Scholarlink

<p>Federated Learning (FL) has become a promising decentralized paradigm in machine learning, but its energy efficiency remains a critical challenge, especially in resource-constrained environments. Existing optimization techniques often improve energy consumption at the cost of model accuracy, and there is a lack of systematic analysis of the trade-offs between these factors. This paper presents a systematic literature review aimed at identifying innovative techniques for improving energy efficiency in FL, while also assessing their impact on model accuracy. The review will explore emerging hardware acceleration strategies, such as TPUs and GPUs, and their role in reducing energy consumption and enhancing scalability in FL deployments. Additionally, the paper will identify key research gaps in balancing energy efficiency, privacy, and scalability in FL frameworks. The findings are expected to provide insights into the most effective methods for optimizing energy use without sacrificing model performance, offering a roadmap for future research and development in this area.</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

Improving energy efficiency and scalable computing in Federated Learning

  • 期刊论文

  • 发表日期：

    2025年11月11日

查看全文
  • Alvina Shahid,
  • Jamil Ahmad,
  • Imran Shafi,
  • Muhammad Muneeb Arshad,
  • Tahir Khurshaid,
  • Hafiz Muhammad Raza ur Rehman,
  • Imran Ashraf

摘要

Federated Learning (FL) has become a promising decentralized paradigm in machine learning, but its energy efficiency remains a critical challenge, especially in resource-constrained environments. Existing optimization techniques often improve energy consumption at the cost of model accuracy, and there is a lack of systematic analysis of the trade-offs between these factors. This paper presents a systematic literature review aimed at identifying innovative techniques for improving energy efficiency in FL, while also assessing their impact on model accuracy. The review will explore emerging hardware acceleration strategies, such as TPUs and GPUs, and their role in reducing energy consumption and enhancing scalability in FL deployments. Additionally, the paper will identify key research gaps in balancing energy efficiency, privacy, and scalability in FL frameworks. The findings are expected to provide insights into the most effective methods for optimizing energy use without sacrificing model performance, offering a roadmap for future research and development in this area.