| Item Type | Web Page |
|---|---|
| Author | Paul Walsh |
| Date | 03 22 2023 |
| URL | https://www.linkedin.com/pulse/here-comes-sun-why-large-language-models-dont-have-cost-paul-walsh/ |
| Accessed | 5/3/2025, 7:44:37 AM |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Newspaper Article |
|---|---|
| Author | Delger Erdenesanaa |
| Abstract | Behind the scenes, the technology relies on thousands of specialized computer chips. |
| Date | 2023-10-10 |
| Language | en-US |
| Library Catalog | NYTimes.com |
| URL | https://www.nytimes.com/2023/10/10/climate/ai-could-soon-need-as-much-electricity-as-an-entire-country.html |
| Accessed | 5/3/2025, 7:39:59 AM |
| Section | Climate |
| Publication | The New York Times |
| ISSN | 0362-4331 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Blog Post |
|---|---|
| Author | Kasia |
| Abstract | With the development of artificial intelligence (AI), increasing attention is being paid not only to its potential benefits but also to the environmental consequences of its use. One key aspect of interest is AI's carbon footprint - a measure of greenhouse gas emissions associated with the operation of AI-based systems. |
| Date | 2024-05-08T14:17:41+00:00 |
| Language | en-GB |
| URL | https://planbe.eco/en/blog/ais-carbon-footprint-how-does-the-popularity-of-artificial-intelligence-affect-the-climate/ |
| Accessed | 5/5/2025, 7:07:55 AM |
| Blog Title | Plan Be Eco |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Blog Post |
|---|---|
| Author | Joram Schwartzmann |
| Abstract | Further reading Fairfuel Criteria Carefree flying with e-kerosene? Illustration: atmosfair fairfuel in comparison to conventional fuel Illustration: CO₂-balance of one tonne of kerosene All images with fuel tanker and video material for TV broadcasters is available here: https://files.atmosfair.de/index.php/s/4fsicpiCmFD9sK2 Press contacts atmosfair und Solarbelt gGmbH: Managing Director Dietrich Brockhagen (brockhagen@atmosfair.de) Wolfdietrich Peiker (Peiker@atmosfair.de, 030-1208480-0) Hauser Exkursionen: Managing |
| Date | 2024-06-28T08:00:10+00:00 |
| Language | en-US |
| Short Title | Air Travel and Climate |
| URL | https://www.atmosfair.de/en/air-travel-and-climate-german-plant-produces-first-quantities-of-carbon-neutral-synthetic-kerosene/ |
| Accessed | 5/4/2025, 8:44:25 AM |
| Blog Title | atmosfair |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Alex de Vries |
| Date | 2018-05-16 |
| Language | English |
| Library Catalog | www.cell.com |
| URL | https://www.cell.com/joule/abstract/S2542-4351(18)30177-6 |
| Accessed | 5/4/2025, 8:44:38 AM |
| Extra | Publisher: Elsevier |
| Volume | 2 |
| Pages | 801-805 |
| Publication | Joule |
| DOI | 10.1016/j.joule.2018.04.016 |
| Issue | 5 |
| Journal Abbr | Joule |
| ISSN | 2542-4785, 2542-4351 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Preprint |
|---|---|
| Author | Qinchan Li |
| Author | Kenneth Chen |
| Author | Changyue Su |
| Author | Qi Sun |
| Abstract | Diffusion models have shown unprecedented success in the task of text-to-image generation. While these models are capable of generating high-quality and realistic images, the complexity of sequential denoising has raised societal concerns regarding high computational demands and energy consumption. In response, various efforts have been made to improve inference efficiency. However, most of the existing efforts have taken a fixed approach with neural network simplification or text prompt optimization. Are the quality improvements from all denoising computations equally perceivable to humans? We observed that images from different text prompts may require different computational efforts given the desired content. The observation motivates us to present BudgetFusion, a novel model that suggests the most perceptually efficient number of diffusion steps before a diffusion model starts to generate an image. This is achieved by predicting multi-level perceptual metrics relative to diffusion steps. With the popular Stable Diffusion as an example, we conduct both numerical analyses and user studies. Our experiments show that BudgetFusion saves up to five seconds per prompt without compromising perceptual similarity. We hope this work can initiate efforts toward answering a core question: how much do humans perceptually gain from images created by a generative model, per watt of energy? |
| Date | 2024-12-23 |
| Short Title | BudgetFusion |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2412.05780 |
| Accessed | 5/5/2025, 8:10:32 AM |
| Extra | arXiv:2412.05780 [cs] |
| DOI | 10.48550/arXiv.2412.05780 |
| Repository | arXiv |
| Archive ID | arXiv:2412.05780 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Kerry Day |
| Abstract | The rapid growth of AI brings hope of unprecedented advancements in many sectors but what is its real carbon footprint? |
| Date | 2025-01-13T08:41:52+00:00 |
| Language | en-US |
| URL | https://www.advancedsciencenews.com/calculating-the-true-environmental-costs-of-ai/ |
| Accessed | 1/18/2025, 3:51:02 PM |
| Website Title | Advanced Science News |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Cambridge Digital Assets Programme (CDAP) |
| Abstract | The Cambridge Blockchain Network Sustainability Index (CBNSI) is created and maintained by the Cambridge Digital Assets Programme (CDAP) Team at the Cambridge Centre for Alternative Finance, an independent research institute based at Cambridge Judge Business School, University of Cambridge. |
| Language | en |
| URL | https://ccaf.io/cbnsi/cbeci/ghg |
| Accessed | 5/3/2025, 7:32:15 AM |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Preprint |
|---|---|
| Author | David Patterson |
| Author | Joseph Gonzalez |
| Author | Quoc Le |
| Author | Chen Liang |
| Author | Lluis-Miquel Munguia |
| Author | Daniel Rothchild |
| Author | David So |
| Author | Maud Texier |
| Author | Jeff Dean |
| Abstract | The computation demand for machine learning (ML) has grown rapidly recently, which comes with a number of costs. Estimating the energy cost helps measure its environmental impact and finding greener strategies, yet it is challenging without detailed information. We calculate the energy use and carbon footprint of several recent large models-T5, Meena, GShard, Switch Transformer, and GPT-3-and refine earlier estimates for the neural architecture search that found Evolved Transformer. We highlight the following opportunities to improve energy efficiency and CO2 equivalent emissions (CO2e): Large but sparsely activated DNNs can consume <1/10th the energy of large, dense DNNs without sacrificing accuracy despite using as many or even more parameters. Geographic location matters for ML workload scheduling since the fraction of carbon-free energy and resulting CO2e vary ~5X-10X, even within the same country and the same organization. We are now optimizing where and when large models are trained. Specific datacenter infrastructure matters, as Cloud datacenters can be ~1.4-2X more energy efficient than typical datacenters, and the ML-oriented accelerators inside them can be ~2-5X more effective than off-the-shelf systems. Remarkably, the choice of DNN, datacenter, and processor can reduce the carbon footprint up to ~100-1000X. These large factors also make retroactive estimates of energy cost difficult. To avoid miscalculations, we believe ML papers requiring large computational resources should make energy consumption and CO2e explicit when practical. We are working to be more transparent about energy use and CO2e in our future research. To help reduce the carbon footprint of ML, we believe energy usage and CO2e should be a key metric in evaluating models, and we are collaborating with MLPerf developers to include energy usage during training and inference in this industry standard benchmark. |
| Date | 2021-04-23 |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2104.10350 |
| Accessed | 5/3/2025, 7:43:18 AM |
| Extra | arXiv:2104.10350 [cs] |
| DOI | 10.48550/arXiv.2104.10350 |
| Repository | arXiv |
| Archive ID | arXiv:2104.10350 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Carbon Trust |
| Abstract | This white paper examines the carbon impact of streaming video on demand in order to improve understanding and inform future decisions. |
| Date | 11 June 2021 |
| Language | en |
| URL | https://www.carbontrust.com/our-work-and-impact/guides-reports-and-tools/carbon-impact-of-video-streaming |
| Accessed | 5/3/2025, 7:31:04 AM |
| Website Title | https://www.carbontrust.com/our-work-and-impact/guides-reports-and-tools/carbon-impact-of-video-streaming |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | International Energy Agency (IEA) |
| Abstract | As the world becomes increasingly digitalised, data centres and data transmission networks are emerging as an important source of energy demand. |
| Language | en-GB |
| URL | https://www.iea.org/energy-system/buildings/data-centres-and-data-transmission-networks |
| Accessed | 5/3/2025, 7:33:28 AM |
| Website Title | IEA |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | International Energy Agency (IEA) |
| Abstract | Energy and AI - Analysis and key findings. A report by the International Energy Agency. |
| Date | 2025-04-10 |
| Language | en-GB |
| URL | https://www.iea.org/reports/energy-and-ai |
| Accessed | 5/6/2025, 8:28:58 AM |
| Website Title | IEA |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Preprint |
|---|---|
| Author | Emma Strubell |
| Author | Ananya Ganesh |
| Author | Andrew McCallum |
| Abstract | Recent progress in hardware and methodology for training neural networks has ushered in a new generation of large networks trained on abundant data. These models have obtained notable gains in accuracy across many NLP tasks. However, these accuracy improvements depend on the availability of exceptionally large computational resources that necessitate similarly substantial energy consumption. As a result these models are costly to train and develop, both financially, due to the cost of hardware and electricity or cloud compute time, and environmentally, due to the carbon footprint required to fuel modern tensor processing hardware. In this paper we bring this issue to the attention of NLP researchers by quantifying the approximate financial and environmental costs of training a variety of recently successful neural network models for NLP. Based on these findings, we propose actionable recommendations to reduce costs and improve equity in NLP research and practice. |
| Date | 2019-06-05 |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/1906.02243 |
| Accessed | 5/3/2025, 7:43:08 AM |
| Extra | arXiv:1906.02243 [cs] |
| DOI | 10.48550/arXiv.1906.02243 |
| Repository | arXiv |
| Archive ID | arXiv:1906.02243 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
Comment: In the 57th Annual Meeting of the Association for Computational Linguistics (ACL). Florence, Italy. July 2019
| Item Type | Journal Article |
|---|---|
| Author | Gregory S. Okin |
| Abstract | In the US, there are more than 163 million dogs and cats that consume, as a significant portion of their diet, animal products and therefore potentially constitute a considerable dietary footprint. Here, the energy and animal-derived product consumption of these pets in the US is evaluated for the first time, as are the environmental impacts from the animal products fed to them, including feces production. In the US, dogs and cats consume about 19% ± 2% of the amount of dietary energy that humans do (203 ± 15 PJ yr-1 vs. 1051 ± 9 PJ yr-1) and 33% ± 9% of the animal-derived energy (67 ± 17 PJ yr-1 vs. 206 ± 2 PJ yr-1). They produce about 30% ± 13%, by mass, as much feces as Americans (5.1 ± Tg yr-1 vs. 17.2 Tg yr-1), and through their diet, constitute about 25–30% of the environmental impacts from animal production in terms of the use of land, water, fossil fuel, phosphate, and biocides. Dog and cat animal product consumption is responsible for release of up to 64 ± 16 million tons CO2-equivalent methane and nitrous oxide, two powerful greenhouse gasses (GHGs). Americans are the largest pet owners in the world, but the tradition of pet ownership in the US has considerable costs. As pet ownership increases in some developing countries, especially China, and trends continue in pet food toward higher content and quality of meat, globally, pet ownership will compound the environmental impacts of human dietary choices. Reducing the rate of dog and cat ownership, perhaps in favor of other pets that offer similar health and emotional benefits would considerably reduce these impacts. Simultaneous industry-wide efforts to reduce overfeeding, reduce waste, and find alternative sources of protein will also reduce these impacts. |
| Date | Aug. 2, 2017 |
| Language | en |
| Library Catalog | PLoS Journals |
| URL | https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0181301 |
| Accessed | 5/3/2025, 7:34:18 AM |
| Extra | Publisher: Public Library of Science |
| Volume | 12 |
| Pages | e0181301 |
| Publication | PLOS ONE |
| DOI | 10.1371/journal.pone.0181301 |
| Issue | 8 |
| Journal Abbr | PLOS ONE |
| ISSN | 1932-6203 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Preprint |
|---|---|
| Author | Alexandra Sasha Luccioni |
| Author | Sylvain Viguier |
| Author | Anne-Laure Ligozat |
| Abstract | Progress in machine learning (ML) comes with a cost to the environment, given that training ML models requires significant computational resources, energy and materials. In the present article, we aim to quantify the carbon footprint of BLOOM, a 176-billion parameter language model, across its life cycle. We estimate that BLOOM's final training emitted approximately 24.7 tonnes of~\carboneq~if we consider only the dynamic power consumption, and 50.5 tonnes if we account for all processes ranging from equipment manufacturing to energy-based operational consumption. We also study the energy requirements and carbon emissions of its deployment for inference via an API endpoint receiving user queries in real-time. We conclude with a discussion regarding the difficulty of precisely estimating the carbon footprint of ML models and future research directions that can contribute towards improving carbon emissions reporting. |
| Date | 2022-11-03 |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2211.02001 |
| Accessed | 5/3/2025, 7:43:32 AM |
| Extra | arXiv:2211.02001 [cs] |
| DOI | 10.48550/arXiv.2211.02001 |
| Repository | arXiv |
| Archive ID | arXiv:2211.02001 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Conference Paper |
|---|---|
| Author | Nesrine Bannour |
| Author | Sahar Ghannay |
| Author | Aurélie Névéol |
| Author | Anne-Laure Ligozat |
| Abstract | Modern Natural Language Processing (NLP) makes intensive use of deep learning methods because of the accuracy they offer for a variety of applications. Due to the significant environmental impact of deep learning, costbenefit analysis including carbon footprint as well as accuracy measures has been suggested to better document the use of NLP methods for research or deployment. In this paper, we review the tools that are available to measure energy use and CO2 emissions of NLP methods. We describe the scope of the measures provided and compare the use of six tools (carbon tracker, experiment impact tracker, green algorithms, ML CO2 impact, energy usage and cumulator) on named entity recognition experiments performed on different computational set-ups (local server vs. computing facility). Based on these findings, we propose actionable recommendations to accurately measure the environmental impact of NLP experiments. |
| Date | 2021 |
| Language | en |
| Short Title | Evaluating the carbon footprint of NLP methods |
| Library Catalog | DOI.org (Crossref) |
| URL | https://aclanthology.org/2021.sustainlp-1.2 |
| Accessed | 5/3/2025, 7:42:14 AM |
| Place | Virtual |
| Publisher | Association for Computational Linguistics |
| Pages | 11-21 |
| Proceedings Title | Proceedings of the Second Workshop on Simple and Efficient Natural Language Processing |
| Conference Name | Proceedings of the Second Workshop on Simple and Efficient Natural Language Processing |
| DOI | 10.18653/v1/2021.sustainlp-1.2 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | OAR US EPA |
| Abstract | This page answers questions about GHG emissions from passenger vehicles and how these emissions are measured and calculated. |
| Date | 2016-01-12T16:29:25-05:00 |
| Language | en |
| URL | https://www.epa.gov/greenvehicles/greenhouse-gas-emissions-typical-passenger-vehicle |
| Accessed | 5/4/2025, 8:44:33 AM |
| Website Type | Overviews and Factsheets |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Oliver Belcher |
| Author | Patrick Bigger |
| Author | Ben Neimark |
| Author | Cara Kennelly |
| Abstract | This paper examines the US military's impact on climate by analysing the geopolitical ecology of its global logistical supply chains. Our geopolitical ecology framework interrogates the material-ecological metabolic flows (hydrocarbon-based fuels, water, sand, concrete) that shape geopolitical and geoeconomic power relations. We argue that to account for the US military as a major climate actor, one must understand the logistical supply chain that makes its acquisition and consumption of hydrocarbon-based fuels possible. Our paper focuses on the US Defense Logistics Agency – Energy (DLA-E), a large yet virtually unresearched sub-agency within the US Department of Defense. The DLA-E is the primary purchase-point for hydrocarbon-based fuels for the US military, as well as a powerful actor in the global oil market. After outlining our geopolitical ecology approach, we detail the scope of the DLA-E's operations, its supply chain, bureaucratic practices, and the physical infrastructure that facilitates the US military's consumption of hydro-based carbons on a global scale. We show several “path dependencies” – warfighting paradigms, weapons systems, bureaucratic requirements, and waste – that are put in place by military supply chains and undergird a heavy reliance on carbon-based fuels by the US military for years to come. The paper, based on comprehensive records of bulk fuel purchases we have gathered from DLA-E through Freedom of Information Act requests, represents a partial yet robust picture of the geopolitical ecology of American imperialism. |
| Date | 2020 |
| Language | en |
| Short Title | Hidden carbon costs of the “everywhere war” |
| Library Catalog | Wiley Online Library |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1111/tran.12319 |
| Accessed | 5/3/2025, 7:31:57 AM |
| Rights | The information, practices and views in this article are those of the author(s) and do not necessarily reflect the opinion of the Royal Geographical Society (with IBG). © 2019 Royal Geographical Society (with the Institute of British Geographers). |
| Extra | _eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/tran.12319 |
| Volume | 45 |
| Pages | 65-80 |
| Publication | Transactions of the Institute of British Geographers |
| DOI | 10.1111/tran.12319 |
| Issue | 1 |
| ISSN | 1475-5661 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Matthias C. Rillig |
| Author | India Mansour |
| Author | Stefan Hempel |
| Author | Mohan Bi |
| Author | Birgitta König-Ries |
| Author | Atoosa Kasirzadeh |
| Abstract | Generative artificial intelligence (AI) models will have broad impacts on society including the scientific enterprise; ecology and environmental science will be no exception. Here, we discuss the potential opportunities and risks of advanced generative AI for visual material (images and video) for the science of ecology and the environment itself. There are clearly opportunities for positive impacts, related to improved communication, for example; we also see possibilities for ecological research to benefit from generative AI (e.g., image gap filling, biodiversity surveys, and improved citizen science). However, there are also risks, threatening to undermine the credibility of our science, mostly related to actions of bad actors, for example in terms of spreading fake information or committing fraud. Risks need to be mitigated at the level of government regulatory measures, but we also highlight what can be done right now, including discussing issues with the next generation of ecologists and transforming towards radically open science workflows. |
| Date | 2024 |
| Language | en |
| Library Catalog | Wiley Online Library |
| URL | https://onlinelibrary.wiley.com/doi/abs/10.1111/ele.14397 |
| Accessed | 5/6/2025, 9:29:43 AM |
| Rights | © 2024 The Authors. Ecology Letters published by John Wiley & Sons Ltd. |
| Extra | _eprint: https://onlinelibrary.wiley.com/doi/pdf/10.1111/ele.14397 |
| Volume | 27 |
| Pages | e14397 |
| Publication | Ecology Letters |
| DOI | 10.1111/ele.14397 |
| Issue | 3 |
| ISSN | 1461-0248 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
e14397 ELE-01224-2023.R1
| Item Type | Preprint |
|---|---|
| Author | Ian Schneider |
| Author | Hui Xu |
| Author | Stephan Benecke |
| Author | David Patterson |
| Author | Keguo Huang |
| Author | Parthasarathy Ranganathan |
| Author | Cooper Elsworth |
| Abstract | Specialized hardware accelerators aid the rapid advancement of artificial intelligence (AI), and their efficiency impacts AI's environmental sustainability. This study presents the first publication of a comprehensive AI accelerator life-cycle assessment (LCA) of greenhouse gas emissions, including the first publication of manufacturing emissions of an AI accelerator. Our analysis of five Tensor Processing Units (TPUs) encompasses all stages of the hardware lifespan - from raw material extraction, manufacturing, and disposal, to energy consumption during development, deployment, and serving of AI models. Using first-party data, it offers the most comprehensive evaluation to date of AI hardware's environmental impact. We include detailed descriptions of our LCA to act as a tutorial, road map, and inspiration for other computer engineers to perform similar LCAs to help us all understand the environmental impacts of our chips and of AI. A byproduct of this study is the new metric compute carbon intensity (CCI) that is helpful in evaluating AI hardware sustainability and in estimating the carbon footprint of training and inference. This study shows that CCI improves 3x from TPU v4i to TPU v6e. Moreover, while this paper's focus is on hardware, software advancements leverage and amplify these gains. |
| Date | 2025-02-01 |
| Short Title | Life-Cycle Emissions of AI Hardware |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2502.01671 |
| Accessed | 5/3/2025, 8:26:46 AM |
| Extra | arXiv:2502.01671 [cs] |
| DOI | 10.48550/arXiv.2502.01671 |
| Repository | arXiv |
| Archive ID | arXiv:2502.01671 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Zhenziao Fu |
| Date | 03 10 2024 |
| URL | https://arxiv.org/html/2410.02950v1 |
| Accessed | 5/5/2025, 7:13:28 AM |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Jacob Parsons |
| Abstract | Large Language Models impact the environment, but also enable companies’ journeys towards a more sustainable tomorrow. Learn how organizations and LLMs can manage their impact on the environment better. |
| Date | 17 11 2024 |
| Language | en-US |
| URL | https://eviden.com/insights/blogs/llms-and-the-effect-on-the-environment/ |
| Accessed | 5/3/2025, 8:12:51 AM |
| Website Title | Eviden |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Preprint |
|---|---|
| Author | Pengfei Li |
| Author | Jianyi Yang |
| Author | Mohammad A. Islam |
| Author | Shaolei Ren |
| Abstract | The growing carbon footprint of artificial intelligence (AI) has been undergoing public scrutiny. Nonetheless, the equally important water (withdrawal and consumption) footprint of AI has largely remained under the radar. For example, training the GPT-3 language model in Microsoft's state-of-the-art U.S. data centers can directly evaporate 700,000 liters of clean freshwater, but such information has been kept a secret. More critically, the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of 4-6 Denmark or half of the United Kingdom. This is concerning, as freshwater scarcity has become one of the most pressing challenges. To respond to the global water challenges, AI can, and also must, take social responsibility and lead by example by addressing its own water footprint. In this paper, we provide a principled methodology to estimate the water footprint of AI, and also discuss the unique spatial-temporal diversities of AI's runtime water efficiency. Finally, we highlight the necessity of holistically addressing water footprint along with carbon footprint to enable truly sustainable AI. |
| Date | 2025-01-15 |
| Short Title | Making AI Less "Thirsty" |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2304.03271 |
| Accessed | 1/18/2025, 3:50:39 PM |
| Extra | arXiv:2304.03271 [cs] |
| DOI | 10.48550/arXiv.2304.03271 |
| Repository | arXiv |
| Archive ID | arXiv:2304.03271 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
Comment: Accepted by Communications of the ACM. Source codes available at: https://github.com/Ren-Research/Making-AI-Less-Thirsty
| Item Type | Web Page |
|---|---|
| Author | Matthew Hutson |
| Abstract | New tools track and reduce emissions from machine learning |
| Date | 26 06 2022 |
| Language | en |
| URL | https://spectrum.ieee.org/ai-carbon-footprint |
| Accessed | 5/3/2025, 7:41:43 AM |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Michael A Clark |
| Author | Marco Springmann |
| Author | Jason Hill |
| Author | David Tilman |
| Abstract | Food choices are shifting globally in ways that are negatively affecting both human health and the environment. Here we consider how consuming an additional serving per day of each of 15 foods is associated with 5 health outcomes in adults and 5 aspects of agriculturally driven environmental degradation. We find that while there is substantial variation in the health outcomes of different foods, foods associated with a larger reduction in disease risk for one health outcome are often associated with larger reductions in disease risk for other health outcomes. Likewise, foods with lower impacts on one metric of environmental harm tend to have lower impacts on others. Additionally, of the foods associated with improved health (whole grain cereals, fruits, vegetables, legumes, nuts, olive oil, and fish), all except fish have among the lowest environmental impacts, and fish has markedly lower impacts than red meats and processed meats. Foods associated with the largest negative environmental impacts—unprocessed and processed red meat—are consistently associated with the largest increases in disease risk. Thus, dietary transitions toward greater consumption of healthier foods would generally improve environmental sustainability, although processed foods high in sugars harm health but can have relatively low environmental impacts. These findings could help consumers, policy makers, and food companies to better understand the multiple health and environmental implications of food choices. |
| Date | 2019-11-12 |
| Library Catalog | pnas.org (Atypon) |
| URL | https://www.pnas.org/doi/full/10.1073/pnas.1906908116 |
| Accessed | 5/3/2025, 7:32:27 AM |
| Extra | Publisher: Proceedings of the National Academy of Sciences |
| Volume | 116 |
| Pages | 23357-23362 |
| Publication | Proceedings of the National Academy of Sciences |
| DOI | 10.1073/pnas.1906908116 |
| Issue | 46 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Conference Paper |
|---|---|
| Author | Emily M. Bender |
| Author | Timnit Gebru |
| Author | Angelina McMillan-Major |
| Author | Shmargaret Shmitchell |
| Abstract | The past 3 years of work in NLP have been characterized by the development and deployment of ever larger language models, especially for English. BERT, its variants, GPT-2/3, and others, most recently Switch-C, have pushed the boundaries of the possible both through architectural innovations and through sheer size. Using these pretrained models and the methodology of fine-tuning them for specific tasks, researchers have extended the state of the art on a wide array of tasks as measured by leaderboards on specific benchmarks for English. In this paper, we take a step back and ask: How big is too big? What are the possible risks associated with this technology and what paths are available for mitigating those risks? We provide recommendations including weighing the environmental and financial costs first, investing resources into curating and carefully documenting datasets rather than ingesting everything on the web, carrying out pre-development exercises evaluating how the planned approach fits into research and development goals and supports stakeholder values, and encouraging research directions beyond ever larger language models. |
| Date | March 1, 2021 |
| Short Title | On the Dangers of Stochastic Parrots |
| Library Catalog | ACM Digital Library |
| URL | https://dl.acm.org/doi/10.1145/3442188.3445922 |
| Accessed | 1/20/2025, 1:00:00 AM |
| Place | New York, NY, USA |
| Publisher | Association for Computing Machinery |
| ISBN | 978-1-4503-8309-7 |
| Pages | 610–623 |
| Series | FAccT '21 |
| Proceedings Title | Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency |
| DOI | 10.1145/3442188.3445922 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Web Page |
|---|---|
| Author | Costs of War Project (Brown University) |
| Abstract | The Costs of War Project is a team of 35 scholars, legal experts, human rights practitioners, and physicians, which began its work in 2011. We use research and a public website to facilitate debate about the costs of the post-9/11 wars in Iraq, Afghanistan, and Pakistan. |
| Language | en |
| URL | https://watson.brown.edu/costsofwar/papers/ClimateChangeandCostofWar |
| Accessed | 5/3/2025, 7:32:37 AM |
| Website Title | The Costs of War |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Conference Paper |
|---|---|
| Author | Alexandra Sasha Luccioni |
| Author | Yacine Jernite |
| Author | Emma Strubell |
| Abstract | Recent years have seen a surge in the popularity of commercial AI products based on generative, multi-purpose AI systems promising a unified approach to building machine learning (ML) models into technology. However, this ambition of ``generality'' comes at a steep cost to the environment, given the amount of energy these systems require and the amount of carbon that they emit. In this work, we propose the first systematic comparison of the ongoing inference cost of various categories of ML systems, covering both task-specific (i.e. finetuned models that carry out a single task) and `general-purpose' models, (i.e. those trained for multiple tasks). We measure deployment cost as the amount of energy and carbon required to perform 1,000 inferences on representative benchmark dataset using these models. We find that multi-purpose, generative architectures are orders of magnitude more expensive than task-specific systems for a variety of tasks, even when controlling for the number of model parameters. We conclude with a discussion around the current trend of deploying multi-purpose generative ML systems, and caution that their utility should be more intentionally weighed against increased costs in terms of energy and emissions. All the data from our study can be accessed via an interactive demo to carry out further exploration and analysis. |
| Date | 2024-06-03 |
| Short Title | Power Hungry Processing |
| Library Catalog | arXiv.org |
| URL | http://arxiv.org/abs/2311.16863 |
| Accessed | 5/4/2025, 8:40:56 AM |
| Extra | arXiv:2311.16863 [cs] |
| Pages | 85-99 |
| Proceedings Title | The 2024 ACM Conference on Fairness, Accountability, and Transparency |
| DOI | 10.1145/3630106.3658542 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Max J. Krause |
| Author | Thabet Tolaymat |
| Abstract | There are now hundreds of cryptocurrencies in existence and the technological backbone of many of these currencies is blockchain—a digital ledger of transactions. The competitive process of adding blocks to the chain is computation-intensive and requires large energy input. Here we demonstrate a methodology for calculating the minimum power requirements of several cryptocurrency networks and the energy consumed to produce one US dollar’s (US$) worth of digital assets. From 1 January 2016 to 30 June 2018, we estimate that mining Bitcoin, Ethereum, Litecoin and Monero consumed an average of 17, 7, 7 and 14 MJ to generate one US$, respectively. Comparatively, conventional mining of aluminium, copper, gold, platinum and rare earth oxides consumed 122, 4, 5, 7 and 9 MJ to generate one US$, respectively, indicating that (with the exception of aluminium) cryptomining consumed more energy than mineral mining to produce an equivalent market value. While the market prices of the coins are quite volatile, the network hashrates for three of the four cryptocurrencies have trended consistently upward, suggesting that energy requirements will continue to increase. During this period, we estimate mining for all 4 cryptocurrencies was responsible for 3–15 million tonnes of CO2 emissions. |
| Date | 2018-11 |
| Language | en |
| Library Catalog | www.nature.com |
| URL | https://www.nature.com/articles/s41893-018-0152-7 |
| Accessed | 5/3/2025, 7:34:07 AM |
| Rights | 2018 The Author(s), under exclusive licence to Springer Nature Limited |
| Extra | Publisher: Nature Publishing Group |
| Volume | 1 |
| Pages | 711-718 |
| Publication | Nature Sustainability |
| DOI | 10.1038/s41893-018-0152-7 |
| Issue | 11 |
| Journal Abbr | Nat Sustain |
| ISSN | 2398-9629 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Bill Tomlinson |
| Author | Rebecca W. Black |
| Author | Donald J. Patterson |
| Author | Andrew W. Torrance |
| Abstract | As AI systems proliferate, their greenhouse gas emissions are an increasingly important concern for human societies. In this article, we present a comparative analysis of the carbon emissions associated with AI systems (ChatGPT, BLOOM, DALL-E2, Midjourney) and human individuals performing equivalent writing and illustrating tasks. Our findings reveal that AI systems emit between 130 and 1500 times less CO2e per page of text generated compared to human writers, while AI illustration systems emit between 310 and 2900 times less CO2e per image than their human counterparts. Emissions analyses do not account for social impacts such as professional displacement, legality, and rebound effects. In addition, AI is not a substitute for all human tasks. Nevertheless, at present, the use of AI holds the potential to carry out several major activities at much lower emission levels than can humans. |
| Date | 2024-02-14 |
| Language | en |
| Library Catalog | www.nature.com |
| URL | https://www.nature.com/articles/s41598-024-54271-x |
| Accessed | 5/5/2025, 7:08:03 AM |
| Rights | 2024 The Author(s) |
| Extra | Publisher: Nature Publishing Group |
| Volume | 14 |
| Pages | 3732 |
| Publication | Scientific Reports |
| DOI | 10.1038/s41598-024-54271-x |
| Issue | 1 |
| Journal Abbr | Sci Rep |
| ISSN | 2045-2322 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Christian Stoll |
| Author | Lena Klaaßen |
| Author | Ulrich Gallersdörfer |
| Abstract | Participation in the Bitcoin blockchain validation process requires specialized hardware and vast amounts of electricity, which translates into a significant carbon footprint. Here, we demonstrate a methodology for estimating the power consumption associated with Bitcoin’s blockchain based on IPO filings of major hardware manufacturers, insights on mining facility operations, and mining pool compositions. We then translate our power consumption estimate into carbon emissions, using the localization of IP addresses. We determine the annual electricity consumption of Bitcoin, as of November 2018, to be 45.8 TWh and estimate that annual carbon emissions range from 22.0 to 22.9 MtCO2. This means that the emissions produced by Bitcoin sit between the levels produced by the nations of Jordan and Sri Lanka, which is comparable to the level of Kansas City. With this article, we aim to gauge the external costs of Bitcoin and inform the broader debate on the costs and benefits of cryptocurrencies. |
| Date | 2019-07-17 |
| Library Catalog | ScienceDirect |
| URL | https://www.sciencedirect.com/science/article/pii/S2542435119302557 |
| Accessed | 5/5/2025, 8:24:49 AM |
| Volume | 3 |
| Pages | 1647-1661 |
| Publication | Joule |
| DOI | 10.1016/j.joule.2019.05.012 |
| Issue | 7 |
| Journal Abbr | Joule |
| ISSN | 2542-4351 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Blog Post |
|---|---|
| Author | Chris Pointon |
| Abstract | An estimate of the carbon emissions from OpenAI’s ChatGPT chatbot service |
| Date | 2023-04-19T15:13:42.548Z |
| Language | en |
| URL | https://medium.com/@chrispointon/the-carbon-footprint-of-chatgpt-e1bc14e4cc2a |
| Accessed | 5/5/2025, 7:21:42 AM |
| Blog Title | Medium |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Seth Wynes |
| Author | Kimberly A Nicholas |
| Abstract | Current anthropogenic climate change is the result of greenhouse gas accumulation in the atmosphere, which records the aggregation of billions of individual decisions. Here we consider a broad range of individual lifestyle choices and calculate their potential to reduce greenhouse gas emissions in developed countries, based on 148 scenarios from 39 sources. We recommend four widely applicable high-impact (i.e. low emissions) actions with the potential to contribute to systemic change and substantially reduce annual personal emissions: having one fewer child (an average for developed countries of 58.6 tonnes CO2-equivalent (tCO2e) emission reductions per year), living car-free (2.4 tCO2e saved per year), avoiding airplane travel (1.6 tCO2e saved per roundtrip transatlantic flight) and eating a plant-based diet (0.8 tCO2e saved per year). These actions have much greater potential to reduce emissions than commonly promoted strategies like comprehensive recycling (four times less effective than a plant-based diet) or changing household lightbulbs (eight times less). Though adolescents poised to establish lifelong patterns are an important target group for promoting high-impact actions, we find that ten high school science textbooks from Canada largely fail to mention these actions (they account for 4% of their recommended actions), instead focusing on incremental changes with much smaller potential emissions reductions. Government resources on climate change from the EU, USA, Canada, and Australia also focus recommendations on lower-impact actions. We conclude that there are opportunities to improve existing educational and communication structures to promote the most effective emission-reduction strategies and close this mitigation gap. |
| Date | 2017-07 |
| Language | en |
| Short Title | The climate mitigation gap |
| Library Catalog | Institute of Physics |
| URL | https://dx.doi.org/10.1088/1748-9326/aa7541 |
| Accessed | 5/3/2025, 7:50:15 AM |
| Extra | Publisher: IOP Publishing |
| Volume | 12 |
| Pages | 074024 |
| Publication | Environmental Research Letters |
| DOI | 10.1088/1748-9326/aa7541 |
| Issue | 7 |
| Journal Abbr | Environ. Res. Lett. |
| ISSN | 1748-9326 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Seth Wynes |
| Author | Kimberly A Nicholas |
| Abstract | Current anthropogenic climate change is the result of greenhouse gas accumulation in the atmosphere, which records the aggregation of billions of individual decisions. Here we consider a broad range of individual lifestyle choices and calculate their potential to reduce greenhouse gas emissions in developed countries, based on 148 scenarios from 39 sources. We recommend four widely applicable high-impact (i.e. low emissions) actions with the potential to contribute to systemic change and substantially reduce annual personal emissions: having one fewer child (an average for developed countries of 58.6 tonnes CO2-equivalent (tCO2e) emission reductions per year), living car-free (2.4 tCO2e saved per year), avoiding airplane travel (1.6 tCO2e saved per roundtrip transatlantic flight) and eating a plant-based diet (0.8 tCO2e saved per year). These actions have much greater potential to reduce emissions than commonly promoted strategies like comprehensive recycling (four times less effective than a plant-based diet) or changing household lightbulbs (eight times less). Though adolescents poised to establish lifelong patterns are an important target group for promoting high-impact actions, we find that ten high school science textbooks from Canada largely fail to mention these actions (they account for 4% of their recommended actions), instead focusing on incremental changes with much smaller potential emissions reductions. Government resources on climate change from the EU, USA, Canada, and Australia also focus recommendations on lower-impact actions. We conclude that there are opportunities to improve existing educational and communication structures to promote the most effective emission-reduction strategies and close this mitigation gap. |
| Date | 2017-07 |
| Language | en |
| Short Title | The climate mitigation gap |
| Library Catalog | Institute of Physics |
| URL | https://dx.doi.org/10.1088/1748-9326/aa7541 |
| Accessed | 1/18/2025, 3:50:54 PM |
| Extra | Publisher: IOP Publishing |
| Volume | 12 |
| Pages | 074024 |
| Publication | Environmental Research Letters |
| DOI | 10.1088/1748-9326/aa7541 |
| Issue | 7 |
| Journal Abbr | Environ. Res. Lett. |
| ISSN | 1748-9326 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | D.S. Lee |
| Author | D.W. Fahey |
| Author | A. Skowron |
| Author | M.R. Allen |
| Author | U. Burkhardt |
| Author | Q. Chen |
| Author | S.J. Doherty |
| Author | S. Freeman |
| Author | P.M. Forster |
| Author | J. Fuglestvedt |
| Author | A. Gettelman |
| Author | R.R. De León |
| Author | L.L. Lim |
| Author | M.T. Lund |
| Author | R.J. Millar |
| Author | B. Owen |
| Author | J.E. Penner |
| Author | G. Pitari |
| Author | M.J. Prather |
| Author | R. Sausen |
| Author | L.J. Wilcox |
| Abstract | Global aviation operations contribute to anthropogenic climate change via a complex set of processes that lead to a net surface warming. Of importance are aviation emissions of carbon dioxide (CO2), nitrogen oxides (NOx), water vapor, soot and sulfate aerosols, and increased cloudiness due to contrail formation. Aviation grew strongly over the past decades (1960–2018) in terms of activity, with revenue passenger kilometers increasing from 109 to 8269 billion km yr−1, and in terms of climate change impacts, with CO2 emissions increasing by a factor of 6.8 to 1034 Tg CO2 yr−1. Over the period 2013–2018, the growth rates in both terms show a marked increase. Here, we present a new comprehensive and quantitative approach for evaluating aviation climate forcing terms. Both radiative forcing (RF) and effective radiative forcing (ERF) terms and their sums are calculated for the years 2000–2018. Contrail cirrus, consisting of linear contrails and the cirrus cloudiness arising from them, yields the largest positive net (warming) ERF term followed by CO2 and NOx emissions. The formation and emission of sulfate aerosol yields a negative (cooling) term. The mean contrail cirrus ERF/RF ratio of 0.42 indicates that contrail cirrus is less effective in surface warming than other terms. For 2018 the net aviation ERF is +100.9 milliwatts (mW) m−2 (5–95% likelihood range of (55, 145)) with major contributions from contrail cirrus (57.4 mW m−2), CO2 (34.3 mW m−2), and NOx (17.5 mW m−2). Non-CO2 terms sum to yield a net positive (warming) ERF that accounts for more than half (66%) of the aviation net ERF in 2018. Using normalization to aviation fuel use, the contribution of global aviation in 2011 was calculated to be 3.5 (4.0, 3.4) % of the net anthropogenic ERF of 2290 (1130, 3330) mW m−2. Uncertainty distributions (5%, 95%) show that non-CO2 forcing terms contribute about 8 times more than CO2 to the uncertainty in the aviation net ERF in 2018. The best estimates of the ERFs from aviation aerosol-cloud interactions for soot and sulfate remain undetermined. CO2-warming-equivalent emissions based on global warming potentials (GWP* method) indicate that aviation emissions are currently warming the climate at approximately three times the rate of that associated with aviation CO2 emissions alone. CO2 and NOx aviation emissions and cloud effects remain a continued focus of anthropogenic climate change research and policy discussions., Image 1 , • Global aviation warms Earth's surface through both CO2 and net non-CO2 contributions. • Global aviation contributes a few percent to anthropogenic radiative forcing. • Non-CO2 impacts comprise about 2/3 of the net radiative forcing. • Comprehensive and quantitative calculations of aviation effects are presented. • Data are made available to analyze past, present and future aviation climate forcing. |
| Date | 2021-1-1 |
| Library Catalog | PubMed Central |
| URL | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7468346/ |
| Accessed | 5/5/2025, 7:41:21 AM |
| Extra | PMID: 32895604 PMCID: PMC7468346 |
| Volume | 244 |
| Pages | 117834 |
| Publication | Atmospheric Environment (Oxford, England : 1994) |
| DOI | 10.1016/j.atmosenv.2020.117834 |
| Journal Abbr | Atmos Environ (1994) |
| ISSN | 1352-2310 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Magazine Article |
|---|---|
| Author | Chris Stokel-Walker |
| Abstract | Integrating large language models into search engines could mean a fivefold increase in computing power and huge carbon emissions. |
| Language | en-US |
| Library Catalog | www.wired.com |
| URL | https://www.wired.com/story/the-generative-ai-search-race-has-a-dirty-secret/ |
| Accessed | 5/3/2025, 7:43:59 AM |
| Extra | Section: tags |
| Publication | Wired |
| ISSN | 1059-1028 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |
| Item Type | Journal Article |
|---|---|
| Author | Francesca Harris |
| Author | Cami Moss |
| Author | Edward J M Joy |
| Author | Ruth Quinn |
| Author | Pauline F D Scheelbeek |
| Author | Alan D Dangour |
| Author | Rosemary Green |
| Abstract | Agricultural water requirements differ between foods. Population-level dietary preferences are therefore a major determinant of agricultural water use. The “water footprint” (WF) represents the volume of water consumed in the production of food items, separated by water source; blue WF represents ground and surface water use, and green WF represents rain water use. We systematically searched for published studies using the WF to assess the water use of diets. We used the available evidence to quantify the WF of diets in different countries, and grouped diets in patterns according to study definition. “Average” patterns equated to those currently consumed, whereas “healthy” patterns included those recommended in national dietary guidelines. We searched 7 online databases and identified 41 eligible studies that reported the dietary green WF, blue WF, or total WF (green plus blue) (1964 estimates for 176 countries). The available evidence suggests that, on average, European (170 estimates) and Oceanian (18 estimates) dietary patterns have the highest green WFs (median per capita: 2999 L/d and 2924 L/d, respectively), whereas Asian dietary patterns (98 estimates) have the highest blue WFs (median: 382 L/d per capita). Foods of animal origin are major contributors to the green WFs of diets, whereas cereals, fruits, nuts, and oils are major contributors to the blue WF of diets. “Healthy” dietary patterns (425 estimates) had green WFs that were 5.9% (95% CI: −7.7, −4.0) lower than those of “average” dietary patterns, but they did not differ in their blue WFs. Our review suggests that changes toward healthier diets could reduce total water use of agriculture, but would not affect blue water use. Rapid dietary change and increasing water security concerns underscore the need for a better understanding of the amount and type of water used in food production to make informed policy decisions. |
| Date | 2020-03-01 |
| Short Title | The Water Footprint of Diets |
| Library Catalog | ScienceDirect |
| URL | https://www.sciencedirect.com/science/article/pii/S2161831322002629 |
| Accessed | 1/20/2025, 1:07:31 PM |
| Volume | 11 |
| Pages | 375-386 |
| Publication | Advances in Nutrition |
| DOI | 10.1093/advances/nmz091 |
| Issue | 2 |
| Journal Abbr | Advances in Nutrition |
| ISSN | 2161-8313 |
| Date Added | 5/6/2025, 4:14:43 PM |
| Modified | 5/6/2025, 4:14:43 PM |