Communications of the ACM
What Deep Learning Algorithms Can Teach Us About Snow
The DeepPrecip algorithm builds predictive models based on data from radar readings of snowfall.
The DeepPrecip neural network developed by Ph.D. student Fraser King and colleagues at the University of Waterloo in Canada models weather in order to measure annual snowfall and snowmelt in the context of climate change. The deep learning computational network builds predictive models from vast datasets derived from radar readings of snowfall.
"It's difficult to accurately measure snow," King says. "There have been other models but they have some limitations. Our new model is helping to move things forward."
There currently is little understanding of how climate change affects snowfall, so "any progress we can make in this area is beneficial," King says.
From University of Waterloo
View Full Article
Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA
No entries found