression events (e.g., a request for a placement of an advertisement) corresponding to forecasted impressions (e.g., a forecast of a request for a placement of an advertisement). The advertisements are allocated according to solution data that includes two disjoint solution sets, with each solution set identifying an advertisement that can be allocated to a forecasted impression.

Abstract: A SPAM detection system is provided. The system includes a graph clustering component to analyze web data. A link analysis component can be associated with the graph clustering component to facilitate SPAM detection in accordance with the web data.

Abstract: The claimed subject matter relates to an architecture that can identify, store, and/or output local contributions to a rank of a vertex in a directed graph. The architecture can receive a directed graph and a parameter, and examine a local subset of vertices (e.g., local to a given vertex) in order to determine a local supporting set. The local supporting set can include a local set of vertices that each contributes a minimum fraction of the parameter to a rank of the vertex. The local supporting set can be the basis for an estimate of the supporting set and/or rank of the vertex for the entire graph and can be employed as a means for detecting link or web spam as well as other influence-based social network applications.

Abstract: In one implementation, a computer-implemented method includes receiving, at a server system, a request for an advertisement to provide to a first user of a social network, and determining, for each of a plurality of advertisements, a probability that the first user will select the advertisement based, at least in part, on previous propagations of the advertisement by one or more second users of the social network. The method can further include scoring, by the server system, the plurality of advertisements based upon the determined probabilities of selection by the first user and bids associated with the plurality of advertisements, and providing one or more of the plurality of advertisements for presentation to the first user based upon the scoring of the plurality of advertisements.