Abstract:The core of federated learning (FL) is to transfer data diversity and distribution knowledge of cross-client domains. Although adopted by most FL methods, model-sharing-based communication has limitations such as unstable optimization and lack of reasonable explanation. In this paper, we propose an innovative approach that exploits a highly abstract local compact representation (LCR) as a communication carrier, paving a new feasible path for FL. LCR is not only more intuitive for multiclient joint training but also insensitive to local privacy, particularly in computer vision tasks. The proposed LCR communication-based FL framework aims to improve performance, mitigate negative transfer, and enhance optimization stability. First, based on the domain adaptation theorem, in-depth theoretical proofs guarantee the contribution of representation communication from other domains, which may lead to a tighter generalization error bound of the local domain. Second, inspired by metric learning, a federated version of the triplet (FedTriplet) loss and distribution similarity reweighting aggregation are proposed to fully digest LCR from other clients and realize the LCRC-based FL framework with better explanations. Cross-client FedTriplet transferring redresses the category boundaries in local latent space, resisting overfitting. The modified Wasserstein distance is employed for reweighting aggregation to overcome the negative transfer problem caused by non-i.i.d. factors. Finally, extensive experiments on MNIST/EMNIST and successful iris recognition applications demonstrate that the proposed LCRC framework is superior in terms of accuracy compared to mainstream model-sharing-based FL methods. The visualization results also show significant improvements in the distinguishability of the representation distribution.

Aligning Before Aggregating: Enabling Communication Efficient Cross-Domain Federated Learning Via Consistent Feature Extraction

FedDGP: Disentangling Global and Personal Models for Federated Learning

FedCross: Towards Accurate Federated Learning via Multi-Model Cross-Aggregation

FedEmb: A Vertical and Hybrid Federated Learning Algorithm using Network And Feature Embedding Aggregation

FedDA: Resource-adaptive Federated Learning with Dual-Alignment Aggregation Optimization for Heterogeneous Edge Devices

Enhancing Edge-Assisted Federated Learning with Asynchronous Aggregation and Cluster Pairing

FedFM: Anchor-based Feature Matching for Data Heterogeneity in Federated Learning

Local Learning Matters: Rethinking Data Heterogeneity in Federated Learning

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

An Aggregation-Free Federated Learning for Tackling Data Heterogeneity

Aligning Model Outputs for Class Imbalanced Non-Iid Federated Learning

Cross-Silo Federated Learning Across Divergent Domains with Iterative Parameter Alignment

Bilateral Improvement in Local Personalization and Global Generalization in Federated Learning.

FedFed: Feature Distillation Against Data Heterogeneity in Federated Learning

Personalized Federated Learning with Feature Alignment and Classifier Collaboration

Federated Local Compact Representation Communication: Framework and Application

Fair Federated Learning under Domain Skew with Local Consistency and Domain Diversity

Federated Learning with Additional Mechanisms on Clients to Reduce Communication Costs

Agglomerative Federated Learning: Empowering Larger Model Training via End-Edge-Cloud Collaboration

Utilizing Free Clients in Federated Learning for Focused Model Enhancement

FedCCRL: Federated Domain Generalization with Cross-Client Representation Learning