Microsoft Technology Licensing, LLC

Robert J Ragno

Eric David Brill

Ewa Dominowska

Matthew R Richardson

Vivek Jawahir Jhaveri

Robert L Rounthwaite

Li Erran Li

Bo Thiesson

Kenneth Ward Church

Silviu-Petru Cucerzan

Hugo Zaragoza

Christopher J Burges

Felipe Luis Naranjo

Jody D Biggs

Adwait Ratnaparkhi

Noah A Smith

Boris Gorodnitsky

An approach is provided for associating structured information as feature-value pairs with advertisements at an advertisement system that provides advertisements for presentation with primary content. Feature-value pairs corresponding with advertisements describe features of the products or services associated with the advertisements. The feature-value pairs may be used by the advertisement system in a number of different manners, including, among other things: using the feature-value pairs during selection of advertisements for presentation with primary content: allowing users to sort and/or filter advertisements during presentation with primary content based on the associated feature-value pairs; exposing additional information with the advertisements; facilitating fraud detection for the advertisement system; and enabling a hybrid pricing model.

Feature-value attachment, reranking and filtering for advertisements

A system for guiding a search for information is presented. The system comprises a user interface that accepts a phrase and receives at least one suggestion based at least in part on the phrase. The system also includes a phrase suggestion engine that matches the phrase with the at least one suggestion. Methods of using the system are also provided.

Dynamic client interaction for search

Methods and systems in search marketing which permit advertisers to provide advertisements to a search engine, as well as bids associated with the advertisements themselves, are provided. Advertisements are selected for presentation when they are determined to be relevant and payment for presentation is due only upon the conditions of the bid being met. If desired, advertisers may be permitted to provide, in addition to an advertisement, one or more destination pages associated with selection of the advertisement. In this case, relevance can be determined not only on the basis of the content of the advertisement itself, but also on the content of the destination page(s).

Selecting directly bid upon advertisements for display

Systems, methods, and computer-readable media for navigating search results are provided. In response to receiving a search query, at least one individual search result is presented in an overlay window, the overlay window being configured to overlay at least a portion of a document viewing window. Navigation between the overlay window and the document viewing window may be performed in response to receiving input of at least one pre-determined command. Content associated with a selected individual search result (or other document identifier) may be presented in the document viewing window and, substantially simultaneously, the overlay window may be hidden from view. Subsequently, in response to receiving at least one pre-determined command, the overlay window may be re-presented such that it again overlays at least a portion of the document viewing window.

Efficient navigation of search results

A unique system and method that facilitates storing virtually all web content that has been viewed or visited by a user on the user's local cache for access anytime by the user. The system and method involve monitoring the user's web activity. Pages that have been viewed or visited can be saved at timed intervals to a local cache. If pages have been updated, those changes can be saved. Thus, multiple versions of a web page can be accessed and the user can select which version to view. The updated page can be visually modified to distinguish between older and newer content. If offline, the user can still view desired pages by accessing them from the local cache. By saving the user's web activity, the user can resume a previous browsing session. A user can also choose to receive periodic notifications when changes to previously viewed pages occur.

Unlimited history store for navigational web applications

A predictive travel log system using one or more ranking schemes to predict the location that a user would seek to navigate to through back and forward navigation in a web browser is provided. The navigation functionality can be exposed through traditional back and forward buttons with drop down travel log menus found in a browser user interface.

Browsing web content using predictive navigation links

A unique ranking system and method that facilitates improving the ranking and ordering of objects to further enhance the quality, accuracy, and delivery of search results in response to a search query. The system and method involve monitoring and tracking an object in terms of the number of times it's been accessed and optionally by whom, when, for how long, and an access rate. The user's interaction with the object can be tracked as well. By tracking the objects, a popularity measure can be determined. Popularity based rankings can be computed based on the popularity measure or some function thereof. The popularity measure can be affected by the access time, who accessed it, access duration or the user's interaction with the object upon access. The popularity based rankings can be utilized by a search component to improve the quality and retrieval of search results.

Using popularity data for ranking

The present invention leverages relevance data to provide enhanced search query results based on relevancy to a specific entity via an entity-specific tunable search. This allows an entity to retrieve information that is of more value to that entity, in a faster and more efficient manner. The entity itself can be an individual user, a grouping of users, and/or an enterprise and the like. In one instance of the present invention, entity-specific relevance information is determined via employment of similarity of the entity to another entity or group of entities. Interest levels and/or satisfaction levels of similar entities can also be utilized along with similarity information to facilitate in deriving the relevance information.

Entity-specific tuned searching

The subject disclosure pertains to systems and methods for facilitating training of machine learning systems utilizing pairwise training. The number of computations required during pairwise training is reduced by grouping the computations. First, a score is generated for each retrieved data item. During processing of the data item pairs, the scores of the data items in the pair are retrieved and used to generate a gradient for each data item. Once all of the pairs have been processed, the gradients for each data item are aggregated and the aggregated gradients are used to update the machine learning system.

Efficiency of training for ranking systems based on pairwise training with aggregated gradients

A method, media and data structure for extraction of key-phrase-content pairs are provided. Query logs having user queries and corresponding web results are processed to generate a content-to-key phrase index that includes key phrases. The query logs are filtered and key phrases are derived from the queries included in the query logs. The key phrases are aggregated based on a repeated content pages in the web results. User behavior and metadata is used to assign scores for each key phrase. The scores are utilized to rank the key phrases. Among other things, the key phrases can be used for advertisement to content matching and content summarization.

Key phrase extraction from query logs

Systems and methods are described that allow programmatic access to search engine results and query logs in a structured form. The search results can be retrieved from the search engine in an intermediary form that contains the information that is in the HTML pages provided to web browsers (potentially with additional information). This intermediary form can then be broken down on the client machine, using local resources, to assemble the structured objects. The library also provides for caching of the search results. This can be provided both on the local machine and on a remote database. When the results for a query exist in the caches, they can be retrieved from such location instead of querying the search engine. Documents and/or web pages can also be cached. The library can also be directed to operate only from the cache, effectively exposing a local data set instead of the remote search engine.

Application programming interface for text mining and search

The present invention leverages query-related information based on a query and/or a search intention to provide a systematic means to facilitate a user in locating desired information despite lacking exact search parameters. This allows users to find information without first formulating an optimum search query. The query graph provides a navigable, graphical notion of the query-related information via nodes representative of the query-related information and edges representative of the associations between the information. In one instance of the present invention, the query graph is derived from an information source such as a query log, a query list, and/or a search engine and the like. Additional instances of the present invention utilize visual and/or audible indicators employed with the query graph to facilitate in relaying the query-related information to the user, including, but not limited to, overlays, icons, colors, and dimensional variances and the like.

Query graphs indicating related queries

Dynamically updated search results based upon continuously-evolving search query that is based at least in part upon phrase suggestion, search engine uses previous result sets performing additional search tasks

Growing community of inventors

Kirkland, WA, United States of America

Robert J Ragno