Sign In

Communications of the ACM

ACM TechNews

Microsoft Predicts Weather for Individual Farms

View as: Print Mobile App Share:
A sensor collecting local climate data for the DeepMC framework, deployed at the Nelson Farm in the Palouse region of southeastern Washington state.

Microsoft Research senior research scientist Peeyush Kumar says DeepMC is more accurate than any comparable model.

Credit: Maryatt Photography

Microsoft’s DeepMC framework uses machine learning and artificial intelligence (AI) to predict local weather.

DeepMC combines data from on-site sensors with standard local weather forecast data accessed through application programming interfaces from sources including the U.S. National Oceanic and Atmospheric Administration and the National Weather Service.

Microsoft Research's Peeyush Kumar said a "fusion mechanism" integrates data from on-site sensors with standard forecast data, which the framework then provides to an AI trained to find errors between local forecasts and micro-climate weather conditions.

The system predicts each weather parameter based on historical forecast and local sensor data, and applies decomposition to determine short- and long-term trends and patterns in weather data.

From IEEE Spectrum
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found