Amazon Technologies, Inc.

Microsoft Technology Licensing, LLC

Amason Technologies, Inc.

Gonzalo Alberto Ramos

Robert Yuji Haitani

Charles Shearer Dorner

Paul Barnhart Sayre, Iii

Janet Ellen Galore

Dominick Khanh Pham

Timothy Andrew Ong

Steven Mark Drucker

Stefan Vant

Jin A Suh

Eyal Ofek

Jenny Ann Blackburn

Matthew Brinkley

Cameron Charles

Babak Amir Parviz

Ting-Hsiang Hwang

Amy Bergsieker

Daniel Kelly

Elliot Chong

Christopher A Meek

Donald Allen Barnett

William R Hazlewood

Timothy Wantland

Shiqian Rachel Ng

Patrice Y Simard

Blaise Aguera Y Arcas

Eva Manolis

Michael Kroepfl

Julian R Walker

Amir Akbarzadeh

Johannes H Verwey

Nan-Chen Chen

Haekyu Park

Avi Bar-Zeev

Billy P Chen

Mary P Czerwinski

Wesley Scott Lauka

Bongshin Lee

Romualdo T Impas

Danyel Aharon Fisher

Hao Lu

Sumit Basu

Nathalie Henry Riche

Tim Phillip Wantland

David Maxwell Gedye

Ido Omer

Adam Fourney

Thomas Lund Dideriksen

Michael L S Dodge

Jonathan Robert Dughi

Samuel Thomas Williams

John-Michael Wiley

Sebastian Jorge Gunningham

Eric Paul Bennett

Peter G Sibley

Arpit Ajay Narechania

Aaron James Arterburn

Michael Chowning Byron

Basak Alper

Kourosh Kaghazian

Napol Rachatasumrit

A document re-finding system generates embeddings for concept clips provided by a user, the concept clips defining a concept for searching for content of interest to the user in a plurality of documents previously seen by the user. The re-finding system determines semantic relationships between the concept and document clips, related to respective document among the plurality of documents, based on the concept embedding and embeddings generated for the document clips. A graphical user interface depicting the semantic relationships is rendered to the user and is operable to enable re-finding a document, among the plurality of documents, having the content of interest to the user.

Document re-finding system

In example, a natural language input is processed to generate a formal query that is executed to retrieve data from a data store. An indication as to which subparts of the natural language input were used to generate the formal query may be presented to a user. The user may debug the query, such that associations between natural language input subparts and formal query subparts may be changed, added, or removed by the user. Further, a set of steps may be generated that provide an indication to the user as to how the formal query is executed, for example using sample data from the data store. The user may manipulate the sample data, thereby causing the presented steps to be updated accordingly. Once the user has made adjustments and has finished debugging the query, the revised formal query may be executed to retrieve updated data from the data store.

Natural language query processing and debugging

In examples, a schema augmentation system for exploratory research leverages intelligence from a machine learning model to augment such tasks by leveraging intelligence derived from machine learning capabilities. Augmenting tasks include schematization of content, such as information units and groupings of information units. Based on the schematization of such content, semantic proximities for information units are determined. The semantic proximities may be used to identify and present potentially relevant information units, for example to accelerate the exploratory research task at hand. As such, users engaged in consumption of heterogeneous content (e.g., across client applications and/or content sources), may receive machine-augmented support to find potential information units. To optimize machine training, user input may be received, such that the system may intelligently augment the user's exploratory research task based on the semantic coherence of the content processed from information units and associated user behavior.

Schema augmentation system for exploratory research

Classification predictions made by a concept classifier may be interactively visualized and explored in a user interface that displays visual representations of a plurality of data items in a star coordinate space spanned by a plurality of anchor concepts each mapping the data items onto respective finite real-valued scores. Positions of the visual representations of the data items in the star coordinate space are based on the scores for the plurality of anchor concepts, and may be updated responsive to a user manipulating the anchor concepts in the user interface, e.g., by moving or modifying definitions of anchor concepts, or by adding or deleting anchor concepts. The visual representations of the data items may reflect labels and/or classification predictions, and may be updated based on updated classification predictions following retraining the of the concept classifier based on added training data or new features.

Interactive semantic data exploration for error discovery

Described are systems and methods that enable users to virtually experience an environment at a destination location from their own user device, control their experience and even interact with others that are physically located within the environment. Likewise, the user's experience may even be improved with the inclusion of enhancements that are presented to the user as the user experiences the environment. Users may interact with the enhancements to obtain additional enhancements, interact with others physically located within the environment, interact with and direct a guide that is physically located within the environment, and receive near real-time video and audio that is transmitted from a guide device located within the environment.

Augmenting transmitted video data

Classification predictions made by a concept classifier may be interactively visualized and explored in a user interface that displays visual representations of a plurality of data items in a star coordinate space spanned by a plurality of anchor concepts each mapping the data items onto respective finite real-valued scores. Positions of the visual representations of the data items in the star coordinate space are based on the scores for the plurality of anchor concepts, and may be updated responsive to a user manipulating the anchor concepts in the user interface, e.g., by moving or modifying definitions of anchor concepts, or by adding or deleting anchor concepts. The visual representations may of the data items may reflect labels and/or classification predictions, and may be updated based on updated classification predictions following retraining the of the concept classifier based on added training data or new features.

Systems and methods are provided for generating social networking recommendations. A color preference of a first user may be determined from a color palette of a first image associated with the user and/or a color palette of an item associated with the user. Other users may be identified that have a similar color preference as the first user based at least in part on the determined color preference of the first user. Interactions between the first user and one or more other users having similar color preferences with respect to the first user may be facilitated. A social networking recommendation may be generated with respect to the one or more other users having similar color preferences with respect to the first user.

Color based social networking recommendations

Described are systems and methods that enable users to virtually experience an environment at a destination location from their own user device, control their experience and even interact with others that are physically located within the environment. Likewise, the user's experience may even be improved with the inclusion of enhancements that are presented to the user as the user experiences the environment. User's may interact with the enhancements to obtain additional enhancements, interact with others physically located within the environment, interact with and direct a guide that is physically located within the environment, and receive near real-time video and audio that is transmitted from a guide device located within the environment.

Synchronizing transmitted video data and enhancements

Features are disclosed for dynamically generating enhanced content showing an item to simulate a physics property of the item via an interaction that will initiate such interactive features. A physics model may be generated to enhance existing content. To generate the physics model, the system analyzes the content such as to detect a feature of the item (e.g., zipper on a clothing item or button on an electronic device item) shown in the content. Some aspects also consider the access device rendering the enhanced content to ensure the enhanced content is appropriately selected and formatted for presentation via the access device.

Physics modeling for interactive content

The display of 3D-representations involves navigation and interaction with users engaged with the display. Points of interest (POIs), as applied to areas of interest for a 3D-representation, are anchored to the 3D-representation on the display and are moveable in synchrony with the 3D-representation navigation. POIs are presented in which interactive actions are associated with the POIs. When a user selects one of the interactive actions associated with a POI, the selection can prompt the display of a new 3D-representation of an item related to the 3D-representation to which the POI is anchored, thereby improving access and analysis about information for related items.

Interactive points of interest for 3D-representations

Systems and methods are described that validate color information in metadata associated with an image. Color information associated with images, in metadata, may need to be validated because there may be mistakes in the metadata. Color names associated with colors may be based on human generated data or human surveys. Individual colors may be extracted from images, which may be associated with the color names. Furthermore, the color names may be retrieved from a data store via a fast index color search. The color information in the metadata may be validated against the color names.

Automatic color validation of image metadata

Systems and methods are provided for generating a perspective view of item images and/or determining collections of similar items. For example, a set and/or collection of item images may be presented in a perspective view. One or more options may be selected by a user to update the set and/or collection of items and/or images. An updated set and/or collection may be determined, for example, by clustering the items and/or comparing items with the base item. One or more dimensions and/or attributes may be used to cluster and/or graph the items to determine new collections of items.

Determining collections of similar items

Growing community of inventors

Bellevue, WA, United States of America

Gonzalo Alberto Ramos