[论文解读] How Green Can AI Be? A Study of Trends in Machine Learning Environmental Impacts

Clément Morand, Anne‐Laure Ligozat+1 more|arXiv (Cornell University)|Dec 23, 2024

Air Quality Monitoring and Forecasting被引用 4

一句话总结

本研究分析了2013年至2023年训练机器学习模型及制造图形处理器的环境影响，发现尽管效率有所提升并采取了如将生产转移至低碳地区等策略，硬件制造与训练能耗均呈指数级增长。效率提升所引发的反弹效应——即效率提升促使模型规模扩大——削弱了减排效果，表明若不减少人工智能本身的规模，当前策略无法遏制人工智能日益增长的环境足迹。

ABSTRACT

The compute requirements associated with training Artificial Intelligence (AI) models have increased exponentially over time. Optimisation strategies aim to reduce the energy consumption and environmental impacts associated with AI, possibly shifting impacts from the use phase to the manufacturing phase in the life-cycle of hardware. This paper investigates the evolution of individual graphics cards production impacts and of the environmental impacts associated with training Machine Learning (ML) models over time. We collect information on graphics cards used to train ML models and released between 2013 and 2023. We assess the environmental impacts associated with the production of each card to visualize the trends on the same period. Then, using information on notable AI systems from the Epoch AI dataset we assess the environmental impacts associated with training each system. The environmental impacts of graphics cards production have increased continuously. The energy consumption and environmental impacts associated with training models have increased exponentially, even when considering reduction strategies such as location shifting to places with less carbon intensive electricity mixes. These results suggest that current impact reduction strategies cannot curb the growth in the environmental impacts of AI. This is consistent with rebound effect, where the efficiency increases fuel the creation of even larger models thereby cancelling the potential impact reduction. Furthermore, these results highlight the importance of considering the impacts of hardware over the entire life-cycle rather than the sole usage phase in order to avoid impact shifting. The environmental impact of AI cannot be reduced without reducing AI activities as well as increasing efficiency.

研究动机与目标

• 评估2013年至2023年图形处理器生产过程的环境影响演变。
• 基于真实世界数据，分析同一时期训练机器学习模型的环境足迹。
• 评估当前减排策略（如硬件升级和地理位置转移）的有效性。
• 研究反弹效应和影响转移在抵消人工智能环境效益方面所起的作用。

提出的方法

• 收集2013年至2023年间发布的NVIDIA工作站级图形处理器数据，以评估其生产阶段的环境影响。
• 使用Epoch AI数据集，将特定机器学习模型与训练所用硬件关联，估算训练能耗。
• 应用MLCA（机器学习碳排放评估）方法，量化硬件生命周期中的碳足迹、金属资源耗竭及一次能源需求。
• 基于GPU数量、训练时长和功耗建模训练能耗，并根据电力结构和PUE（电力使用效率）进行调整。
• 评估将训练迁移至低碳电力区域的影响，发现其不足以抵消指数级增长。
• 通过敏感性分析考虑不确定性，包括PUE、训练时长和电力结构等假设。

实验结果

研究问题

• RQ1 2013年至2023年，图形处理器生产的环境影响如何变化？
• RQ2 当前的效率策略在多大程度上减少了机器学习模型训练的碳足迹？
• RQ3 反弹效应是否抵消了硬件效率提升在人工智能训练中的环境效益？
• RQ4 使用阶段向制造阶段的影响转移如何影响人工智能系统的整体环境足迹？
• RQ5 更清洁的电力结构能否完全抵消人工智能训练日益增长的能源需求？

主要发现

• 2013年至2023年，图形处理器生产的环境影响持续上升，主要受复杂度和材料强度增加的驱动。
• 尽管GPU能效有所提升，但训练机器学习模型的能耗和碳足迹在同一时期呈指数级增长。
• 将训练迁移至低碳电力区域等策略因反弹效应而未能遏制碳排放的指数级上升。
• 反弹效应普遍存在：能效提升导致模型规模扩大，整体能耗上升，从而抵消了潜在的环境收益。
• 使用阶段向制造阶段的影响转移显著，凸显了仅关注运行能耗的全生命周期评估的必要性。
• 当前减排策略不足以扭转趋势；唯有减少人工智能活动，才能显著降低环境影响。