Federated Learning (FL) presents a paradigm shift towards distributed model training across isolated data repositories or edge devices without explicit data sharing. Despite of its advantages, FL is inherently less efficient than centralized training models, leading to increased energy consumption and, consequently, higher carbon emissions. In this paper, we propose CAMA, a carbon-aware FL framework, promoting the operation on renewable excess energy and spare computing capacity, aiming to minimize operational carbon emissions. CAMA introduces a dynamic model adaptation strategy which adapts the model sizes based on the availability of energy and computing resources. Ordered dropout is integratged to enable the aggregation with varying model sizes. Empirical evaluations on real-world energy and load traces demonstrate that our method achieves faster convergence and ensures equitable client participation, while scaling efficiently to handle large numbers of clients. The source code of CAMA is available at https://github.com/denoslab/CAMA.

翻译：联邦学习（FL）提出了一种分布式模型训练的新范式，可在隔离的数据存储库或边缘设备上进行训练，而无需显式共享数据。尽管具有诸多优势，联邦学习本质上比集中式训练模型效率更低，导致能耗增加，进而产生更高的碳排放。本文提出CAMA，一种碳感知的联邦学习框架，旨在利用可再生能源过剩能量和空闲计算能力进行运算，以最小化运行碳排放。CAMA引入了一种动态模型适应策略，该策略根据能源和计算资源的可用性调整模型规模。通过集成有序丢弃技术，实现了不同规模模型的聚合。基于真实世界能源与负载轨迹的实证评估表明，我们的方法实现了更快的收敛速度，确保了客户端参与的公平性，并能高效扩展以处理大量客户端。CAMA的源代码可在 https://github.com/denoslab/CAMA 获取。