Asia Pacific > Asia > AI model recommendation framework could slash machine learning selection emissions by 98% -study

AI model recommendation framework could slash machine learning selection emissions by 98% -study

Published 05:13 on May 26, 2026 / Last updated at 05:13 on May 26, 2026 / Americas (US & Canada), Asia Pacific (Asia), EMEA (Europe), Net Zero Transition (Industrial Decarbonisation, Power/Electrification), Voluntary (VCM Developments)

Carbon Pulse PremiumNet Zero Pulse

A new academic study has proposed a machine learning recommendation framework designed to minimise carbon emissions from AI model development by predicting the environmental impact of training runs before they occur, with the researchers claiming the approach could cut emissions from model selection by more than 98%.
A new academic study has proposed a machine learning recommendation framework designed to minimise carbon emissions from AI model development by predicting the environmental impact of training runs before they occur, with the researchers claiming the approach could cut emissions from model selection by more than 98%.


A subscription is required to read this content. Subscribe today to Carbon Pulse Premium or Net Zero Pulse to access our unrivalled news and intelligence, as well as other content including all job listings. Click here for details.

We offer a FREE TRIAL to each of our subscription services and it only takes a minute to register. If you already have a Carbon Pulse account, login here.

This page is intended to be viewed online and may not be printed.
As per our terms and conditions, the republication or redistribution of Carbon Pulse content can result in the suspension or termination of your subscription.