调研

实验

adaptive Federated learning in resource constrained edge computing systems

• networked prototype system with 5 nodes
• a simulated environment with the number of nodes varying from 5 to 500. The prototype system consists of three Raspberry Pi (version 3) devices and two laptop computers, which are all interconnected via Wi-Fi in an office building. This represents an edge computing environment where the computational capabilities of edge nodes are heterogeneous
• we train each model for a fixed amount of time budget.
• 模拟资源消耗服从高斯分布
• baseline
  • Centralized gradient descent
  • Canonical federated learning approach 规范的联邦学习方法
  • Synchronous distributed gradient descent
• 为了避免与损失计算相关的资源消耗，集中式梯度下降使用最后的模型参数w（T）（而不是wf）作为结果，因为w（T）的收敛可以在集中式情况下得到证实
• SGD中 minibatch
• The models include squared-SVM, linear regression, K-means, and deep convolutional neural networks (CNN).
• 分布式配置
  • In Case 1, each data sample is randomly assigned to a node, thus each node has uniform (but not full) information.
  • In Case 2, all the data samples in each node have the same label.
  • Case 3, each node has the entire dataset (thus full information).
  • In Case 4, data samples with the first half of the labels are distributed to the first half of the nodes as in Case 1; the other samples are distributed to the second half of the nodes as in Case 2. 中和平衡和非平衡的情况

真实实验

模拟实验