Researchers evaluate a hybrid quantum-inspired Kolmogorov-Arnold network (HQKAN) against a standard multilayer perceptron for arrhythmia classification using federated learning on electrocardiogram data. On the MIT-BIH dataset, HQKAN achieved 37.35% fewer trainable parameters and reduced communication cost by 24.89% while improving most performance metrics, with similar efficiency gains observed on the INCART dataset and demonstrated robustness to non-identically-distributed data across federated clients.
