{"id":12548,"date":"2024-06-07T09:51:29","date_gmt":"2024-06-07T01:51:29","guid":{"rendered":"https:\/\/www.1ai.net\/?p=12548"},"modified":"2024-06-07T09:51:29","modified_gmt":"2024-06-07T01:51:29","slug":"%e5%be%ae%e8%bd%af%e6%8e%a8%e5%87%ba%e9%a6%96%e4%b8%aa%e5%9f%ba%e4%ba%8e-ai-%e7%9a%84%e5%a4%a9%e6%b0%94%e9%a2%84%e6%8a%a5%e7%b3%bb%e7%bb%9f-aurora%ef%bc%8c%e8%bf%98%e8%83%bd%e9%a2%84%e6%b5%8b%e7%a9%ba","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/12548.html","title":{"rendered":"Microsoft launches Aurora, the first AI-based weather forecasting system that can also predict air pollution levels"},"content":{"rendered":"<p><a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%be%ae%e8%bd%af\" title=\"[View articles tagged with [Microsoft]]\" target=\"_blank\" >Microsoft<\/a>The first weather forecast was released on Monday.<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%a4%a7%e5%9e%8b%e5%a4%a7%e6%b0%94%e5%9f%ba%e7%a1%80%e6%a8%a1%e5%9e%8b\" title=\"[Sees articles with labels]\" target=\"_blank\" >Large atmospheric base model<\/a> <a href=\"https:\/\/www.1ai.net\/en\/tag\/aurora\" title=\"_Other Organiser\" target=\"_blank\" >Aurora<\/a>(Aurora), which enables high-precision forecasts of weather and atmospheric changes.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-12549\" title=\"22068ea2-740a-4cd6-b213-49c15221fd1a\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2024\/06\/22068ea2-740a-4cd6-b213-49c15221fd1a.png\" alt=\"22068ea2-740a-4cd6-b213-49c15221fd1a\" width=\"488\" height=\"1028\" \/><\/p>\n<p>Aurora is a 1.3 billion parameter base model for high-resolution forecasting of weather and atmospheric processes. Aurora is a flexible 3D Swin Transformer with a 3D perceptron-based encoder and decoder. While pre-training, Aurora is optimized to minimize loss on multiple heterogeneous datasets with different resolutions, variables, and pressure levels. The model is then fine-tuned in two stages: 1) short-lead time fine-tuning of pre-trained weights (2) long-lead time (rollout) fine-tuning using Low Rank Adaptation (LoRA). The fine-tuned model is then deployed to solve a variety of operational forecasting scenarios at different resolutions.<\/p>\n<p>According to Microsoft, Aurora is a basic model with 1.3 billion parameters, and also has an encoder and decoder based on 3D perception.<\/p>\n<p>Aurora is trained on over one million hours of diverse weather and climate data. Aurora leverages the strengths of fundamental modeling approaches to deliver operational forecasts for a wide range of atmospheric forecasting problems, including those with limited training data, heterogeneous variability, and extreme events.<\/p>\n<p>The team optimized the Aurora model in two stages during pre-training to minimize the loss of multiple heterogeneous data sets with different resolutions, variables, and pressure levels, so that it can adapt to various operational forecasting scenarios with different resolutions. In addition, the Aurora model has the following features, which are officially introduced as follows:<\/p>\n<ul class=\"list-paddingleft-2\">\n<li>\n<p data-vmark=\"da81\"><span class=\"accentTextColor\">Aurora can generate a 5-day global air pollution forecast and a 10-day high-resolution weather forecast in less than a minute.<\/span>.<\/p>\n<\/li>\n<li>\n<p data-vmark=\"36a5\">Aurora excels in air quality forecasting, providing forecasts comparable to CAMS (Copernicus Atmosphere Monitoring Service) at a smaller computational cost.<\/p>\n<\/li>\n<li>\n<p data-vmark=\"fea3\">Aurora makes weather forecasts with 0.1\u00b0 accuracy, enabling it to accurately predict extreme weather events, such as Storm Ciar\u00e1n.<\/p>\n<\/li>\n<li>\n<p data-vmark=\"57d0\">Aurora outperforms existing AI models and numerical weather forecast models in predicting extreme values.<\/p>\n<\/li>\n<li>\n<p data-vmark=\"8daf\">Aurora is able to perform well when pre-trained on diverse datasets, outperforming specialized deep learning models and state-of-the-art numerical weather forecasting models.<\/p>\n<\/li>\n<\/ul>\n<p>&nbsp;<\/p>","protected":false},"excerpt":{"rendered":"<p>Microsoft on Monday released Aurora, the first large-scale atmospheric base model for weather prediction, enabling high-precision forecasts of weather and atmospheric changes. Aurora is a 1.3 billion-parameter base model for high-resolution forecasting of weather and atmospheric processes.Aurora is a flexible 3D Swin Transformer with a 3D perceptron-based encoder and decoder. During pre-training, Aurora is optimized to minimize losses from multiple heterogeneous datasets with different resolutions, variables and pressure levels. The model is then fine-tuned in two stages \ud83d\ude41 1) Short lead time fine-tuning of the pre-trained weights (2) Long lead time (rollout) using low-rank adaptive (L<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[2961,2962,280],"collection":[],"class_list":["post-12548","post","type-post","status-publish","format-standard","hentry","category-news","tag-aurora","tag-2962","tag-280"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/12548","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=12548"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/12548\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=12548"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=12548"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=12548"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=12548"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}