Collaborative Research: Conference: DESC: Type III: Eco Edge - Advancing Sustainable Machine Learning at the Edge

协作研究：会议：DESC：类型 III：生态边缘 - 推进边缘的可持续机器学习

• 批准号: 2342497
• 负责人: Vijay Janapa Reddi
• 金额: $ 5.5万
• 依托单位: Harvard University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2024
• 资助国家: 美国
• 起止时间: 2024-01-01 至 2024-09-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2342497&HistoricalAwards=false
• 关键词: Collaborative; Research; Conference; DESC; Type

The proliferation of edge-computing devices and machine-learning algorithms promises to transform technology infrastructure and enable real-time analytics across sectors like healthcare, manufacturing, and smart cities. However, with tens of billions of edge devices expected by 2030, it is imperative to study their sustainability implications. This workshop brings together over 50 experts across industry, academia, and government to draft strategies for energy-efficient, environmentally-sustainable edge machine learning. Through invited talks and interdisciplinary working groups, participants will identify challenges and opportunities in assessing and minimizing the carbon footprint of edge devices throughout their lifecycle. The workshop will produce actionable recommendations on optimized model design, resource-efficiency benchmarks, policy incentives for sustainability, and more. By taking a holistic approach encompassing technology, metrics, tools, and governance, this effort lays the foundation to make edge machine learning a driver for a circular green economy. The workshop facilitates cutting-edge, collaborative research on sustainable edge machine learning. Technical working groups will investigate methods to improve energy efficiency, minimize electronic waste, and reduce the environmental impact at each stage of edge systems’ lifecycles. Discussions will address designing specialized modeling tools for comprehensive impact assessment, creating realistic scenarios to simulate long-term effects, building emulation platforms to accelerate sustainable design choices, and developing efficiency and carbon-footprint benchmarks tailored to edge machine learning. Workgroups will also explore policy incentives, environmental standards for responsible edge-computing practices, and societal considerations beyond carbon emissions, such as biosphere integrity. The workshop develops actionable strategies for sustainable innovation through data-driven studies and multi-stakeholder dialogue as edge intelligence transforms how we live and work.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

边缘计算设备和机器学习算法的激增有望改变技术基础设施，并实现医疗保健、制造和智能城市等行业的实时分析。然而，预计到 2030 年边缘设备将达到数百亿，这一点势在必行。该研讨会汇集了 50 多位来自行业、学术界和政府的专家，通过受邀演讲和跨学科工作组，起草节能、环境可持续的边缘机器学习战略。评估和最小化该研讨会将通过采取涵盖技术、指标、工具和治理的整体方法，提出关于优化模型设计、资源效率基准、可持续发展政策激励等的可行建议。该研讨会旨在促进可持续边缘机器学习的前沿协作研究，为使边缘机器学习成为循环绿色经济的驱动力奠定基础。边缘系统每个阶段的环境影响讨论还将涉及设计用于全面影响评估的专用建模工具、创建模拟长期影响的现实场景、构建仿真平台以加速可持续设计选择，以及开发适合边缘机器学习的效率和碳足迹基准。探讨负责任的边缘计算实践的政策激励措施、环境标准以及碳排放之外的社会考虑因素，例如双完整性圈。随着边缘智能改变我们的生活方式，研讨会通过数据驱动的研究和多方利益相关者对话制定可持续创新的可行战略。和该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。