FjORD: Fair and Accurate Federated Learning under heterogeneous targets with Ordered Dropout

TL;DR

FjORD introduces Ordered Dropout to create nested, adaptable models in federated learning, improving fairness and accuracy by tailoring model size to client capabilities without retraining.

Contribution

The paper proposes Ordered Dropout and a federated learning framework called FjORD, enabling efficient, fair, and accurate training across heterogeneous client systems.

Findings

FjORD outperforms state-of-the-art baselines in diverse modalities.

Ordered Dropout produces nested models without retraining.

FjORD improves fairness and performance in federated learning.

Abstract

Federated Learning (FL) has been gaining significant traction across different ML tasks, ranging from vision to keyboard predictions. In large-scale deployments, client heterogeneity is a fact and constitutes a primary problem for fairness, training performance and accuracy. Although significant efforts have been made into tackling statistical data heterogeneity, the diversity in the processing capabilities and network bandwidth of clients, termed as system heterogeneity, has remained largely unexplored. Current solutions either disregard a large portion of available devices or set a uniform limit on the model's capacity, restricted by the least capable participants. In this work, we introduce Ordered Dropout, a mechanism that achieves an ordered, nested representation of knowledge in deep neural networks (DNNs) and enables the extraction of lower footprint submodels without the need of retraining. We further show that for linear maps our Ordered Dropout is equivalent to SVD. We employ this technique, along with a self-distillation methodology, in the realm of FL in a framework called FjORD. FjORD alleviates the problem of client system heterogeneity by tailoring the model width to the client's capabilities. Extensive evaluation on both CNNs and RNNs across diverse modalities shows that FjORD consistently leads to significant performance gains over state-of-the-art baselines, while maintaining its nested structure.