Microsoft Technology Licensing, LLC

Amazon Technologies, Inc.

Erik Bergman

Jeremy S Knudsen

Gaurang Prajapati

Holger Kuehnle

Michael John Bray

Sandy Wong

Yu-ting Kuo

Evan Ming Lew

Michael Brian Palmer

Britt Hillary Hansing

Christer Garbis

Katherine M Ryan

Robb Dean Anderson

Andrew Salituri

Edward Michael Leonard

Anh Q Ly

Brian Kropf

Tanya Marie Test

Jeffrey Pratt

James Rosenblum

Alper Aras

Cindy Wong

Rich Zarick

Jera Vincelli

Avrum Stephen Laurie

Systems and methods are described herein for providing predictive user interface elements. A computing system may train a machine-learning model to identify issues that are likely being experienced by users contacting a customer service system based at least in part on historical user account data of a plurality of user accounts. When a request for assistance is received, user account data corresponding to the request may be obtained and provided to the model to identify issues likely experienced by a user. A number of graphical user interface (GUI) elements (e.g., 'match cards'), each corresponding to one of the identified issues, may be generated, ranked, and presented in accordance with the ranking. Each GUI element may be selectable. Upon selection additional data likely to be pertinent to the selected issue may be presented alleviating a need to search for this data as would be the case in conventional systems.

Techniques for providing predictive interface elements

One or more techniques and/or systems are provided for query suggestion formulation for multi-source queries, for ranking multi-source search results, and/or for exposing contextual launch functionality through multi-source search results of a multi-source search interface. In an example, a query suggestion may be provided for a partial search query based upon an implied content source that corresponds to a search intent of a user (e.g., an intent to view videos, as opposed to images, of houses). In another example, relevancy ranks may be assigned to content sources based upon a content type preference of a user, and search results may be provided from content sources having relevancy rankings above a relevancy threshold. In another example, links to applications and/or execution contexts may be embedded within search results so that applications may be launched into contextually aware states from the search results.

Multi-source search

One or more techniques and/or systems are provided for indexing referenced content and/or for deep content searching. In an example, parent content (e.g., an instant message from a friend about a celebrity) may be evaluated to identify a reference (e.g., a URL) to referenced content hosted by a content source (e.g., a photo shared through a photo sharing service). The referenced content may be acquired from the content source, and may be evaluated to identify a search term that is descriptive of the referenced content (e.g., a name of the celebrity in the photo). The parent content and the referenced content may be indexed into a search index using the search term. In an example, responsive to a search query corresponding to the parent content and/or the search term, the parent content and/or the referenced content may be provided as search results.

Referenced content indexing

Display screen with graphical user interface

Technologies are described herein for providing a context pane with semantic zoom. A content interface including a content pane and a context pane is provided. The content pane may be adapted to provide content, and the context pane may be adapted to display a richness of contextual data. A selectable interface element is also provided in the content interface. When the interface element is selected, the context pane may be transformed from displaying the richness of contextual data into displaying another richness of contextual data.

Context pane with semantic zoom

Growing community of inventors

Seattle, WA, United States of America

Erik Bergman