[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-126542-en":3,"doc-seo-126542-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},126542,13056712833777,"Logic","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","Counting Carbon - A Survey of Factors Influencing the Emissions of Machine Learning","Machine learning model training consumes energy, and that energy generation carries greenhouse-gas emissions whose magnitude depends on both the energy quantity used and its source. Existing studies often cover only a limited set of models and tasks, leaving the field’s emission drivers insufficiently characterized. This work surveys 95 ML models over time across natural language processing and computer vision. Energy sources, CO2 emissions, temporal evolution, and relationships to model performance are analyzed, concluding with a proposal for a centralized reporting repository.","arXiv :2302 .08476v1 [ cs .LG] 16 Feb 2023  \nCounting Carbon: A Survey of Factors Influencing the Emissions of Machine Learning  \nALEXANDRA SASHA LUCCIONI, Hugging Face, Montreal, Canada  \nALEX HERNANDEZ-GARCIA, Mila, Université de Montréal, Montreal, Canada  \nMachine learning (ML) requires using energy to carry out computations during the model training process. The generation of this energy comes with an environmental cost in terms of greenhouse gas emissions, depending on quantity used and the energy source. Existing research on the environmental impacts of ML has been limited to analyses covering a small number of models and does not adequately represent the diversity of ML models and tasks. In the current study, we present a survey of the carbon emissions of 95 ML models across time and different tasks in natural language processing and computer vision. We analyze them in terms of the energy sources used, the amount of CO2 emissions produced, how these emissions evolve across time and how they relate to model performance. We conclude with a discussion regarding the carbon footprint of our field and propose the creation of a centralized repository for reporting and tracking these emissions.  \n1 INTRODUCTION  \nIn recent years, machine learning (ML) models have achieved high performance in a multitude of tasks such as image classification, machine translation, and object detection. However, this progress also comes with a cost in terms of energy, since developing and deploying ML models requires access to computational resources such as Graphical Processing Units (GPUs) and therefore energy to power them. In turn, producing this energy comes with a cost to the environment, given that energy generation often entails the emission of greenhouse gases (GHG) such as carbon dioxide (CO2) [40] . On a global scale, electricity generation represents over a quarter of the global GHG emissions, adding up to 33.1 gigatonnes of CO2 in 2019 [24] . Recent estimates put the contribution of the information and communications technology (ICT) sector – which includes the data centers, devices and networks used for training and deploying ML models – at 2–6 % of global GHG emissions, although the exact number is still debated [25, 32, 36] . In fact, there is limited information about the overall energy consumption and carbon footprint of our field, how it is evolving, and how it correlates with performance on different tasks.  \nThe goal of the current paper is to analyze the main factors influencing the carbon emissions of our field, to study the evolution across time, and to contribute towards a better understanding of the carbon emissions generated by ML models trained on different tasks and as a function of their performance. As such, our research aims to answer the following research questions:  \n(1) What are the main sources of energy used for training ML models?  \n(2) What is the order of magnitude of CO2 emissions produced by training ML models?  \n(3) How do the CO2 emissions produced by training ML models evolve over time?  \n(4) Does more energy and CO2 lead to better model performance?  \nAuthors’ addresses: Alexandra Sasha Luccioni, Hugging Face, Montreal, Canada, [sasha.luccioni@huggingface.co](sasha.luccioni@huggingface.co); Alex Hernandez-Garcia, Mila, Université de Montréal, Montreal, Canada, [alex.hernandez-garcia@mila.quebec](alex.hernandez-garcia@mila.quebec).  \n2023. Manuscript pending review  \n2 Alexandra Sasha Luccioni and Alex Hernandez-Garcia  \nWe start our article with a survey of related work in Section 2, followed by a presentation of our methodology in Section 3 . In Section 4 we present our analysis, and we conclude with our proposals for future work, including a centralized hub for reporting the carbon footprint of machine learning..  \n2 RELATED WORK  \nMeasuring the environmental impact of ML models is a relatively new undertaking, but one that has been gathering momentum in recent years. In the current section, we pre","cbCaik3Ro1GEyfpK","https://ap.wps.com/l/cbCaik3Ro1GEyfpK","pdf",1157402,4,1,19,"English","en",105,"# Introduction\n## Goals and research questions\n# Related work\n## Empirical studies on carbon emissions\n## Tools and approaches for measuring carbon emissions","[{\"question\":\"Why does machine learning training produce environmental emissions?\",\"answer\":\"Training and deploying ML models require computational resources such as GPUs, which consume energy. Energy generation typically involves greenhouse-gas emissions like CO2.\"},{\"question\":\"What does the survey in this paper analyze?\",\"answer\":\"It analyzes carbon emissions across 95 ML models over time, considering energy sources, produced CO2 amounts, temporal emission evolution, and how emissions relate to model performance.\"},{\"question\":\"What gap in existing research motivates this study?\",\"answer\":\"Prior work has largely examined a small number of models and tasks, limiting coverage of the diversity of ML models and leaving many emission-related factors insufficiently explored.\"}]","Counting Carbon - A Survey of Factors Influencing the Emissions of Machine Learning | PDF",1785933230,48,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"counting-carbon-a-survey-of-factors-influencing-the-emissions-of-machine-learning","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,49,52],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":48},"https://docshare.wps.com/document/","Document",2,{"item":50,"name":12,"@type":44,"position":51},"https://docshare.wps.com/document/research-report/",3,{"item":53,"name":13,"@type":44,"position":20},"https://docshare.wps.com/document/counting-carbon-a-survey-of-factors-influencing-the-emissions-of-machine-learning/126542/",{"url":53,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-27","2026-08-05",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"Why does machine learning training produce environmental emissions?","Question",{"text":76,"@type":77},"Training and deploying ML models require computational resources such as GPUs, which consume energy. Energy generation typically involves greenhouse-gas emissions like CO2.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"What does the survey in this paper analyze?",{"text":81,"@type":77},"It analyzes carbon emissions across 95 ML models over time, considering energy sources, produced CO2 amounts, temporal emission evolution, and how emissions relate to model performance.",{"name":83,"@type":74,"acceptedAnswer":84},"What gap in existing research motivates this study?",{"text":85,"@type":77},"Prior work has largely examined a small number of models and tasks, limiting coverage of the diversity of ML models and leaving many emission-related factors insufficiently explored.","https://schema.org",{"og:url":53,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":53},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,116,121,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":48,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":112,"doc_module":4,"doc_module_name":47,"category_name":113,"show_sort_weight":114,"slug":115},6,"Technology",50,"technology",{"id":117,"doc_module":4,"doc_module_name":47,"category_name":118,"show_sort_weight":119,"slug":120},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":122,"slug":123},30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":22,"doc_module":4,"doc_module_name":47,"category_name":137,"show_sort_weight":107,"slug":138},"General","general"]