June 28, 2022

The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink

Key Points

Key points are not available for this paper at this time.

Abstract

Machine learning (ML) workloads have rapidly grown, raising concerns about their carbon footprint. We show four best practices to reduce ML training energy and carbon dioxide emissions. If the whole ML field adopts best practices, we predict that by 2030, total carbon emissions from training will decline.

Demander à l'IA

Bookmark

Cite This Study

Patterson et al. (Tue,) studied this question.

synapsesocial.com/papers/69f25a11d4c794b733abc38e https://doi.org/https://doi.org/10.1109/mc.2022.3148714

Also Consider

Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context:

1Towards the Systematic Reporting of the Energy and Carbon Footprints of\n Machine Learning2020 · 308 citations
2GLaM: Efficient Scaling of Language Models with Mixture-of-Experts2021 · 168 citations
3Searching for Fast Model Families on Datacenter Accelerators2021 · 31 citations
4Deep Learning's Diminishing Returns: The Cost of Improvement is Becoming Unsustainable2021 · 136 citations

Demander à l'IA

Bookmark