Copula-based synthetic data generation for machine learning emulators in weather and climate: application to a simple radiation model

12/16/2020
by   David Meyer, et al.
37

Can we improve machine learning (ML) emulators with synthetic data? The use of real data for training ML models is often the cause of major limitations. For example, real data may be (a) only representative of a subset of situations and domains, (b) expensive to source, (c) limited to specific individuals due to licensing restrictions. Although the use of synthetic data is becoming increasingly popular in computer vision, the training of ML emulators in weather and climate still relies on the use of real data datasets. Here we investigate whether the use of copula-based synthetically-augmented datasets improves the prediction of ML emulators for estimating the downwelling longwave radiation. Results show that bulk errors are cut by up to 75 bias error (from 0.08 to -0.02 W m^-2 and by up to 62 W m^-2) for the mean absolute error, thus showing potential for improving the generalization of future ML emulators.

READ FULL TEXT

page 4

page 6

research
07/19/2021

Introducing a Family of Synthetic Datasets for Research on Bias in Machine Learning

A significant impediment to progress in research on bias in machine lear...
research
04/28/2023

Temporal Subsampling Diminishes Small Spatial Scales in Recurrent Neural Network Emulators of Geophysical Turbulence

The immense computational cost of traditional numerical weather and clim...
research
12/07/2021

Estimating Quality of Transmission in a Live Production Network using Machine Learning

We demonstrate QoT estimation in a live network utilizing neural network...
research
11/29/2022

Synthetic data enable experiments in atomistic machine learning

Machine-learning models are increasingly used to predict properties of a...
research
11/29/2021

ClimART: A Benchmark Dataset for Emulating Atmospheric Radiative Transfer in Weather and Climate Models

Numerical simulations of Earth's weather and climate require substantial...
research
06/14/2020

Sub-Seasonal Climate Forecasting via Machine Learning: Challenges, Analysis, and Advances

Sub-seasonal climate forecasting (SSF) focuses on predicting key climate...
research
09/14/2021

Conditional Synthetic Data Generation for Robust Machine Learning Applications with Limited Pandemic Data

Background: At the onset of a pandemic, such as COVID-19, data with prop...

Please sign up or login with your details

Forgot password? Click here to reset