Microsoft Technology Licensing, LLC

Amazon Technologies, Inc.

Jiazhen Chen

Calvin Yue-Ren Kuo

Dharshan Rangegowda

Morgan Asher Brown

Anthony S Chavez

Anatoliy Panasyuk

Geoffrey Hugh Outhred

Jonathan I Turow

Hasan Alkhatib

Murali Krishna Sangubhatla

Ram Viswanathan

Jingwei Sun

Haiyang Liu

Hasan S Alkhatib

A provider network implements a machine learning deployment service for generating and deploying packages to implement machine learning at connected devices. The service may receive from a client an indication of an inference application, a machine learning framework to be used by the inference application, a machine learning model to be used by the inference application, and an edge device to run the inference application. The service may then generate a package based on the inference application, the machine learning framework, the machine learning model, and a hardware platform of the edge device. To generate the package, the service may optimize the model based on the hardware platform of the edge device and/or the machine learning framework. The service may then deploy the package to the edge device. The edge device then installs the inference application and performs actions based on inference data generated by the machine learning model.

Generating and deploying packages for machine learning at edge devices

Technology is described for providing machine learning (ML) models. A plurality of candidate ML models that are derived from a primary ML model may be generated in a service provider environment. The primary ML model may be associated with a set of parameters and a candidate ML model in the plurality of candidate ML models may be associated with a subset of the parameters associated with the primary ML model. The plurality of candidate ML models may be run against validation data to evaluate performance criteria for the candidate ML models. A performance representation of the candidate ML models with respect to performance results for the candidate ML models may be provided. An ML model may be selected from the performance representation based on the performance results for the candidate ML models.

Transformation of machine learning models for computing hubs

Architecture that facilitates the virtual specification of a connection between physical endpoints. A network can be defined as an abstract connectivity model expressed in terms of the connectivity intent, rather than any specific technology. The connectivity model is translated into configuration settings, policies, firewall rules, etc., to implement the connectivity intent based on available physical networks and devices capabilities. The connectivity model defines the connectivity semantics of the network and controls the communication between the physical nodes in the physical network. The resultant virtual network may be a virtual overlay that is independent of the physical layer. Alternatively, the virtual overlay can also include elements and abstracts of the physical network(s). Moreover, automatic network security rules (e.g., Internet Protocol security-IPSec) can be derived from the connectivity model of the network.

Model-based virtual networking

In a multi-tenant environment, machines across the Internet, belonging to a particular subscription are securely enrolled with the tenant's subscription. Authentication of the machines is delegated to each of the tenant's own on-premise authentication mechanism The trust relationship with the tenant's authentication service is used to validate the security token presented by the machine being authenticated. Once authenticated, the machine has authorization (e.g. SSL machine cert for identity, security token, etc.,) to access the subscription. Each tenant within the multi-tenant environment can provide its own level of authentication. The machine presents the security token to the multi-tenant environment for requests for resources (e.g. services/content) from a user. When a request is received from a machine to access a resource, the multi-tenant environment determines from the issued token whether or not the machine is authorized to access the requested resources.

Secure machine enrollment in multi-tenant subscription environment

Growing community of inventors

Bellevue, WA, United States of America

Jiazhen Chen