Microsoft Technology Licensing, LLC

David B Perry

Richard Jennings Duncan

Leroy B Keely

Anshul Rawat

Elizabeth Picchietti Salowitz

Mohammed Samji

Raju Jain

Robert Paul St Pierre

Chris Devaraj

Steven P Dodge

David Andrew Matthews

Aaron Naoyoshi Sheung Yan Woo

John Heinrich Lueders

Sanjeev Kewal Verma

Rick Duncan

Thomas H Alphin

Steven Nabil Bathiche

Michael John Novak

Neal Robert Christiansen

Gavin M Gear

Jonathan David Friedman

Eric B Fox

Tyler Kien Beam

Shawna Julie Davis

Vivek Pradeep

Emily K Rimas-Ribikauskas

Zoltan C Szilagyi

Eric Chris Wolfgang Sommerlade

Ping Xie

Weerapan Wilairat

Craig Ashley Barkhouse

Sharath Viswanathan

Hugh Edward McLoone, Jr

Kyle Thomas Beck

Ernest L Pennington, Ii

Richard L Spencer, Ii

John Michael Calandrino

Sally Liu

Nathan James Luquetta-fish

Carlos A C Pessoa

Srinivasan Malayala

Kenneth Sean Gilmour

Sripresanna Sripresanna

Jose Antonio Lara Silva

Tom Alphin

Machine learning techniques are leveraged to provide personalized assistance on a computing device. In some configurations a timeline of a user's interactions with the computing device is generated. For example, screenshots and audio streams may be saved as entries in the timeline. Context—the state of the computing device when the entry is created, such as which documents and websites are open—is also stored. Entries in the timeline are processed by a model to generate embedding vectors. The timeline may be searched by finding the embedding vector that is closest to an embedding vector derived from a search query. The user may select a query result, causing the associated context to be restored. For example, if the query is 'show me all documents related to my upcoming trip to Japan', the query result may open documents and websites that were open when booking a flight to Japan.

User activity history experiences powered by a machine learning model

Determination logic for a cloud content states framework is described herein. In implementations, a framework is configured to perform operations for management of cloud content files at the computing device on behalf of cloud storage providers. Determination logic for the framework is utilized to make state determinations in dependence upon a defined set of file sync properties. The framework operates to make assessments of the file sync properties and set states for files in accordance with the assessments. In implementations, file sync properties are related to corresponding states through a mapping data structure that maps file sync property values to different states and enables look-ups of states based on property values. The framework utilizes and references the mapping data structure to match file sync properties identified for cloud content files to corresponding states and then set states for the cloud content files accordingly.

Cloud content states determination logic

A contextual file manager is described herein. Contextual information regarding files and folders of a file system of an operating system is collected. The contextual information may be collected from a variety of different sources, such as contents of the file or folder, an application associated with the file, a service provider, or a cloud storage service. The contextual information is then exposed in a file system interface of the operating system along with representations of the files and folders. In some cases, the contextual information is exposed in response to detection of an intuitive user action to a representation of one of the files or folders via the file system interface.

Contextual file manager

A cloud content states framework is described herein. In implementations, a framework is configured to perform operations for management of cloud content files at the computing device on behalf of cloud storage providers. The framework is operable to manage states for cloud content files accessible from multiple cloud storage services. The framework also manages and outputs state indicators with file system representations of the cloud content files. The states include at least a set of primary states indicative of synchronization status controlled by the framework. The states may also include custom states associated with cloud storage services that are defined via an interface exposed by the framework. In operation, states associated with cloud content files are recognized in connection with rendering a user interface with cloud content file representations. Then, state indicators that reflect the recognized states are exposed along with cloud content file representations in the user interface.

Cloud content states framework

A method for processing an image such as a computer wallpaper identifies a characteristic color representative of the image. Image pixels with similar colors are separated into groups, and the average value of the R,G,B color components in each group is determined, after filtering out pixels with R,G,B values representing white, black, or grey. The group with the maximum difference between the highest average color component value and the lowest average color component value is identified as the characteristic color. Groups representing a number of pixels less than a certain percentage of all of the pixels are not considered. The characteristic color can be used in other displayed images at an intensity &#x3b1; determined by setting maximum and minimum values of &#x3b1;, with &#x3b1; being the lesser of &#x3b1;and &#x3b1;plus the average color span of all pixels in the image.

Automatic adjustment of computer interface colors using image processing

Embodiments described herein are directed to simultaneously presenting a virtual touchpad and a graphical user interface of an operating system on the same display. A user can interact with the OS using the virtual touchpad, and user touches are read by a touch-screen portion of the display assigned to the touchpad. Touch input software converts touch packets from user touches into data packets relevant to the screen and display resolution of either the display or a portion of the display presenting the OS. Gesture-recognition software applies rules to the converted packets to determine which actions the user meant by the touches. And an application controlling a mouse cursor manipulates the mouse cursor according to the mouse actions.

Virtual touchpad

Methods, systems, and computer-storage media for customizing the appearance of a desktop background are provided. The desktop is the bottom surface drawn behind the application windows, gadgets, a task bar, and other user interface elements displayed on a computing screen. Embodiments of the present invention allow a user to customize the background by dragging an image representation from a desktop control panel onto a portion of the desktop. Once dropped onto the desktop, the image associated with the representation is used to generate the desktop background. When the desktop is extended to different monitors, the portion of desktop on each monitor may be customized separately.

Directly assigning desktop backgrounds

Methods are provided for interpreting a touch in relation to touch targets displayed on a touch-screen display, the touch targets associated with an application. A touch is detected at a first touch-screen location. The application is queried to determined a first touch target located within a predetermined touch region of the first touch-screen location. The application is then queries to determine whether a second touch target is located within the predetermined touch region. The touch is them disambiguated to determine whether the touch was intended for the first touch target or the second touch target.

Interpreting ambiguous inputs on a touch-screen

A method for interpreting ambiguous click events in relation to click targets on a touch-screen display disposed at respective click target locations includes detecting a touch at a touch location and determining whether the touch location corresponds to a click target location. The method also includes searching for nearby click target locations in a predetermined click region and, upon locating a first click target location within the predetermined click region, associating the touch with the first click target location to identify a click event. A computing device for facilitating accurate touch input targeting with respect to a touch-screen display includes a display component, a touch detection component, a targeting component that associates a touch with a click target, and an event detection component that associates the touch with one of a right click event, a left click event, or a drag event.

A system and method for training a user to perform a flick gesture is described. A flick gesture is a simple gesture that may be easily detected and is characterized by minimal interference with other applications or gestures.

Gesture training

A processing device may include a display screen and one or more buttons located near the display screen. Pressing one of the one or more buttons may cause the processing device to perform a first function and pressing and holding the one of the one or more buttons may cause the processing device to perform a second function. When the one of the one or more buttons is pressed, the processing device may provide a first feedback describing the first function. When the one of the one or more buttons is pressed and held for at least a predetermined amount of time, the processing device may provide a second feedback describing the second function. In some implementations, a time indicator may be displayed indicating an amount of passing time until the one of the one or more buttons is pressed and held for the predetermined amount of time.

Button discoverability

Data associated with pixels obscured by annotations in an annotated image is stored using an annotation mask enabling annotation-aware applications to reconstruct the original image, while annotation-unaware applications can display the annotated image with optimized increase in extra data to be stored. The annotation mask, same size as the image, is initialized to a monochrome default and the color depth information for obscured pixels saved within the annotation mask. Because a majority area of the annotation mask is monochrome, it compresses well and the obscured image content can be saved in a side channel, metadata header, or a separate file. During reconstruction, changes to annotated image such as rotation, size change, and the like, may be detected and accounted for when obscured image content is carried over to the annotated image for reconstruction.

Sharing editable ink annotated images with annotation-unaware applications

Growing community of inventors

Bothell, WA, United States of America

David B Perry