Abstract:Efficient collaboration between collaborative machine learning and wireless communication technology, forming a Federated Edge Learning (FEEL), has spawned a series of next-generation intelligent applications. However, due to the openness of network connections, the FEEL framework generally involves hundreds of remote devices (or clients), resulting in expensive communication costs, which is not friendly to resource-constrained FEEL. To address this issue, we propose a distributed approximate Newton-type algorithm with fast convergence speed to alleviate the problem of FEEL resource (in terms of communication resources) constraints. Specifically, the proposed algorithm is improved based on distributed L-BFGS algorithm and allows each client to approximate the high-cost Hessian matrix by computing the low-cost Fisher matrix in a distributed manner to find a better descent direction, thereby speeding up convergence. Second, we prove that the proposed algorithm has linear convergence in strongly convex and non-convex cases and analyze its computational and communication complexity. Similarly, due to the heterogeneity of the connected remote devices, FEEL faces the challenge of heterogeneous data and non-IID (Independent and Identically Distributed) data. To this end, we design a simple but elegant training scheme, namely FedOVA (Federated One-vs-All), to solve the heterogeneous statistical challenge brought by heterogeneous data. In this way, FedOVA first decomposes a multi-class classification problem into more straightforward binary classification problems and then combines their respective outputs using ensemble learning. In particular, the scheme can be well integrated with our communication efficient algorithm to serve FEEL. Numerical results verify the effectiveness and superiority of the proposed algorithm.

Federated Learning Meets Multi-objective Optimization

Optimize Scheduling of Federated Learning on Battery-powered Mobile Devices

FedDGP: Disentangling Global and Personal Models for Federated Learning

Federated Multi-Objective Learning

A Field Guide to Federated Optimization

Federated Optimization Via Knowledge Codistillation.

Multi-objective Evolutionary Federated Learning

Federated Optimization in Heterogeneous Networks

Accelerated Federated Learning with Decoupled Adaptive Optimization

A Federated Learning Framework for Nonconvex-PL Minimax Problems

Joint Optimization of Resource Allocation and Data Selection for Fast and Cost-Efficient Federated Edge Learning

Federated Learning for Non-IID Data Via Unified Feature Learning and Optimization Objective Alignment

Optimizing Privacy, Utility and Efficiency in Constrained Multi-Objective Federated Learning

Federated Communication-Efficient Multi-Objective Optimization

Multi-Model Federated Learning

Fairness and Accuracy in Federated Learning

Multi-Dimensional Fair Federated Learning

Resource-constrained Federated Learning with Heterogeneous Data: Formulation and Analysis

FedGrad: Optimisation in Decentralised Machine Learning

FedDANE: A Federated Newton-Type Method

Fairness-aware Federated Minimax Optimization with Convergence Guarantee