Microsoft Technology Licensing, LLC

Pallavi Baral

Bhuvan Malladihalli Shashidhara

Emily Nicole Wilson

Brett Kenneth Dodds

Daniel Sebastian Berger

Avnish Kumar Chhabra

Michael Charles Stephenson

John A Strange

Ramakoti R Bhimanadhuni

Huaming Huang

Peter Hanpeng Jiang

Jasmine Grace Schlichting

Sisi Xiong

Yilan Zhang

Farah Farzana

Michael Wayne Stephenson

Prateek Punj

Sebastien Nathan R Levy

Josh Charles Moore

Ahu Oral

Arhatha Bramhanand

Juan Arturo Herrera Ortiz

Abhishek Kumar

Christina Margaret Tobias

Hanyi Xu

Charlotte Gauchet

Mayank Meghwanshi

Juan-Arturo Herrera Ortiz

The disclosure herein describes training and using an uncorrectable error (UE) state prediction model based on telemetry error data. Sets of UE state labels and non-UE state labels are generated from a first set of collected telemetry data, wherein the UE state labels each reference a UE and telemetry data of an interval prior to the referenced UE. Statistical features are extracted from telemetry data of the sets of UE state labels and non-UE state labels, and the extracted statistical features are used to train a UE state prediction model. A second set of collected telemetry data is obtained, and a UE event is predicted based on the second set of collected telemetry data using the trained UE state prediction model. A preventative operation is performed on a memory page of the system based on the predicted UE event, whereby the predicted UE event is prevented from occurring.

Training and using a memory failure prediction model

The present concepts may automate and optimize deployment to a cloud computing fleet. Artificial intelligence and/or optimization algorithms may be used to find optimal deployment parameters (e.g., deployment order of computers in the fleet) that minimize deployment time and minimize deployment risk. For example, machine-learning prediction models may be used to generate a shortest path graph problem models a deployment planning problem. Then, optimization algorithms may be used to efficiently find approximations of Pareto-optimal solutions to the shortest path graph problem. Depending on the preferred level of time and risk, one of the solutions may be used to run the deployment.

Smart deployment using graph optimization

Growing community of inventors

Redmond, WA, United States of America

Pallavi Baral