인공지능 발전에 따른 기후 변화 대응: 도전과 기회
Artificial Intelligence and Climate Change: Challenges and Opportunities for Mitigation
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
인공지능, 특히 대규모 언어 모형의 발전은 막대한 에너지 소비를 초래해 기후 변화 대응에 새로운 과제를 제시한다. 인공지능 모형의 학습, 추론 및 응용 서비스에는 엄청난 전산 자원과 냉각 시스템이 필요하며, 이는 막대한 전력 소비와 온실가스 배출 증가로 이어질 수 있기 때문이다. 이러한 문제를 해결하기 위해 효율적인시스템 냉각 기술 개발, 에너지 효율적인 전산 설비 개발, 모형의 경량화 및 소형화, 재생 에너지 사용 확대 등 다양한 노력이 필요하다. 또한, 인공지능 기술 자체를 활용해 기후 변화 완화 방안을 연구하고 개발하는 노력도 중요하다. 인공지능은기후 변화 대응에 도전과 기회를 동시에 제공한다. 모형의 에너지 효율성을 높이고재생 에너지 사용을 확대하는 노력을 통해 인공지능이 기후 변화 대응에 긍정적으로 기여할 수 있도록 노력해야 한다. 물론, 데이터센터 냉각 기술의 환경 오염 문제와 인공지능 관련 기술의 전자 폐기물 문제도 함께 해결해야 할 과제다.
Abstract
The rapid advancement of artificial intelligence, particularly large language models, has led to a surge in energy consumption, posing new challenges for mitigating climate change. The training, inference, and application services of these models necessitate substantial computing resources and cooling systems, which can result in significant power consumption and increased greenhouse gas emissions. To address these issues, a multifaceted approach is required, encompassing the development of energy-efficient system cooling technologies, energy-efficient computing facilities, smaller and more efficient AI models, and the expanded use of renewable energy. Furthermore, harnessing AI itself for research and development of climate change mitigation strategies is crucial. AI presents both challenges and opportunities in the fight against climate change. We must strive to enhance the energy efficiency of models and expand the utilization of renewable energy to ensure that AI contributes positively to climate action. Additionally, environmental concerns related to data center cooling technologies and e-waste generated by AI technologies must be proactively addressed.
참고문헌 (44)
- 과학기술정보통신부, "대한민국 인공지능 주요 3개국 도약을 위한, 대한민국 대표 인공지능 연구거점 구축 착수", 보도 자료, ~ (2024)
- 초거대 AI 경쟁력 강화 방안, "관계부처 합동", ~ (2023)
- 전 국민 AI 일상화 실행 계획, "관계부처 합동", ~ (2023)
- 데이터센터 수도권 집중 완화 방안, "산업통상자원부", ~ (2023)
- 산업통상자원부, "제11차 전력수급기본계획 실무안", ~ (2024)
- "한국 데이터센터 시장. 세빌스 코리아 리서치 자료" (2024), 세빌스 코리아[URL]
- 10차(2019년 기준) 가구 에너지 패널 조사, "에너지경제연구원", ~ (2022)
- 2023년도 발전 설비 현황, "전력거래소", ~ (2024)
- 한국전력공사, "2023년 12월 전력통계월보(제542호)", ~ (2024)
- Azarifar, M., "Liquid cooling of data centers : A necessity facing challenges", Applied Thermal Engineering, 247, pp.123112~ (2024)
- "Tiny but mighty: The Phi-3 small language models with big potential" (2024)[URL]
- 2024 Global Data Center Investor Intentions Survey, "CBRE", CBRE (Coldwell Banker & Richard Ellis), ~ (2024)
- Energy Institute Statistical Review of World Energy 2024, "EI", Energy Institute (EI), ~ (2024)
- Ember, "Yearly electricity data — Updated on 8 August 2024", ~ (2024)
- "Large-Scale AI Models—August 10, 2024" (2024)[URL]
- "Notable AI Models—August 11, 2024" (2024)[URL]
- 2024 Environmental Report, "Google", ~ (2024)
- "Cloud locations" (2024)[URL]
- "Neural Network Accelerator Comparison" (2024)[URL]
- IEA(International Energy Agency), "Electricity Mid-Year Update: July 2024", IEA Publications, ~ (2024)
- IEA(International Energy Agency), "Monthly Electricity Statistics (data up to April 2024)", IEA Publications, ~ (2024)
- IPCC, "Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change" (2021), Cambridge University Press
- Climate Change 2023: Synthesis Report. Contribution of Working Groups I, II and III to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change, "IPCC", Intergovernmental Panel on Climate Change, ~ (2023)
- Global E-waste Monitor 2024, "ITU(International Telecommunication Union)", United Nations Institute for Training and Research (UNITAR), ~ (2024)
- Jouppi, N., "TPU v4: An optically reconfigurable supercomputer for machine learning with hardware support for embeddings", Proceedings of the 50th Annual International Symposium on Computer Architecture, pp.82~ (2023)
- Kochkov, D., "Neural general circulation models for weather and climate", Nature, pp.1~7 (2024)
- Krizhevsky, A., "ImageNet classification with deep convolutional neural networks", Proceedings of the 25th International Conference on Neural Information Processing Systems, 1 (2012)
- Net Zero Tracker, "Net Zero Tracker", Energy and Climate Intelligence Unit, Data-Driven EnviroLab, NewClimate Institute, Oxford Net Zero, ~ (2024)
- Ortiz, A., "An artificial intelligence dataset for solar energy locations in India", Scientific Data, 9(1), pp.497~510 (2022)
- Ott, B., "Analysis of cooling technologies in the data center sector on the basis of patent applications", Energies, 17(15), pp.3615~ (2024)
- Peckham, O, "PFAS regulations, 3M exit to impact two-phase cooling in HPC", HPCwire, ~ (2023)
- Petrova, V., "Amazon unveils 60-MW solar PPA with S Korea’s SK E&S", Renewables Now, ~ (2023)
- Raina, R., "Large-scale deep unsupervised learning using graphics processors", Proceedings of the 26th Annual International Conference on Machine Learning, pp.873~880 (2009), Association for Computing Machinery
- Sevilla, J., "Compute trends across three eras of machine learning", 2022 International Joint Conference on Neural Networks (IJCNN), pp.1~8 (2022)
- Shao, X., "A review of energy efficiency evaluation metrics for data centers", Energy and Buildings, 271, pp.112308~ (2022)
- Stoddard, I., "Three decades of climate mitigation : Why haven’t we bent the global emissions curve?", Annual Review of Environment and Resources, 46(1), pp.653~689 (2021)
- "The CHIP Dataset" (2024)[URL]
- "TOP500 list —June 2024" (2024)[URL]
- Vaswani, A., "Attention is all you need", Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS 2017), pp.6000~6010 (2017)