Microsoft Technology Licensing, LLC

X Development LLC

Om Prakash Ravi

Hubert Van Hoof

Phillip Ellsworth Stahlfeld

Dmitry Denisov

Ananya Gupta

Arthur Robert Pope

Alina Mihaela Stanciu

Alexandre Igorevich Mineev

Stanislaw Wiktor Swierc

Xinyue Li

Supradha Sankaran

Matthew Robert Darsney

Aviva Cheryl Shwaid

Conor Edmund Kelly

Venkatasubramanian Jayaraman

Kshitij Naresh Nikhal

Methods, computer systems, and apparatus, including computer programs encoded on computer storage media, for training a classification neural network. The system generates, from a set of object-specific data, one or more meta-learning datasets for one or more respective initial training tasks. The system determines values for a set of meta parameters by performing meta-learning with a classification neural network on the one or more meta-learning datasets. The system obtains a set of labeled training examples for a characteristic-detection task. The system determines based at least on one of the values for the set of meta parameters and using the set of labeled training examples, target values for the network parameters for the classification neural network to perform the characteristic-detection task.

Meta-learning for detecting object anomaly from images

A technique is described herein for using computing technology to intelligently manage the consumption of a resource in a physical environment and/or controlling the physical environment in other ways. The technique maintains environment information that describes entities within the physical environment, together with the relationships among the entities. The technique leverages the environment information and collected sensor data to generate forecast data using one or more machine-trained models. The technique then leverages the environment information, sensor data, and forecast data to generate a control plan. The control plan provides a strategy for controlling the physical environment that satisfies a specified optimization objective. In one use case, the technique contributes to the efficient consumption of power provided by a distribution system by avoiding consumption of power in periods in which the distribution system is expected to experience high loads.

Controlling devices using a rich representation of their environment

A method for processing events comprising time series data may include inferring different schema associated with the events. The method may also include storing property definitions corresponding to the events. Each property definition may include a name and a data type. The method may also include storing schema definitions corresponding to the different schema that are inferred. Each schema definition may include a set of one or more properties. The method may also include updating at least one data structure for storing information about the events based on the different schema that are inferred.

Dynamic schema for storing events comprising time series data

Growing community of inventors

Sammamish, WA, United States of America

Om Prakash Ravi