Google Inc.

Yahoo! Inc.

Excalibur Ip, LLC

Vladimir Ofitserov

Navneet Panda

Marc Alexander Najork

Mike Bendersky

Hang Su

Siva Gurumurthy

Kaihua Zhu

Spurthi Amba Hombaiah

Shanmugasundaram Ravikumar

Amit Kumar

Priyank Shanker Garg

Oleg Kislyuk

Igor Krivokon

Conrad Newman

Amit Sasturkar

Rajat Ahuja

Implementations relate to training a model that can be used to process values for defined features, where the values are specific to a user account, to generate a predicted user measure that reflects both popularity and quality of the user account. The model is trained based on losses that are each generated as a function of both a corresponding generated popularity measure and a corresponding generated quality measure of a corresponding training instance. Accordingly, the model can be trained to generate, based on values for a given user account, a single measure that reflects both quality and popularity of the given user account. Implementations are additionally or alternatively directed to utilizing such predicted user measures to restrict provisioning of content items that are from user accounts having respective predicted user measures that fail to satisfy a threshold.

Training and/or utilizing a model for predicting measures reflecting both quality and popularity of content

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for receiving a search query from a client device; receiving search result data identifying a first plurality of search result resources and respective initial scores for each of the first plurality of search result resources; identifying a respective group of resources to which each of the search result resources belongs; determining a respective group-based modification factor for each group of resources; and adjusting the initial score for each of the search result resources based at least in part on the group-specific modification factor for the group of resources to which the search result resource belongs to generate a respective second score for each of the search result resources.

Ranking search results

Disclosed herein is use of a preview of content from a target document, as provided by a content preview source such as a Really Simple Syndication (RSS) feed, by a search engine. The content preview source includes the preview of the target document's content and a reference, e.g., a Universal Resource Locator (URL) or other link. A content preview document is generated using data extracted from the content preview source. The content preview document is made available in a searchable index used by a search engine to respond to a search query. A fetch operation is scheduled to fetch the target document using the reference provided in the content preview source. Once fetched, the data extracted from the content preview source can be associated with the target document, and can be used in presenting the target document in search results.

Search engine recency using content preview

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for determining site quality scores. In one aspect, a method includes obtaining a plurality of measurements of durations of user visits to resources included in a particular site; and determining a site quality score for the particular site based at least in part on the plurality of measurements, wherein determining the site quality score for the particular site comprises computing a statistical measure from the plurality of measurements.

Scoring site quality

Methods, systems, and apparatus for computing quality scores based on neighbor features. In one aspect, a method includes obtaining a quality model that was trained using a set of training entities; identifying a set of candidate entities that are different from each of the training entities; for each candidate entity: obtaining a first quality score for the candidate entity; obtaining one or more neighbor features for neighbor entities of the candidate entity, where each neighbor entity of the candidate entity is linked to the candidate entity; obtaining one or more entity specific feature values for the candidate entity, where each entity specific feature value is determined independent of the neighbor entities of the candidate entity; and determining a second quality score for the candidate entity using the quality model, the second quality score being computed based on the first quality score, the neighbor features, and the entity specific feature values.

Evaluating quality based on neighbor features

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for ranking search results. One of the methods includes determining, for each of a plurality of groups of resources, a respective count of independent incoming links to resources in the group; determining, for each of the plurality of groups of resources, a respective count of reference queries; determining, for each of the plurality of groups of resources, a respective group-specific modification factor, wherein the group-specific modification factor for each group is based on the count of independent links and the count of reference queries for the group; and associating, with each of the plurality of groups of resources, the respective group-specific modification factor for the group, wherein the respective group-specific modification for the group modifies initial scores generated for resources in the group in response to received search queries.

Disclosed herein is a technique for providing an interface that allows a user to navigate backwards through linked webpages. Initially, a request to display inlinks of linking webpages that contain a link to a particular webpage is received. In response to the request, a new page that contains a set of inlinks that correspond to a set of linking webpages that each contain a link to the particular webpage is provided. Each of the inlinks may be associated with a particular clickable item. An indication of a selection of a clickable item, associated with a particular inlink is received. In response, a second new page which contains a second set of inlinks that correspond to a second set of linking webpages that each contain a link to the webpage that corresponds to the particular inlink is provided. Some of the displayed inlinks may correspond to webpages that redirect to the particular webpage.

Providing an interface to browse links or redirects to a particular webpage

Techniques are disclosed for detecting web pages with duplicate content. In one embodiment, a set of shingles is computed for each page of a group of pages. An aggregate set of shingles is determined based on the sets of shingles computed for the group of pages. A first subset from the aggregate set of shingles is determined by selecting, from the aggregate set, shingles whose frequencies in the aggregate set exceed a specified threshold. A modified set of shingles is generated for each page of the group of pages by removing, from the set of shingles for that page, any shingle included in the first subset. One or more duplicate pages in the group of pages are determined based at least in part on the modified sets of shingles generated for the group of pages.

Growing community of inventors

Foster City, CA, United States of America

Vladimir Ofitserov