Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
80 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Towards Environmentally Equitable AI via Geographical Load Balancing (2307.05494v2)

Published 20 Jun 2023 in cs.AI and cs.CY

Abstract: Fueled by the soaring popularity of large language and foundation models, the accelerated growth of AI models' enormous environmental footprint has come under increased scrutiny. While many approaches have been proposed to make AI more energy-efficient and environmentally friendly, environmental inequity -- the fact that AI's environmental footprint can be disproportionately higher in certain regions than in others -- has emerged, raising social-ecological justice concerns. This paper takes a first step toward addressing AI's environmental inequity by balancing its regional negative environmental impact. Concretely, we focus on the carbon and water footprints of AI model inference and propose equity-aware geographical load balancing (GLB) to explicitly address AI's environmental impacts on the most disadvantaged regions. We run trace-based simulations by considering a set of 10 geographically-distributed data centers that serve inference requests for a large language AI model. The results demonstrate that existing GLB approaches may amplify environmental inequity while our proposed equity-aware GLB can significantly reduce the regional disparity in terms of carbon and water footprints.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (79)
  1. Deep Learning. MIT Press, 2016.
  2. Tackling climate change with machine learning. ACM Comput. Surv., 55(2), feb 2022.
  3. Language models are few-shot learners. In H. Larochelle, M. Ranzato, R. Hadsell, M.F. Balcan, and H. Lin, editors, Advances in Neural Information Processing Systems, volume 33, pages 1877–1901. Curran Associates, Inc., 2020.
  4. The carbon footprint of machine learning training will plateau, then shrink. Computer, 55(7):18–28, 2022.
  5. Green ai. Commun. ACM, 63(12):54–63, nov 2020.
  6. Energy and policy considerations for deep learning in NLP. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 3645–3650, Florence, Italy, July 2019. Association for Computational Linguistics.
  7. Towards the systematic reporting of the energy and carbon footprints of machine learning. J. Mach. Learn. Res., 21(1), jan 2020.
  8. Estimating the carbon footprint of bloom, a 176b parameter language model. In arXiv:2211.02001, 2022.
  9. Junkyard computing: Repurposing discarded smartphones to minimize carbon. In Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2, ASPLOS 2023, page 400–412, New York, NY, USA, 2023. Association for Computing Machinery.
  10. OECD. Measuring the environmental impacts of artificial intelligence compute and applications: The AI footprint. OECD Digital Economy Papers, (341), 2022.
  11. Act: Designing sustainable computer systems with an architectural carbon modeling tool. In Proceedings of the 49th Annual International Symposium on Computer Architecture, ISCA ’22, page 784–799, New York, NY, USA, 2022. Association for Computing Machinery.
  12. Steven Gonzalez Monserrate. The Cloud Is Material: On the Environmental Impacts of Computation and Data Storage. MIT Case Studies in Social and Ethical Responsibilities of Computing, (Winter 2022), January 2022. https://mit-serc.pubpub.org/pub/the-cloud-is-material.
  13. LaMDA: Language models for dialog applications, 2022.
  14. Making AI less “thirsty”: Uncovering and addressing the secret water footprint of ai models, 2023.
  15. Local carbon policy. Working Paper 30027, National Bureau of Economic Research, May 2022.
  16. Mark Z. Jacobson. Enhancement of local air pollution by urban CO2 domes. Environmental Science & Technology, 44(7):2497–2502, 2010. PMID: 20218542.
  17. U.S. Environmental Protection Agency. About the U.S. electricity system and its impact on the environment.
  18. Consumptive water use for US power production. National Renewable Energy Laboratory Technical Paper (TP-550-33905), December 2003.
  19. Philipp Hacker. Sustainable AI regulation. In Privacy Law Scholars Conference, 2023, https://arxiv.org/abs/2306.00292.
  20. Sustainable AI: Environmental implications, challenges and opportunities. In Proceedings of Machine Learning and Systems, volume 4, pages 795–813, 2022.
  21. Deepspeed-MOE: Advancing mixture-of-experts inference and training to power next-generation AI scale. In ICML, 2022.
  22. Frugalgpt: How to use large language models while reducing cost and improving performance, 2023.
  23. AutoDNNchip: An automated DNN chip predictor and builder for both FPGAs and ASICs. In FPGA, 2020.
  24. Accelerator-aware neural network design using AutoML. arXiv preprint arXiv:2003.02838, 2020.
  25. Metrics for sustainability in data centers. In HotCarbon, 2022.
  26. Enabling sustainable clouds: The case for virtualizing the energy system. In Proceedings of the ACM Symposium on Cloud Computing, SoCC ’21, page 350–358, New York, NY, USA, 2021. Association for Computing Machinery.
  27. Carbon explorer: A holistic framework for designing carbon aware datacenters. In Proceedings of the 28th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2, ASPLOS 2023, page 118–132, New York, NY, USA, 2023. Association for Computing Machinery.
  28. Google. Environmental report, 2022, https://sustainability.google/reports/.
  29. Meta. Sustainability report, 2021, https://sustainability.fb.com/.
  30. Google. Water commitments, 2023, https://sustainability.google/commitments/water/.
  31. Meta. Sustainability — water, https://sustainability.fb.com/water/.
  32. U.S. Department of Engergy. What is environmental justice?.
  33. Exploiting spatio-temporal diversity for water saving in geo-distributed data centers. IEEE Transactions on Cloud Computing, 6(3):734–746, 2018.
  34. Water-energy tradeoffs in data centers: A case study in hot-arid climates. Resources, Conservation and Recycling, 181:106194, 2022.
  35. Minimax group fairness: Algorithms and experiments. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, AIES ’21, page 66–76, New York, NY, USA, 2021. Association for Computing Machinery.
  36. A review on fairness in machine learning. ACM Comput. Surv., 55(3), feb 2022.
  37. Reuben Binns. On the apparent conflict between individual and group fairness. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, FAT* ’20, page 514–524, New York, NY, USA, 2020. Association for Computing Machinery.
  38. Fairness through awareness. In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference, ITCS ’12, page 214–226, New York, NY, USA, 2012. Association for Computing Machinery.
  39. Fairness in Learning-Based Sequential Decision Algorithms: A Survey, pages 525–555. Springer International Publishing, Cham, 2021.
  40. UNESCO. Recommendation on the ethics of artificial intelligence. In Policy Recommendation, 2022.
  41. Algorithms as social-ecological-technological systems: an environmental justice lens on algorithmic audits. In Proceedings of the 2023 Conference on Fairness, Accountability, and Transparency, FAccT ’23, New York, NY, USA, 2023. Association for Computing Machinery.
  42. U.S. White House Office of Science and Technology Policy. Climate and energy implications of crypto-assets in the United States, September 2022.
  43. Chasing carbon: The elusive environmental footprint of computing. IEEE Micro, 42(4):37–47, jul 2022.
  44. Carbon-aware computing for datacenters. IEEE Transactions on Power Systems, 38(2):1270–1280, 2023.
  45. Carbon-aware energy capacity planning for datacenters. In MASCOTS, 2012.
  46. Cutting the cost of hosting online services using cloud spot markets. In HPDC, 2015.
  47. Cost minimization using renewable cooling and thermal energy storage in CDNs. In ICAC, 2015.
  48. Balance your bids before your bits: The economics of geographic load-balancing. In e-Energy, 2014.
  49. Maximizing the revenues of data centers in regulation market by coordinating with electric vehicles. to appear in Sustainable Computing: Informatics and Systems, 2014.
  50. Reducing electricity cost: Optimization of distributed internet data centers in a multi-electricity-market environment. In INFOCOM, 2010.
  51. Greening geographical load balancing. In SIGMETRICS, 2011.
  52. It’s not easy being green. SIGCOMM Comput. Commun. Rev., 2012.
  53. Cutting the electric bill for internet-scale systems. In Proceedings of the ACM SIGCOMM 2009 Conference on Data Communication, SIGCOMM ’09, page 123–134, New York, NY, USA, 2009. Association for Computing Machinery.
  54. Reducing electricity cost through virtual machine placement in high performance computing clouds. In SuperComputing, 2011.
  55. Carbon emissions and large neural network training, 2021.
  56. iSwitch: Coordinating and optimizing renewable energy powered server clusters. In ISCA, 2012.
  57. Shefy Manayil Kareem. Introducing critical new water data capabilities in Microsoft Cloud for Sustainability, 2023.
  58. Treehouse: A case for carbon-aware datacenter software. In HotCarbon, 2022.
  59. Microsoft. Carbon-aware computing: Measuring and reducing the carbon footprint associated with software in execution. In Whitepaper, 2023.
  60. Fair resource allocation in federated learning. In International Conference on Learning Representations, 2020.
  61. Learning fair representations. In Sanjoy Dasgupta and David McAllester, editors, Proceedings of the 30th International Conference on Machine Learning, volume 28 of Proceedings of Machine Learning Research, pages 325–333, Atlanta, Georgia, USA, 17–19 Jun 2013. PMLR.
  62. Bias in computer systems. ACM Trans. Inf. Syst., 14(3):330–347, jul 1996.
  63. Minimax pareto fairness: A multi objective perspective. In Proceedings of the 37th International Conference on Machine Learning, ICML’20. JMLR.org, 2020.
  64. Discrimination-aware data mining. In Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’08, page 560–568, New York, NY, USA, 2008. Association for Computing Machinery.
  65. How do fairness definitions fare? testing public attitudes towards three algorithmic definitions of fairness in loan allocations. Artificial Intelligence, 283:103238, 2020.
  66. Fast&fair: Training acceleration and bias mitigation for GNNs. Transactions on Machine Learning Research, 2023.
  67. The environmental footprint of data centers in the United States. Environmental Research Letters, 16(6):064017, 2021.
  68. Water-constrained geographic load balancing in data centers. IEEE Trans. Cloud Computing, 2015.
  69. United States data center energy usage report. Lawrence Berkeley National Laboratory, Berkeley, California. LBNL-1005775, 2016.
  70. The Green Grid. Water usage effectiveness (WUE): A green grid data center sustainability metric. Whitepaper, 2011.
  71. L. Tassiulas and S. Sarkar. Maxmin fair scheduling in wireless networks. In Proceedings 21st Annual Joint Conference of the IEEE Computer and Communications Societies, volume 2, pages 763–772 vol.2, 2002.
  72. Capping the brown energy consumption of internet services at low cost. In IGCC, 2010.
  73. Deep compression: Compressing deep neural networks with pruning, trained quantization and huffman coding. In ICLR, 2016.
  74. International Energy Agency (IEA). Data and statistics, https://www.iea.org/data-and-statistics.
  75. U.S. Energy Information Administration. Open data, https://www.eia.gov/opendata/.
  76. Singapore Public Utilities Board. Water efficiency benchmarks.
  77. Iowa State University. Iowa environmental mesonet, https://mesonet.agron.iastate.edu/.
  78. D. Meyer and D. Thevenard. PsychroLib: A library of psychrometric functions to calculate thermodynamic properties of air. Journal of Open Source Software, 4(33):1137, 2019.
  79. A review of operational water consumption and withdrawal factors for electricity generating technologies. NREL Tech. Report: NREL/TP-6A20-50900, 2011.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Pengfei Li (185 papers)
  2. Jianyi Yang (37 papers)
  3. Adam Wierman (132 papers)
  4. Shaolei Ren (56 papers)
Citations (6)
View on Semantic Scholar