Yahoo! Inc.

Google Inc.

Excalibur Ip, LLC

Cong Yu

Sihem Amer-Yahia

Ronny Lempel

Nadav Golbandi

Philip Lewis Bohannon

Moran Feldman

Arun Sundaresan Iyer

Mridul Muralidharan

Anish Das Sarma

Munmun De Choudhury

Bo Pang

Alon Yitzchak Halevy

Evgeniy Gabrilovich

Sathiya Keerthi Selvaraj

Sergei Vassilvitskii

Ramana Yerneni

Fei Wu

Colin Keogh

Sundararajan Sellamanickam

Alpa Jain

Ziad Raymond Sultan

Lujun Fang

Abraham P Ittycheriah

Ashwin Machanavajjhala

Vinh Quoc Tran

Adam Lelkes

Julia Stoyanovich

Simon Baumgartner

Shan Jiang

Regis McHugh Gaughan, Iii

Xiao Ling

Alban Galland

Katrina Tempero

Munmum De Choudhury

Laks Lakshmanan

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, that facilitate automatic identification of a set of fact-check factors from digital documents. Digital documents can be identified from a plurality of sources. For each digital document, a set of fact check factors are identified using a trained sequence tagging model. Based on the sequence tagging model, a confidence value representing a likelihood that the set of fact check factors identified from the digital document are an actual set of fact check factors for the digital document is determined. The set of fact check factors is stored in association with the digital document. A request for fact check factors for a particular digital document among the digital documents is received from a fact checking entity. In response, the set of fact check factors identified from the particular digital document are provided to the fact checking entity.

Automatic identification of fact check factors

Methods and systems that facilitate generating and presenting a crossword puzzle. Methods include obtaining, from a content source, a plurality of data items. Using the plurality of data items, a dictionary of clue-word pairs are generated. A crossword layout is generated using a random crossword layout generator. The generated crossword layout is input to a model that outputs a likelihood that the input crossword layout results in a valid crossword. If the likelihood that the crossword layout results in a valid crossword puzzle satisfies a first threshold, a plurality of words in a set of clue-word pairs from among the plurality of clue-word pairs is inserted into the crossword layout. If the plurality of words inserted into the crossword layout results in a valid crossword puzzle, the crossword puzzle is provided to a user on a user device using the set of clue-word pairs.

Crossword puzzle generator

Systems and techniques are provided for generating a union table with from stitchable tables. Tables may be extracted from web pages to obtain extracted tables. Stitchable tables may be determined from the extracted tables. Hidden attributes for the stitchable tables may be extracted from the web pages from which the stitchable tables were extracted using segmentation of text for contextual data from the web pages into segment sequences, and alignment of the segment sequences. Iterative pairwise alignment may be used to align the segment sequences and obtain aligned segments. The stitchable tables may be joined into a union table. Hidden attributes from the aligned segments may be added to the union table. Headers for the hidden attributes in the union table may be labeled using a database of entities and class labels.

Synthesizing union tables from the web

Method, system, and programs for detecting dynamic relationship and discovering dynamic events. Data from a first data source is first received. At least one dynamic relation candidate is identified and each dynamic relation candidate involves multiple entities. The at least one dynamic relation candidate is identified based on temporal properties with respect to the entities exhibited in the data from the first data source. Dynamic relations are then extracted by corroborating the temporal properties of the entities involved in the at least one dynamic relation candidate with that of the same entities exhibited in data from a second data source. Then, a dynamic event that gives rise to the dynamic relations among different entities is detected.

Method and system for discovering dynamic relations among entities

A system and method is described for large scale entity-specific classification of each entity-specific set of candidates in a collection of candidates for each specific entity in a collection of entities. The collection of entities may comprise a specific category or domain of entities (e.g. schools, restaurants, manufacturers, products, events, people). Candidates may comprise webpages or other resources with resource identifiers. Entity specific sets of candidates may be found by leveraging search engine query results and user interaction therewith for queries based on entity-specific attributes. The relationship(s) or class(es) for which candidate resources are being classified relative to a specific entity may comprise an authoritative, official home page (OHP), or other class (e.g. fan page, review, aggregator) relative to a specific entity. A feature generator generates entity-specific features for candidates. In accordance with its features, one or more classifiers rank each candidate for a specific class for a specific entity.

Large scale entity-specific resource classification

Methods and systems are provided for providing recommendations to users of a computer-based network of items of potential interest to the users. Items and people of potential interest to users may be determined using obtained word-based social behavior information, semantically-sensitive vocabulary taxonomies, and determined implied topic-specific social networks. The user may be presented with a graphical user interface including the recommendation, an explanation of the rationale relating to the recommendation, and an opportunity for the user to provide feedback relating to the recommendation or the rationale. The feedback may be used to improve future recommendations.

Recommendation system using social behavior analysis and vocabulary taxonomies

One embodiment accesses two or more photos taken by one or more travelers at one or more destinations and one or more points-of-interest located within the destinations; constructs one or more photo streams for each unique traveler-destination combination, wherein each one of the photo streams comprises two or more of the photos taken by the corresponding traveler at the corresponding destination; maps each one of the photos to one of the points-of-interest; constructs one or more timed paths for each unique traveler-destination combination based on the photo streams and the mapping between the photos and the points-of-interest, wherein each one of the timed paths comprises one or more of the points-of-interest located within the corresponding destination and visited by the corresponding travel; and constructs an itinerary based on a start point-of-interest, an end point-of-interest, a time constraint, and the timed paths.

Constructing travel itineraries from tagged geo-temporal photographs

Method, system, and programs for providing one or more explanations. An inquiry is received via a communication platform where the inquiry is about how a set of entities are related. Information is retrieved from a knowledge storage in accordance with the set of entities and such information describes a plurality of entities and relationships existing among the plurality of entities. Based on such retrieved information, one or more explanations with respect to each relationship by which the set of entities are connected are generated. The one or more explanations are then transmitted as a response to the inquiry.

Display entity relationship

After receipt of a training and execution plan, a trainer operator is automatically trained based on specified training documents so as to generate a new trained operator for extracting information from documents. The new trained operator is a new version of the trainee operator. Both trainee operators are automatically retained for later use in extracting information from one or more unknown documents. After receipt of the training and execution plan, the new trained operator is automatically executed on one or more unknown documents so as to extract information from such one or more unknown documents.

Apparatus and methods for operator training in information extraction

Methods and apparatus for making recommendations of content items to users of computer systems include compiling a database relating a list of items and corresponding explanations; receiving from a user, through a computer user interface, a request for a recommendation; extracting from the database a preliminary list of items related to the request; identifying distances between the extracted items based on the explanation corresponding to each item; and identifying a subset of the preliminary list to form a recommendation list having a limited number of recommendation results with a desired balance of both high relevancy and high diversity relative to each other.

Diversifying recommendation results through explanation

An improved method for information retrieval in web query and recommendation systems, where items that are likely unfamiliar to the users of the system, but potentially relevant, are recommended. In a recommendation system having ratings by a plurality of users for a plurality of items, items are assigned to one or more data regions based on item attributes or user activity. Source regions are identified for each of the data regions. For a given user, data regions with which both the user and the user's social network are unfamiliar are identified. Within a given data region, the relevance of items to the user within such regions is evaluated using ratings provided by other users who have entered ratings similar to the user in source regions for the data region. Items receiving the highest relevance score are recommended to the user.

System and method for finding unexpected, but relevant content in an information retrieval system

Systems and methods are provided for determining items or people of potential interest to recommend to users in a computer-based network. Implied social networks may be determined based at least in part on obtained social behavior information. Items or people of potential interest to users may be determined based at least in part on implied social network information. Vocabulary taxonomies may be associated with, or used in determining, implied social networks.

Social behavior analysis and inferring social networks for a recommendation system

Growing community of inventors

Hoboken, NJ, United States of America

Cong Yu