Saved in:
Bibliographic Details
Main Authors: Asan, Busra, Akgül, Abdullah, Unal, Alper, Kandemir, Melih, Unal, Gozde
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.16612
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917630048206848
author Asan, Busra
Akgül, Abdullah
Unal, Alper
Kandemir, Melih
Unal, Gozde
author_facet Asan, Busra
Akgül, Abdullah
Unal, Alper
Kandemir, Melih
Unal, Gozde
contents Seasonal forecasting is a crucial task when it comes to detecting the extreme heat and colds that occur due to climate change. Confidence in the predictions should be reliable since a small increase in the temperatures in a year has a big impact on the world. Calibration of the neural networks provides a way to ensure our confidence in the predictions. However, calibrating regression models is an under-researched topic, especially in forecasters. We calibrate a UNet++ based architecture, which was shown to outperform physics-based models in temperature anomalies. We show that with a slight trade-off between prediction error and calibration error, it is possible to get more reliable and sharper forecasts. We believe that calibration should be an important part of safety-critical machine learning applications such as weather forecasters.
format Preprint
id arxiv_https___arxiv_org_abs_2403_16612
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Calibrating Bayesian UNet++ for Sub-Seasonal Forecasting
Asan, Busra
Akgül, Abdullah
Unal, Alper
Kandemir, Melih
Unal, Gozde
Machine Learning
Computer Vision and Pattern Recognition
Seasonal forecasting is a crucial task when it comes to detecting the extreme heat and colds that occur due to climate change. Confidence in the predictions should be reliable since a small increase in the temperatures in a year has a big impact on the world. Calibration of the neural networks provides a way to ensure our confidence in the predictions. However, calibrating regression models is an under-researched topic, especially in forecasters. We calibrate a UNet++ based architecture, which was shown to outperform physics-based models in temperature anomalies. We show that with a slight trade-off between prediction error and calibration error, it is possible to get more reliable and sharper forecasts. We believe that calibration should be an important part of safety-critical machine learning applications such as weather forecasters.
title Calibrating Bayesian UNet++ for Sub-Seasonal Forecasting
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.16612