Abstract:Federated learning (FL) is an emerging technique used to preserve the privacy of users’ data in training by a distributed machine learning approach. Previously, a client-edge-cloud hierarchical federated learning (HierFL) system was proposed to reduce the long communication latency between the client and the cloud. However, HierFL performs very poorly at extracting better features from the data in the client, which is also a common drawback of traditional FL architectures. HierFL also constrains the its hierarchy to server-edge-client three levels which is very restrictive. This paper proposes that the specifically designed FL architecture can naturally have an excellent effect on feature extraction. Specifically, this paper proposes a multibranch multilevel federated learning (MBMLFL) framework to perform a better job at feature extraction, and each branch and level has its own self specific effect. The proposed framework is also friendlier as well as further enhances the privacy. By extending FedAvg, we design an M2DCFedAvg algorithm for the framework to optimize its objective function distributedly, and by conducting numerous experiments and analyses, we propose a general epoch selection principle for all the FL methods: the ${1^{T}_{d}}$ -n principle. Experiments with various data distributions are performed on MBMLFL to comprehensively research its characteristics. Moreover, to complete the personalization ability of our framework, we propose a plug-and-play dynamic-adjusting double flow personalization approach (DADFPA) for our MBMLFL method, which will further enhance the ability of the MBMLFL, such that it comprehensively and adaptively satisfies the generalization and personalization needs of clients. Experiments show that the MBMLFL and DADFPA can improve their baselines by an average of 5.26% and 2.24% points, which demonstrates their excellent performance.

PerMl-Fed: enabling personalized multi-level federated learning within heterogenous IoT environments for activity recognition

Multi-level Personalized Federated Learning on Heterogeneous and Long-Tailed Data

Federated Mutual Learning: a Collaborative Machine Learning Method for Heterogeneous Data, Models, and Objectives

Federated mutual learning

Personalized Multi-tier Federated Learning

Realizing the Heterogeneity: A Self-Organized Federated Learning Framework for IoT

MDFL: Model-Distance Federated Learning on Non-IID Data

PervasiveFL: Pervasive Federated Learning for Heterogeneous IoT Systems.

Personalized Federated Learning with Model Interpolation among Client Clusters and Its Application in Smart Home

Correction To: Multibranch Multilevel Federated Learning for a Better Feature Extraction and a Plug-and-play Dynamic-Adjusting Double Flow Personalization Approach

FedMLLM: Federated Fine-tuning MLLM on Multimodal Heterogeneity Data

Perfednilm: a practical personalized federated learning-based non-intrusive load monitoring

Heterogeneous Federated Learning with Cross-layer Model Fusion.

Sparse Federated Learning with Hierarchical Personalization Models.

FedCML: Federated Clustering Mutual Learning with non-IID Data.

AsyFed: Accelerated Federated Learning with Asynchronous Communication Mechanism

FedPartial: Enabling Model-Heterogeneous Federated Learning Via Partial Model Transmission and Aggregation

Robust Multi-model Personalized Federated Learning Via Model Distillation

FlexiFed: Personalized Federated Learning for Edge Clients with Heterogeneous Model Architectures

FLAME: Federated Learning Across Multi-device Environments

Federated Multi-Task Learning under a Mixture of Distributions