Scholarlink

<p>Green Artificial Intelligence (Green AI), as a new generation of intelligent technologies, aims to reduce energy consumption and optimize computational efficiency, opening new horizons in the development of algorithms, hardware, and machine learning-based infrastructures. This review article, with a focus on technological innovations in this field, examines advanced solutions in the design of low-power architectures, model compression algorithms, utilization of specialized hardware such as TPUs and Spiking Neural Networks (SNNs), and the application of edge computing to lower energy costs and enhance efficiency. Moreover, intelligent tools for evaluating energy consumption, carbon footprint, and lifecycle indicators of AI models are introduced and their role in monitoring and optimizing computational resources is analyzed. This study, while identifying the current technological challenges including scalability limitations, lifecycle assessment complexities, and the need for convergence between data-driven approaches, hardware, and software, emphasizes the necessity of developing integrated and multidimensional solutions. Finally, by presenting emerging trends such as self-powered computing, biodegradable hardware, and explainable AI models, the article outlines key pathways toward achieving sustainable and technologically advanced Green AI.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Green Artificial Intelligence: A Comprehensive Review of Metrics, Tools, Challenges, Trends, and Future Prospects

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月1日

查看全文
  • Pejman Peykani,
  • Ali Emrouznejad,
  • Sanly Ghanidel,
  • Iman Javadi-Sisi,
  • Seyedali Mirjalili

摘要

Green Artificial Intelligence (Green AI), as a new generation of intelligent technologies, aims to reduce energy consumption and optimize computational efficiency, opening new horizons in the development of algorithms, hardware, and machine learning-based infrastructures. This review article, with a focus on technological innovations in this field, examines advanced solutions in the design of low-power architectures, model compression algorithms, utilization of specialized hardware such as TPUs and Spiking Neural Networks (SNNs), and the application of edge computing to lower energy costs and enhance efficiency. Moreover, intelligent tools for evaluating energy consumption, carbon footprint, and lifecycle indicators of AI models are introduced and their role in monitoring and optimizing computational resources is analyzed. This study, while identifying the current technological challenges including scalability limitations, lifecycle assessment complexities, and the need for convergence between data-driven approaches, hardware, and software, emphasizes the necessity of developing integrated and multidimensional solutions. Finally, by presenting emerging trends such as self-powered computing, biodegradable hardware, and explainable AI models, the article outlines key pathways toward achieving sustainable and technologically advanced Green AI.