Hybrid Quantum-Inspired Kolmogorov-Arnold Networks for Privacy-Aware Federated Biosignal Learning

A new research paper on arXiv (2608.13914) evaluates a hybrid quantum-inspired Kolmogorov-Arnold network (HQKAN) for federated learning of ECG signals, addressing privacy constraints in sharing raw biomedical data. The study compares HQKAN against a multilayer perceptron (MLP) for arrhythmia classification using the MIT-BIH and INCART datasets under federated averaging (FedAvg). The research tackles challenges such as limited client-side samples, imbalanced labels, and non-IID data across clients, aiming for communication-efficient and robust classifiers. The paper is categorized as a cross-type announcement and was published on arXiv.