Abstract:With rapid advancement of Internet of Things (IoT) and social networking applications generating large amounts of data at or close to the network edge, Mobile Edge Computing (MEC) has naturally been proposed to bring model training closer to where data is produced. However, there still exists privacy concern since typical MEC frameworks need to transfer sensitive data resources from data collection end devices/clients to MEC server. So the concept of Federated Learning (FL) has been introduced which supports privacy-preserved collaborative machine learning involving multiple clients coordinated by the MEC server without centralizing the private data. Unfortunately, FL is prone to multiple challenges: 1) systems heterogeneity between clients causes straggler issue, and 2) statistical heterogeneity between clients brings about objective inconsistency problem, both of which may lead to a significant slow-down in the convergence speed in heterogeneous MEC environment. In this paper, we propose a novel framework, FedAda (Federated Adaptive Training), that incorporates systems capabilities and data characteristics of the clients to adaptively assign appropriate workload to each client. The key idea is that instead of running a fixed number of local training iterations as in Federated Averaging (FedAvg), our algorithm adopts an adaptive workload assignment strategy by minimizing the runtime gap between clients and maximizing convergence gain in heterogeneous MEC environment. Moreover, we design a light mechanism extending FedAda to accelerate the convergence speed by further fine-tuning the workload assignment based on the global convergence status in each communication round. We evaluate FedAda on CIFAR-10 dataset to explore the performance of the algorithm in the simulated heterogeneous MEC environment. Experimental results show that FedAda is able to assign appropriate amount of workload to each client and substantially reduces the convergence time by up to 49.5% compared to FedAvg in heterogeneous MEC environment. In addition, we demonstrate that fine-tuning the workload assignment can help FedAda improve the learning performance in heterogeneous mobile edge computing environment.

Federated learning with workload-aware client scheduling in heterogeneous systems

FedSAE: A Novel Self-Adaptive Federated Learning Framework in Heterogeneous Systems

FedSA: A Semi-Asynchronous Federated Learning Mechanism in Heterogeneous Edge Computing

FedSCS: Client Selection for Federated Learning under System Heterogeneity and Client Fairness with a Stackelberg Game Approach.

AsyncFedED: Asynchronous Federated Learning with Euclidean Distance Based Adaptive Weight Aggregation

Latency Aware Semi-synchronous Client Selection and Model Aggregation for Wireless Federated Learning

An Aggregation-Free Federated Learning for Tackling Data Heterogeneity

FED-SDS: Adaptive Structured Dynamic Sparsity for Federated Learning under Heterogeneous Clients

FedAda: Fast-convergent adaptive federated learning in heterogeneous mobile edge computing environment

FedAEB: Deep Reinforcement Learning Based Joint Client Selection and Resource Allocation Strategy for Heterogeneous Federated Learning

FedRich: Towards Efficient Federated Learning for Heterogeneous Clients Using Heuristic Scheduling

Federated learning based on stratified sampling and regularization

Tackling Data-Heterogeneity Variations in Federated Learning Via Adaptive Aggregate Weights

Federated Learning Client Selection Mechanism under System and Data Heterogeneity

Federated Dynamic Client Selection for Fairness Guarantee in Heterogeneous Edge Computing

Heterogeneity-Aware Federated Learning with Adaptive Client Selection and Gradient Compression.

FedASA: A Personalized Federated Learning with Adaptive Model Aggregation for Heterogeneous Mobile Edge Computing

On the Power of Adaptive Weighted Aggregation in Heterogeneous Federated Learning and Beyond

An EMD-Based Adaptive Client Selection Algorithm for Federated Learning in Heterogeneous Data Scenarios

AEDFL: Efficient Asynchronous Decentralized Federated Learning with Heterogeneous Devices

FedASMU: Efficient Asynchronous Federated Learning with Dynamic Staleness-aware Model Update