The DOI matches its DataCite record; 18 of the 18 references it cites were found in a registry or an index.
Quantifying the Carbon Emissions of Machine Learning
Open accessAbstract
From an environmental standpoint, there are a few crucial aspects of training a neural network that have a major impact on the quantity of carbon that it emits. These factors include: the location of the server used for training and the energy grid that it uses, the length of the training procedure, and even the make and model of hardware on which the training takes place. In order to approximate these emissions, we present our Machine Learning Emissions Calculator, a tool for our community to better understand the environmental impact of training ML models. We accompany this tool with an explanation of the factors cited above, as well as concrete actions that individual practitioners and organizations can take to mitigate their carbon emissions.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“the 'Machine Learning Emissions Calculator' for estimating the carbon footprint of GPU compute through specifying hardware type, hours used, cloud provider, and region [12], Anthony et al. [3]”
From Sustainable AI: AI for sustainability and the sustainability of AI (2021) · cites this paper as [12]
In the citing paper
“In Section 5, we discuss how while the estimation methods of these works provide some understanding of carbon and energy impacts, nuances in the estimation methods may make them inaccurate-particularly in experiments which utilize combined CPU and GPU workloads heavily.”
From Towards the Systematic Reporting of the Energy and Carbon Footprints of Machine Learning (2020) · cites this paper as Lacoste et al., 2019
In the citing paper
“Some GPUs have substantially higher efficiency in terms of floating point operations per second (FLOPS) per watt of power usage compared to others (Lacoste et al., 2019).”
From Carbontracker: Tracking and Predicting the Carbon Footprint of Training Deep Learning Models (2020) · cites this paper as Lacoste et al., 2019
In the citing paper
“Estimations were conducted using the Machine Learning Impact calculator presented in [66].”
From SatMAE: Pre-Training Transformers for Temporal and Multi-Spectral Satellite Imagery (2022) · cites this paper as [66]
In the citing paper
“For a model pre-trained and finetuned on fMoW Sentinel dataset, total emissions are estimated to be 109.44 kg CO 2 eq.”
From SatMAE: Pre-Training Transformers for Temporal and Multi-Spectral Satellite Imagery (2022) · cites this paper as [66]
In the citing paper
“Total emissions are estimated to be 136.8 kg CO 2 eq. of which 100 percent was directly offset by the cloud provider.”
From SatMAE: Pre-Training Transformers for Temporal and Multi-Spectral Satellite Imagery (2022) · cites this paper as [66]
Citations
Cites
Very Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“For a model such as VGG [18] or BERT [19], which are trained on multiple GPUs for several weeks, this can correspond to avoiding emitting several hundreds of kilograms of CO 2 eq by training on a server powered by hydroelectricity instead of fossil fuels.”
MethodBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325 · cited 1×
“For a model such as VGG [18] or BERT [19], which are trained on multiple GPUs for several weeks, this can correspond to avoiding emitting several hundreds of kilograms of CO 2 eq by training on a server powered by hydroelectricity instead of fossil fuels.”
Method(2012) · Cited by 8,009
Universal Language Model Fine-tuning for Text Classification
Howard, J., & Ruder, S. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/p18-1031 · Cited by 4,100 · cited 1×
“Notably, recent research has shown that using pre-trained models with task-specific fine-tuning performs as well as training from scratch, while being more robust, for tasks in image recognition [7, 8] and NLP [9].”
BackgroundEnergy and Policy Considerations for Deep Learning in NLP
Strubell, E., Ganesh, A., & McCallum, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1355 · Cited by 3,684 · cited 2×
“This trend has been the subject of recent studies aiming to evaluate the climate impact of AI, which have predominantly put the focus on the environmental cost of training large-scale models connected to grids powered by fossil fuels [1, 2].”
BackgroundConvolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?
Tajbakhsh, N., Shin, J. Y., Gurudu, S. R., Hurst, R. T., Kendall, C. B., Gotway, M. B., & Liang, J. (2016) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2016.2535302 · Cited by 3,277 · cited 1×
“Notably, recent research has shown that using pre-trained models with task-specific fine-tuning performs as well as training from scratch, while being more robust, for tasks in image recognition [7, 8] and NLP [9].”
BackgroundHyperband: a novel bandit-based approach to hyperparameter optimization
Li, L., Jamieson, K., DeSalvo, G., Rostamizadeh, A., & Talwalkar, A. (2017) · Journal of Machine Learning Research · Cited by 930
Fine-tuned Language Models for Text Classification.Preprint
Howard, J., & Ruder, S. (2018) · arXiv (Cornell University) · Cited by 357
Food image recognition using deep convolutional network with pre-training and fine-tuning
Yanai, ., & Kawano, Y. (2015) · IEEE International Conference on Multimedia & Expo Workshops (ICMEW) · doi:10.1109/icmew.2015.7169816 · Cited by 339 · cited 1×
“Notably, recent research has shown that using pre-trained models with task-specific fine-tuning performs as well as training from scratch, while being more robust, for tasks in image recognition [7, 8] and NLP [9].”
Background
Cited by
Sorted by citations
Items 201–300 of 414
Sundaram, G. A. S., Gandhiraj, R., Binoy, B. N., Ilango, H. S., & Surya, S. N. (2022) · IEEE Access · doi:10.1109/access.2022.3206371 · Cited by 8
“The machine learning emissions calculator [65] has been 1254 used to estimate the equivalent black carbon footprint [66] of 1255 prominent cloud-based computing resources and on reduced 1256 scales by the many other bespoke ML initiatives – including 1257 the one considered in this work.”
MethodEnergAt: Fine-Grained Energy Attribution for Multi-Tenancy
Hè, H., Friedman, M., & Rekatsinas, T. (2023) · Workshop on Sustainable Computer Systems (HotCarbon) · doi:10.1145/3604930.3605716 · Cited by 10
“In recent years, several tools have been developed to measure software energy consumption (e.g., [6, 12, 21, 25, 35, 53]).”
BackgroundMethodMulti-Task Training with In-Domain Language Models for Diagnostic Reasoning
Sharma, B., Gao, Y., Miller, T. A., Churpek, M. M., Afshar, M., & Dligach, D. (2023) · Clinical Natural Language Processing Workshop (ClinicalNLP) · doi:10.18653/v1/2023.clinicalnlp-1.10 · Cited by 10
G3: Geolocation via Guidebook Grounding
Luo, G., Biamby, G., Darrell, T., Fried, D., & Rohrbach, A. (2022) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2022.findings-emnlp.430 · Cited by 9
Smith, M. J. (2020) · University of Hertfordshire Research Archive (University of Hertfordshire) · doi:10.5281/zenodo.4072940 · Cited by 8
Machine Learning Methods in Climate Finance: A Systematic ReviewPreprint
Alonso, A., Carbó, J. M., & Marqués, J. M. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4352569 · Cited by 9
Rebound Effects in Methods of Artificial Intelligence
Willenbacher, M., Hornauer, T., & Wohlgemuth, V. (2022) · Progress in IS · doi:10.1007/978-3-030-88063-7_5 · Cited by 7
Evaluating station, satellite, & combined data for XGBoost-based visibility forecast
Schütz, M., Schütz, A., Bendix, J., & Thies, J. M. U. B. (2025) · Atmospheric Research · doi:10.1016/j.atmosres.2025.108395 · Cited by 7
Martino, V. D., Martínez‐Fernández, S., & Palomba, F. (2025) · IEEE/ACM International Conference on Software Engineering: Software Engineering in Society (ICSE-SEIS) · doi:10.1109/icse-seis66351.2025.00019 · Cited by 7
“In recent years, the software engineering (SE) community has made considerable efforts in promoting software sustainability, with specific attention now being paid to ML-enabled systems [6], [7], [8].”
Douwes, C., Bindi, G., Caillon, A., Esling, P., & Briot, J.-P. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10096975 · Cited by 8
“Recently, [11] proposed an online tool called the Machine Learning Impact Calculator , which estimates carbon emissions produced while training deep leaning models according to the overall time, hardware and geographic position.”
MethodWhen Layers Play the Lottery, all Tickets Win at Initialization
Jordão, A., Araújo, G. H. S., Maia, H. D. A., & Pedrini, H. (2023) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw60793.2023.00130 · Cited by 7
Gutierrez-Bustamante, M., & Espinosa-Leal, L. (2022) · Preprints.org · doi:10.20944/preprints202207.0090.v1 · Cited by 7
Environmental Impacts of Machine Learning Applications in Protein SciencePreprint
Lannelongue, L., & Inouye, M. (2023) · Cold Spring Harbor Perspectives in Biology · doi:10.1101/cshperspect.a041473 · Cited by 7
Assessing the Impact of Artificial Intelligence on Healthcare Sector for Enhancing Sustainability
Crowther, D., Singh, R., Singh, A. P., & Rani, S. (2024) · Approaches to global sustainability, markets, and governance · doi:10.1007/978-981-97-9555-0_1 · Cited by 6
FieldNet: Efficient real-time shadow removal for enhanced vision in field robotics
Saleh, A., Olsen, A., Wood, J., Philippa, B. W., & Azghadi, M. R. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127442 · Cited by 6
“Estimations were conducted using the Machine Learning Impact calculator presented in (Lacoste et al., 2019).”
Deep learning-based object detection for environmental monitoring using big data
Lin, W., Li, T., & Li, X. (2025) · Frontiers in Environmental Science · doi:10.3389/fenvs.2025.1566224 · Cited by 11
Green metrics for AI: A hybrid strategy for environmental impact assessment
Borraccia, S., Masciari, E., & Napolitano, E. V. (2025) · Array · doi:10.1016/j.array.2025.100528 · Cited by 8
GreenDFL: A framework for assessing the sustainability of Decentralized Federated Learning systems
Feng, C., Celdrán, A. H., Cheng, X., Bovet, G., & Stiller, B. (2025) · Information and Software Technology · doi:10.1016/j.infsof.2025.107937 · Cited by 8
Highlighting the Importance of Reducing Research Bias and Carbon Emissions in CNNs
Badar, A., Varma, A., Staniec, A., Gamal, M., Magdy, O., Iqbal, H., Arani, E., & Zonooz, B. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-08421-8_36 · Cited by 7
Gauging Carbon Footprint of AI/ML Implementations in Smart Cities: Methods and Challenges
Rajkumar, P. V. (2022) · International Conference on Fog and Mobile Edge Computing (FMEC) · doi:10.1109/fmec57183.2022.10062634 · Cited by 7
“the location of the server used for training and the energy grid that it uses, the length of the training procedure, and even the make and model of hardware on which the training takes place [1].”
BackgroundMethodInfluentialSaxena, V., Rethmeier, N., DIJCK, G. V., & Spanakis, G. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.481 · Cited by 6
“These estimations were conducted on Tesla V100-SXM2-32GB (TDP of 300W) using the MachineLearn-ing Impact calculator presented in (Lacoste et al., 2019).”
MethodReal-Time Event Detection Using Self-Evolving Contextual Analysis (SECA) Approach
Sulaimani, S. A., & Starkey, A. J. (2023) · IEEE Access · doi:10.1109/access.2023.3331219 · Cited by 6
“Such activities have motivated many researchers in the community [33], [34], [35], [36], [37], [38] (with terminologies such as “Green AI”, “Red AI”, “Environmentally-Friendly AI”, “Carbon Footprints of Machine Learning”), whether to prepare for any current or future recommendations or regulations,…”
BackgroundM6GP: Multiobjective Feature Engineering
Batista, J. E., Rodrigues, N. M., Vanneschi, L., & Silva, S. (2024) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec60901.2024.10612107 · Cited by 6
“Due to the current trend in producing larger and more complex models, evaluating their environmental impact during training and deployment [46], [47] sustainable AI has become a rising field within AI.”
BackgroundNeuroevolution-enhanced multi-objective optimization for mixed-precision quantization
Miret, S., Chua, V. S., Marder, M., Phielipp, M., Jain, N. K., & Majumdar, S. (2022) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3512290.3528692 · Cited by 7
Lu, X., Liu, Z., Guan, Y., Zhang, H., Zhuang, C., Ma, W., Tan, Y., Gu, J., & Zhang, G. (2023) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2023/677 · Cited by 7
“As for energy and carbon emission, following the method proposed in [Lacoste et al. , 2019], we first calculate the energy consumption (EC) mainly taking the RAM, CPU and GPU into consideration: where PUE is a constant determined by the energy efficiency of a data center (the worldwide average is…”
MethodAttanasio, G., Plaza-del-Arco, F. M., Nozza, D., & Lauscher, A. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.243 · Cited by 8
Creating AI Art Responsibly: A Field Guide for Artists
Partnership on, A. I., Leibowicz, C., Saltz, E., Times, T. N. Y., & Coleman, L. (2021) · Revista Diseña · doi:10.7764/disena.19.article.5 · Cited by 6
NeurIPS 2020 NLC2CMD Competition: Translating Natural Language to Bash CommandsPreprint
Agarwal, M., Chakraborti, T., Fu, Q., Gros, D., Lin, X., Maene, J., Talamadupula, K., Teng, Z., & White, J. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.02523 · Cited by 9
Supervising the Transfer of Reasoning Patterns in VQAPreprint
Kervadec, C., Wolf, C., Antipov, G., Baccouche, M., & Nadri, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.05597 · Cited by 6
Silverman, A. L., Bhasuran, B., Mosenia, A., Yasini, F., Ramasamy, G., Banerjee, I., Gupta, S., Mardirossian, T., Narain, R., Sewell, J. L., et al. (2024) · Inflammatory Bowel Diseases · doi:10.1093/ibd/izae068 · Cited by 6
EcoMLS: A Self-Adaptation Approach for Architecting Green ML-Enabled Systems
Tedla, M., Kulkarni, S., & Vaidhyanathan, K. (2024) · IEEE International Conference on Software Architecture Companion (ICSA-C) · doi:10.1109/icsa-c63560.2024.00048 · Cited by 5
Special Session: End-To-End Carbon Footprint Assessment and Modeling of Deep Learning
Faiz, A., Jiang, L., & Chen, F. (2024) · International Conference on Hardware/Software Codesign and System Synthesis · doi:10.1109/codes-isss60120.2024.00011 · Cited by 5
Serizel, R., Cornell, S., & Turpault, N. (2023) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49357.2023.10095938 · Cited by 5
“However as argued by many works [1–4] some deep learning recent trends, both in industry and academia, raise fundamental concerns about the environmental impact and energy consumption of deep learning approaches.”
BackgroundRevisiting Lightweight Compiler Provenance Recovery on ARM Binaries
Kim, J. Z., Genkin, D., & Leach, K. (2023) · IEEE International Conference on Program Comprehension (ICPC) · doi:10.1109/icpc58990.2023.00044 · Cited by 5
“Their complexity makes them expensive to train not only in time, but also energy [20, 32, 45, 46]: the CO 2 footprint of training a large AI model can exceed 626,000 pounds [45].”
BackgroundActiveAED: A Human in the Loop Improves Annotation Error Detection
Weber, L., & Plank, B. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.562 · Cited by 5
A Unified Framework of Ethical Principles for AI
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0004 · Cited by 9
Data Augmentation for Code Translation with Comparable Corpora and Multiple References
Xie, Y., Naik, A., Fried, D., & Rosé, C. P. (2023) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2023.findings-emnlp.917 · Cited by 6
Rigutini, L., Globo, A., Stefanelli, M., Zugarini, A., Gültekin, S., & Ernandes, M. (2024) · International Journal on Natural Language Computing · doi:10.5121/ijnlc.2024.13102 · Cited by 6
“Simultaneously, the literature has introduced various tools for assessing the Carbon-Footprint, such as those presented by Luccioni et al. [14] and Code-Carbon [13].”
MethodUsing NLP to quantify the environmental cost and diversity benefits of in-person NLP conferences
Przybyła, P., & Shardlow, M. (2022) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2022.findings-acl.304 · Cited by 5
“Work has also been undertaken to estimate the overall cost of training machine learning (ML) models — taking into account not only the training time, but also the age of the hardware and server location (García-Martín et al., 2019; Lacoste et al., 2019).”
BackgroundHan, Y., Wang, X., & Zhe, C. (2022) · Scientific Programming · doi:10.1155/2022/2216781 · Cited by 5
“A wide number of studies have examined the total amount [2, 3], footprint [4, 5], intensity [6], and performance [7] of CO2 emissions from various perspectives and geographical scales, revealing the evolutionary characteristics and drivers of the spatial and temporal patterns of CO2 emissions [8–10].”
BackgroundArtefact removal from micrographs with deep learning based inpainting
Squires, I., Dahari, A., Cooper, S. J., & Kench, S. (2023) · Digital Discovery · doi:10.1039/d2dd00120a · Cited by 5
The EcoIndex metric, reviewed from the perspective of Data Science techniques
Cérin, C., Trystram, D., & Menouer, T. (2023) · IEEE Annual International Computer Software and Applications Conference (COMPSAC) · doi:10.1109/compsac57700.2023.00172 · Cited by 5
“More details can be found in [4].”
Kamruzzaman, M., & Kim, G. L. (2023) · International Workshop on Natural Language Processing for Social Media (SocialNLP) · doi:10.18653/v1/2023.socialnlp-1.2 · Cited by 9
Curb Your Carbon Emissions: Benchmarking Carbon Emissions in Machine TranslationPreprint
Yusuf, M., Surana, P., Gupta, G., & Ramesh, K. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.12584 · Cited by 6
“We compare the performance of these systems for our language pairs using the BLEU score as a standard metric and the CodeCarbon package [1] to track the carbon emissions released during the time of training.”
BackgroundMethodChallenges and opportunities in applying AI to evolutionary morphologyPreprint
He, Y., Mulqueeney, J. M., Watt, E., Salili‐James, A., Barber, N., Camaiti, M., Hunt, E. S. E., Kippax-Chui, O., Knapp, A., Lanzetti, A., et al. (2024) · doi:10.32942/x2s315 · Cited by 5
Görücü, S., Ren, Y., SAMUEL, G. G., & Panagiotidou, G. (2025) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3715275.3732088 · Cited by 4
Comparing data reduction strategies for energy-efficient green recommender systems
Spillo, G., Filippo, A. D., Musto, C., Milano, M., & Semeraro, G. M. (2025) · Journal of Intelligent Information Systems · doi:10.1007/s10844-025-00965-1 · Cited by 5
“Concerning the tools to quantify the carbon footprint of the algorithms, it is worth men - tioning the Machine Learning Emissions Calculator (Lacoste et al., 2019), and the framework introduced in Lannelongue et al. ( 2021), which aims to estimate the carbon footprint of any computational task.”
Evaluating LLMs' divergent thinking capabilities for scientific idea generation with minimal context
Ruan, K., Wang, X., Hong, J., Wang, P., Liu, Y., & Sun, H. (2026) · Nature Communications · doi:10.1038/s41467-026-70245-1 · Cited by 50
“Finally, a practical consideration is the environmental cost associated with the extensive LLM usage required for comprehensive benchmarking 68 .”
Reliable recommendations for CCTV sewer inspections through multi-label image classification
Cuingnet, R., Bernard, M., Sampaio, P. R., Sakhri, I., Chelouche, K., Jossent, J., Doumi, I., Gaudichet, E., Chenu, D., Maitrot, A., et al. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2025.103317 · Cited by 4
Not Enough Data to Pre-train Your Language Model? MT to the Rescue!
Urbizu, G., Vicente, I. S., Saralegi, X., & Corral, A. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.235 · Cited by 4
“8e+19 FLOPs of computation, and has an estimated 196 kg CO2 emissions, estimated with Machine-Learning Impact calculator7 (Lacoste et al., 2019).”
MethodMonitor the energy and carbon emissions of process-based models: ProcessC
Li, Z., Qi, Z., Li, B., Xu, J., Wu, R., Liu, Y., & Smith, W. (2024) · Resources Conservation and Recycling · doi:10.1016/j.resconrec.2024.108101 · Cited by 4
Bieser, J. C. T. (2024) · International Conference on ICT for Sustainability (ICT4S) · doi:10.1109/ict4s64576.2024.00010 · Cited by 4
“There is a growing body of GHG assessments for different types of ML models, which cover an increasing scope of GHG effects along the entire life cycle of ML models and the required hardware [10], [11], [12], [13], [14].”
Toward Drought Modeling in South Asia: Machine Learning Approaches, Challenges, and Opportunities
Reshan, M. S. A., Raza, M. O., Mahoto, N. A., Rajab, A. D., Shaikh, A., Elmagzoub, M. A., & Rajab, K. D. (2025) · IEEE Access · doi:10.1109/access.2025.3567257 · Cited by 4
“Explainable Artificial Intelligence [147] is essential for creating black-box deep learning models [139] interpretable models so that stakeholder can gain better insights into the underlying mechanisms of drought.”
Optimizing Datasets for Code Summarization: Is Code-Comment Coherence Enough?
Vitale, A., Mastropaolo, A., Oliveto, R., Penta, M. D., & Scalabrino, S. (2025) · IEEE International Conference on Program Comprehension (ICPC) · doi:10.1109/icpc66645.2025.00033 · Cited by 8
Return on Investment in Machine Learning: Crossing the Chasm between Academia and Business
Mizgajski, J. T., Szymczak, A., Morzy, M., Augustyniak, Ł., Szymański, P., & Żelasko, P. (2020) · Foundations of Computing and Decision Sciences · doi:10.2478/fcds-2020-0015 · Cited by 8
“• real environmental costs of ML training [76],”
BackgroundGenerative Adversarial Networks: A Brief History and Overview
Gunasekaran, A. (2023) · Journal of Student Research · doi:10.47611/jsr.v12i1.1848 · Cited by 4
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0002 · Cited by 6
On Bilingual Lexicon Induction with Large Language Models
Li, Y., Korhonen, A., & Vulić, I. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.595 · Cited by 9
“Based on a publicly available ‘machine learning emissions calculator’ (Luccioni et al., 2019) 27 and our computational infrastructure, we estimate that our work causes the emission of circa 200 kg CO 2 equivalents.”
MethodGreenformer: Factorization Toolkit for Efficient Deep Neural NetworksPreprint
Cahyawijaya, S., Winata, G. I., Lovenia, H., Wilie, B., Dai, W., Ishii, E., & Fung, P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.06762 · Cited by 6
“Many recent works (Strubell, Ganesh, and McCallum 2019; Lacoste et al. 2019; Patterson et al. 2021; Dai et al. 2021; Menghani 2021) address the environmental challenges such as energy usage and carbon emission level of DNN models and develop more ef-ficient deep learning solutions.”
Estimating the Sustainability of AI Models Based on Theoretical Models and Experimental DataPreprint
Gitzel, R., Platenius‐Mohr, M., & Burger, A. (2023) · Preprints.org · doi:10.20944/preprints202301.0406.v1 · Cited by 4
PosAx-O: Exploring Operator-level Approximations for Posit Arithmetic in Embedded AI/ML
Immaneni, A., Ullah, S., Nambi, S., Sahoo, S. S., & Kumar, A. S. (2022) · Euromicro Conference on Digital System Design (DSD) · doi:10.1109/dsd57027.2022.00037 · Cited by 3
Kaplan, A., Keim, J., Greiner, L., Sieger, R., Mirandola, R., & Reussner, R. (2025) · IEEE/ACM International Workshop on Green and Sustainable Software (GREENS) · doi:10.1109/greens66463.2025.00016 · Cited by 4
“CodeCarbon (2020, MIT) is a Python package and the successor of ML CO2 Impact [32] for estimating and tracking carbon emissions by considering computing infrastructure, location, hardware, usage, and runtime [33].”
A systematic review of Green Machine Learning: practices and challenges for sustainability
Santos, S., Ottoni, A. L. C., Borgo, R., Ferreira, D., & Nepomuceno, E. (2026) · Artificial Intelligence Review · doi:10.1007/s10462-026-11515-8 · Cited by 6
Eryılmaz, B., Koraş, O. A., Schlötterer, J., & Seifert, C. (2024) · IEEE International Conference on Cognitive Machine Intelligence (CogMI) · doi:10.1109/cogmi62246.2024.00042 · Cited by 6
Evaluating Energy Efficiency of Ai Accelerators Using Two Mlperf Benchmarks
Ferdaus, F., Wu, X., Taylor, V., Lan, Z., Shanmugavelu, S., Vishwanath, V., & Papka, M. E. (2025) · IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid) · doi:10.1109/ccgrid64434.2025.00035 · Cited by 4
Paracolli, A., & Arquilla, V. (2023) · Linköping electronic conference proceedings · doi:10.3384/ecp203067 · Cited by 3
“Lacoste et al. (2019) and Kaack et al. (2022) elaborate on the carbon emission sources of an artificial intelligence service, offering a more comprehensive view of the direct impacts in terms of carbon footprint.”
BackgroundInformation Technology for Generating Lyrics for Song Extensions Based on Transformers
Mediakov, O., Vysotska, V., Uhryn, D. I., Ushenko, Y., & Hu, C. (2024) · International Journal of Modern Education and Computer Science · doi:10.5815/ijmecs.2024.01.03 · Cited by 3
The Environmental Price of Intelligence: Evaluating the Social Cost of Carbon in Machine Learning
Hasan, S. M., Shahid, A. R., & Imteaj, A. (2024) · IEEE Conference on Technologies for Sustainability · doi:10.1109/sustech60925.2024.10553496 · Cited by 3
“There are multiple emission trackers in the open-source community, such as machine learning emissions calculator [28], carbontrcaker [31], experiment-impact-tracker [26] etc.”
BackgroundA Protocol to Assess the Accuracy of Process-Level Power Models
Cadorel, E., & Saingre, D. (2024) · IEEE International Conference on Cluster Computing (CLUSTER) · doi:10.1109/cluster59578.2024.00014 · Cited by 3
Li, M., & Mohammadi, J. (2024) · North American Power Symposium (NAPS) · doi:10.1109/naps61145.2024.10741623 · Cited by 3
“Several key factors, including the geographic location of the server, the type of GPU, and the duration of the training, are considered to estimate the approximate CO2 equivalent emissions produced [8].”
BackgroundImandi, R., Roy, A., Kim, Y.-G., & N, P. K. B. (2025) · Sustainable Computing Informatics and Systems · doi:10.1016/j.suscom.2025.101133 · Cited by 3
Dynamic Neural Diversification: Path to Computationally Sustainable Neural Networks
Kovalenko, A., Kordík, P., & Friedjungová, M. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-86340-1_19 · Cited by 3
“Recently, researchers started questioning energy consumption of machine learning algorithms and their carbon footprint [3].”
BackgroundMUNI-NLP Systems for Low-resource Indic Machine Translation
Signoroni, E., & Rychlý, P. (2023) · Conference on Machine Translation (WMT) · doi:10.18653/v1/2023.wmt-1.91 · Cited by 3
Data augmentation and image understandingPreprint
Hernández-García, Á. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.14185 · Cited by 7
“Therefore, the contribution of deep learning research to global warming and climate change should not be neglected (Schwartz et al., 2019; Lacoste et al., 2019; Lannelongue et al., 2020).”
BackgroundOn Contrastive Representations of Stochastic ProcessesPreprint
Mathieu, É., Foster, A., & Teh, Y. W. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.10052 · Cited by 4
“Estimations were conducted using the Machine Learning Impact calculator presented in Lacoste et al. (2019).”
MethodMDP Playground: A Design and Debug Testbed for Reinforcement LearningPreprint
Rajan, R., Diaz, J. L. B., Guttikonda, S., Ferreira, F., Biedenkapp, A., von Hartz, J. O., & Hutter, F. (2021) · arXiv (Cornell University) · Cited by 2
The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink
Patterson, D., Gonzalez, J., Hölzle, U., Le, Q. H., Liang, C., Munguia, L.-M., Rothchild, D., So, D., Texier, M., & Dean, J. (2022) · doi:10.36227/techrxiv.19139645 · Cited by 3
“More successful attempts to simplify calculation of energy are online calculators, such as the ML Emissions Calculator [5,19].”
MethodPedersen, M., Pardoe, H., de Weger, A., Hutchison, D. M., Abbott, D., Verspoor, K. M., & Jackson, G. D. (2023) · doi:10.31219/osf.io/kag75 · Cited by 3
Machine Intelligence in Africa: a surveyPreprint
Tembiné, H., Tapo, A. A., Danioko, S., & Traoré, A. (2024) · doi:10.36227/techrxiv.170555182.20418305/v1 · Cited by 3
On Generative Artificial Intelligence: Open-Source is the WayPreprint
Merilehto, J. (2024) · SocArXiv (OSF Preprints) · doi:10.31235/osf.io/jnmzg · Cited by 5
EcoMLS: A Self-Adaptation Approach for Architecting Green ML-Enabled SystemsPreprint
Tedla, M., Kulkarni, S., & Vaidhyanathan, K. (2024) · doi:10.36227/techrxiv.171177355.53837507/v1 · Cited by 3
Rodríguez-Garavito, C. A., Gruber, D. F., Nemeth, A., & Beguš, G. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5165527 · Cited by 3
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0005 · Cited by 2
The low-carbon fruit tree for primary care
Sergeant, M. D., Zahid, D., Zvereva, E., Douglass, R. A., Kroeker, L., Tieu, M., Mckenna, B., & Nash, W. S. (2025) · Healthcare Management Forum · doi:10.1177/08404704251333639 · Cited by 2
“68,69 70-73 It also mitigates environmental costs associated with accessing care such as transportation.”
Flexico: Sustainable Machine Translation via Self-Adaptation
Casimiro, M., Romano, P., de Souza, J. G. C., Khan, A. M., & Garlan, D. (2025) · International Symposium on Software Engineering for Adaptive and Self-Managing Systems (SEAMS) · doi:10.1109/seams66627.2025.00024 · Cited by 2
“…ML models [37]–[39] (ii) creating tools for self-reporting and evaluating the energy consumption and carbon emissions of training ML models [40]–[42]; (iii) researching green architectural tactics [43] and creating systems that account for the ML model’s energy consumption when selecting…”
Deep GreenAI: Effective Layer Pruning Method for Modern Neural Networks
Pons, I., Yamamoto, B. L., & Correia, A. J. L. (2025) · Revista Eletrônica de Iniciação Científica em Computação · doi:10.5753/reic.2025.6038 · Cited by 2
An Energy-Based Comparative Analysis of Common Approaches to Text Classification in the Legal Domain
Gültekin, S., Globo, A., Zugarini, A., Ernandes, M., & Rigutini, L. (2024) · International Conference on AI, Machine Learning and Applications (AIMLA) · doi:10.5121/csit.2024.140203 · Cited by 4
“The charts were recreated with data from [17]. in the literature, as by [12] and [11].”
MethodActivation Control of Vision Models for Sustainable AI Systems
Burton-Barr, J., Fernando, B., & Rajan, D. (2024) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2024.3372935 · Cited by 4
“With a rising percentage of global energy consumption [5] and an increasing carbon footprint [6], [7], reports have suggested that the unregulated adoption of AI technology could become unsustainable in the near future [8].”
Exploring Human-AI Collaboration and Explainability for Sustainable ML
chaaben, E. B. (2024) · Frontiers in artificial intelligence and applications · doi:10.3233/faia240209 · Cited by 2
Energy Consumption Trends in Sound Event Detection Systems
Douwes, C., & Serizel, R. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889562 · Cited by 2
“This growing energy demand raises concerns about the associated environmental impact [5]–[7], concerns that first emerged in the field of natural language processing [8] and, more recently, within the audio processing and machine listening communities [9], [10].”
Towards Net-Zero Carbon Emissions in Federated Edge Intelligence
Cai, H., Xiao, Y. L., Li, Y., Zhou, Y., Niyato, D. T., & Sun, S. (2024) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2024-spring62846.2024.10683033 · Cited by 3
“Some frameworks have effectively reduced emissions [2], [3], there is still lacking a unified framework that can analyze the overall environmental impact, especially the greenhouse emissions, of AI-based services that may depend on a range of key elements and processes including data collecting,…”
BackgroundEnhancing Monocular Depth Estimation with Multi-Source Auxiliary Tasks
Quercia, A., Yildiz, E., Cao, Z., Krajsek, K., Morrison, A., Assent, I., & Scharr, H. (2025) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv61041.2025.00627 · Cited by 2
Grah, J. S., Irrgang, C., Schaade, L., Ladewig, K., & Körber, N. (2025) · Bundesgesundheitsblatt - Gesundheitsforschung - Gesundheitsschutz · doi:10.1007/s00103-025-04098-2 · Cited by 2
Measuring the Environmental Cost in the Evaluation of Metaheuristics
Hernández, P. N., Puris, A., & Pelta, D. A. (2021) · Atlantis studies in uncertainty modelling/Atlantis Studies in Uncertainty Modelling · doi:10.2991/asum.k.210827.028 · Cited by 2
Flexible Parallel Learning in Edge Scenarios: Communication, Computational and Energy Cost
Malandrino, F., & Chiasserini, C. F. (2022) · IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops) · doi:10.1109/percomworkshops53856.2022.9767275 · Cited by 2
“I summarizes the energy cost and carbon footprint of each approach, computed according to [21] for private servers running in northern Italy and supplied by the national energy provider Enel, which has a carbon efficiency of 0.”
MethodScaling Laws for BERT in Low-Resource Settings
Urbizu, G., Vicente, I. S., Saralegi, X., Agerri, R., & Soroa, A. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.492 · Cited by 4
“CO 2 emissions were estimated with Machine-Learning Impact calculator 15 (Lacoste et al., 2019).”
MethodFloridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0003 · Cited by 2
AI and the UN Sustainable Development Goals
Floridi, L. (2023) · Oxford University Press eBooks · doi:10.1093/oso/9780198883098.003.0012 · Cited by 2
Metadata
- Authors
- Lacoste, A., Luccioni, A., Schmidt, V., & Dandres, T.
- Affiliations
- Polytechnique Montréal (CA)
- Venue
- arXiv
- Publisher
- arXiv
- Published
- 2019
- ISSN
- 2331-8422 (ISSN-L)
- arXiv
- 1910.09700
- OpenAlex
- W2981540061
- Cited by
- 408
- References
- 18
- Citations by year
- 2026: 13 · 2025: 55 · 2024: 71 · 2023: 139 · 2022: 48 · 2021: 64 · 2020: 16
- Type
- posted content
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Green IT and Sustainability
- Physical Sciences › Computer Science › Hardware and Architecture › Energy Efficiency in Computing
- Physical Sciences › Environmental Science › Environmental Engineering › Environmental Impact and Sustainability
- Keywords
- CO2 emissions, machine learning training, environmental impact of AI, neural network training, power grid, training duration, carbon mitigation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
