论文地址:Federated Learning with Non-IID Data 一、 Introduction 介绍 这部分内容先是介绍了FL的由来和发展,简单介绍了Fedavg算法(不了解的小伙伴需要看一下2016年谷歌那篇论文,流程比较简单),说明了一下FL通信的问题和研究,最后引出了FL的Non-IID问题,在一些特定的Non-IID数据集上Fedavg是可以收敛的,但是其他情...
Federated Learning enables clients to train a joint model collaboratively without disclosing raw data. However, learning over non-IID data may raise performance degeneration, which has become a fundamental bottleneck. Despite numerous efforts to address this issue, challenges such as excessive local compu...
我们进一步表明,这种精度降低可以用重量差异来解释,重量差异可以通过每个设备上各个类别的分布与人口分布之间的推土机距离(EMD)进行量化。作为解决方案,我们提出了一种策略,可通过创建在所有边缘设备之间全局共享的一小部分数据来改善对非IID数据的训练。实验表明,对于仅包含5%全局共享数据的CIFAR-10数据集,其准确性可以提...
Federated Learning with Non-IID Data 论文笔记 /104632718 论文通过实验验证了,在non-IID数据中,使用FedAvg算法训练的模型会使准确率降低。 从图中可以看出在non-IID使用FedAvg算法训练的模型准确率有了明显的下降,但是对于IID...因为数据分布的不同。 FedAvg算法训练的模型准确率收到数据分布偏态性的影响。 研究...
Federated Learning with Non-IID Data IID:独立同分布 (idependently and identically distributed, IID) 论文链接 Abstract 联合学习使资源受限的边缘计算设备(例如移动电话和IoT设备)能够学习共享的预测模型,同时将训练数据保持在本地。这种去中心化的训练模型方法提供了隐私,安全性,监管和经济利益。在这项工作中,我们...
《Achieving linear speedup with partial worker participation in non-iid federated learning》 这些方法与我们的方法是兼容的,可以很容易地集成到我们的方法中。 然而,Zhao等[32]的理论表明,参数偏差会累积,导致次优解。 Local Drift in Federated Learning ...
FEDERATED learningSTATISTICSPRIVACYIn Differentially Private Federated Learning (DPFL), gradient clipping and random noise addition disproportionately affect statistically heterogeneous data. As a consequence, DPFL has a disparate impact: the accuracy of models trained with DPFL tends to decrease more on ...
Optimizing Federated Learning on Non-IID Data with Reinforcement Learning 这张示意图和对应的文字描述了一个双深度Q学习网络(Double DQN, DDQN)智能体如何与联邦学习(FL)服务器交互。以下是详细解读: 图中主要组件: 智能体(Agent): 智能体负责根据当前状态选择行动(Action)。它利用一个双深度Q学习网络(DDQN)计...
In Section 3, we present the definition and classification of non-IID data in a federated environment, and we also discuss the different strategies to deal with it. In Section 4, we introduce the Continual Learning framework and the multiple ways data can evolve over time. In Section 5, we...
Decoupled Federated Learning for ASR with Non-IID Data Automatic speech recognition (ASR) with federated learning (FL) makes it possible to leverage data from multiple clients without compromising privacy. The quality of FL-based ASR could be measured by recognition performance, communication and ...