Yahoo! Inc.

Facebook, Inc.

Excalibur Ip, LLC

Deepayan Chakrabarti

Kunal Punera

Shanmugasundaram Ravikumar

Vanja Josifovski

Jonathan D Chang

Swapnil Hajela

Emre Velipasaoglu

Andrew Stephen Tomkins

Srinivas Vadrevu

Lei Duan

Aditya Pal

Avery Li Kuang Ching

Fan Li

Maja Kabiljo

Deepak K Agrawal

Andrei Z Broder

Deepak Agarwal

Anitha Kannan

Amaç Herdagdelen

Sofus Attila Macskassy

Sourav Chatterji

Xuanhui Wang

Stanislav Funiak

Karthik Subbian

Sumit Taank

Manav Ratan Mital

Dejan Diklic

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.

Semi-supervised learning via deep label propagation

Systems, methods, and non-transitory computer-readable media can determine one or more respective topics of interest for at least some users of a social networking system. At least some of the topics can be propagated to at least a first user, wherein the propagated topics were determined to be of interest to users that follow the first user in the social networking system. At least one topic from the propagated topics for which the first user is a topical authority is determined.

Systems and methods for recommendation of topical authorities

A social networking system identifies users to receive a recommendation to establish a connection to an object maintained by the social networking system. The social networking system determines one or more classifiers identifying attributes of users to receive the recommendation based on attributes of users connected to the object and additional users connected to those users. The attributes of an additional user may be weighted by a factor that provides a measure of the overlap between the attributes of the additional user and a user connected to the object.

Selecting users to receive a recommendation to establish connection to an object in a social networking system

A method and apparatus are provided for better web ad matching by combining relevance with consumer click feedback. In one example, the method includes receiving a query page, extracting features from the query page, re-weighting the query page, evaluating the query page in light of each ad in order to score each ad and pick substantially best ad matches of the indexed ads, and returning the substantially best ad matches to the consumer computer.

Method and apparatus for web ad matching

Embodiments are disclosed for striping a directed graph, e.g., a social graph, so as to efficiently perform an operation to each node in the directed graph. At least some of the embodiments can select first and second sets of nodes from the directed graph to form first and second stripes. The first and second sets of nodes are selected, for example, based on available computing resources. First and second intermediate results can be generated by performing the operation to each node of the first and the second stripes, respectively. The operation iteratively performs a superstep. The first and the second intermediate results are combined to form a collective result as an output of the superstep.

Striping of directed graphs and nodes with improved functionality

At least one embodiment of this disclosure includes a method of inferring attribute labels for a user in a social networking system based on the user's social connections and user-specified attribute labels in the social networking system. The method can include: establishing variational equations based on attribute labels of nodes in an ego network in a social graph of a social networking system; determining likelihood scores for at least a portion of the attribute labels of neighboring nodes from a focal user node in the ego network based on user-specified attribute labels from the social networking system; and calculating probability distributions of possible attribute labels for the focal user node of the ego network based on the variational equations and the likelihood scores.

Label inference in a social network

Exemplary methods and apparatuses are provided which may be used for classifying and indexing segmented portions of web pages and providing related information for use in information extraction and/or information retrieval systems. In an embodiment, an index of segmented portions may be used by a search engine to respond to a search query. In an embodiment, one or more machine learned models may be used to identify one or more feature properties of a plurality of segmented portions within one or more files, or otherwise inferable from the one or more files. In an embodiment, one or more machine learned models may be used to classify one or more of a plurality of segmented portions as being at least one of a plurality of segment types.

Automatic classification of segmented portions of web pages

An optimization-based framework is utilized to extract broad query aspects from query reformulations performed by users in historical user session logs. Objective functions are optimized to yield query aspects. At run-time, the best broad but unspecified query aspects relevant to any user query are presented along with the results of the run time query.

Mining broad hidden query aspects from user search sessions

A method and apparatus for receiving training data that comprise a plurality of event-and-time-specific texts that are contextually related to a plurality of events; iteratively processing the training data to generate a modified network model that defines a plurality of states; receiving additional data that comprise a plurality of additional event-and-time-specific texts that are contextually related to a particular event; processing the additional data by applying the modified network model to the additional data to identify, within the plurality of additional event-and-time specific texts, a particular set of texts that belong to a particular state of the plurality of states; identifying, within the particular set of texts, one or more texts that are most representative of all texts in the particular set of texts that belong to the particular state; wherein the method is performed by one or more special-purpose computing devices.

Method for summarizing event-related texts to answer search queries

An improved system and method for evolutionary clustering of sequential data sets is provided. A snapshot cost may be determined for representing the data set for a particular clustering method used and may determine the cost of clustering the data set independently of a series of clusterings of the data sets in the sequence. A history cost may also be determined for measuring the distance between corresponding clusters of the data set and the previous data set in the sequence of data sets to determine a cost of clustering the data set as part of a series of clusterings of the data sets in the sequence. An overall cost may be determined for clustering the data set by minimizing the combination of the snapshot cost and the history cost. Any clustering method may be used, including flat clustering and hierarchical clustering.

System and method for evolutionary clustering of sequential data sets

Exemplary methods and apparatuses are provided which may be used for classifying and indexing segmented portions of web pages and providing related information for use in information extraction and/or information retrieval systems.

Method, apparatus and computer readable medium for indexing advertisements to combine relevance with consumer click feedback

Methods, computer programs, and systems for generating a link title for a URL (Uniform Resource Locator) within a context webpage to be shown as a web result are provided. The method evaluates generation parameters for a plurality of sources for picking words from the link title. Further, the method generates candidates for the link title, and a likelihood is computed for each candidate. When computing the likelihood, the generation parameters, the context webpage and the words are considered. In addition, the method selects a candidate with the highest likelihood from all the computed likelihoods, and presents the URL with the selected candidate as the title.

Growing community of inventors

Mountain View, CA, United States of America

Deepayan Chakrabarti