FedSkipTwin: Digital-Twin-Guided Client Skipping for Communication-Efficient Federated Learning
This addresses bandwidth constraints for mobile and IoT devices in federated learning, presenting an incremental improvement over existing methods.
The paper tackles communication overhead in federated learning by introducing FedSkipTwin, a client-skipping algorithm that uses server-side digital twins to predict client updates, reducing total communication by 12-15.5% and improving model accuracy by up to 0.5 percentage points compared to FedAvg.
Communication overhead remains a primary bottleneck in federated learning (FL), particularly for applications involving mobile and IoT devices with constrained bandwidth. This work introduces FedSkipTwin, a novel client-skipping algorithm driven by lightweight, server-side digital twins. Each twin, implemented as a simple LSTM, observes a client's historical sequence of gradient norms to forecast both the magnitude and the epistemic uncertainty of its next update. The server leverages these predictions, requesting communication only when either value exceeds a predefined threshold; otherwise, it instructs the client to skip the round, thereby saving bandwidth. Experiments are conducted on the UCI-HAR and MNIST datasets with 10 clients under a non-IID data distribution. The results demonstrate that FedSkipTwin reduces total communication by 12-15.5% across 20 rounds while simultaneously improving final model accuracy by up to 0.5 percentage points compared to the standard FedAvg algorithm. These findings establish that prediction-guided skipping is a practical and effective strategy for resource-aware FL in bandwidth-constrained edge environments.