Hybrid Quantum-inspired Kolmogorov-Arnold Networks for Privacy-Aware Federated Biosignal Learning
AuthorsChun-Hua Lin, Samuel Yen-Chi Chen, Yu-Chao Hsu, Kuo-Chung Peng, Jiun-Cheng Jiang, Chi-Sheng Chen, Tai-Yue Li, Nan-Yow Chen, En-Jui Kuo, Hsi-Sheng Goan
Resources
A compact quantum-inspired neural network improves federated arrhythmia detection while reducing communication costs across hospital and wearable-device clients.
Key results
HQKAN uses fewer trainable parameters than the MLP baseline.
FedAvg communication cost reduction achieved by HQKAN.
HQKAN parameter reduction relative to the MLP baseline.
FedAvg communication cost reduction achieved by HQKAN.
HQKAN macro-F1 versus 0.698 for the MLP baseline.
HQKAN macro-F1 versus 0.838 for the MLP baseline.
What the paper found
This paper evaluates a hybrid quantum-inspired Kolmogorov–Arnold Network, or HQKAN, for privacy-aware federated ECG classification, where hospitals and wearable devices keep raw signals local and exchange model updates through FedAvg. HQKAN combines fully connected encoder and decoder layers with a QKAN latent processor using DARUAN trainable quantum-inspired edge functions, replacing a conventional ReLU multilayer perceptron while substantially reducing model size. Experiments use five-class MIT-BIH Arrhythmia and three-class INCART classification across 8, 16, and 32 clients, with IID and Dirichlet non-IID partitions, 30 communication rounds, and 5 local epochs. On MIT-BIH, HQKAN reduces trainable parameters by 37.35% and communication cost by 24.89%; on INCART, the reductions reach 44.81% and 36.41%. Under the most fragmented 32-client non-IID setting, HQKAN reaches a macro-F1 of 0.761 versus 0.698 for the MLP on MIT-BIH, while lowering the Brier score from 0.121 to 0.094. On INCART, it achieves macro-F1 of 0.850 versus 0.838 and lowers the Brier score from 0.040 to 0.032. Using class-frequency-weighted cross-entropy and AdamW, HQKAN also improves Cohen’s kappa, AUROC, AUPRC, and many minority-class sensitivity and precision measures, with robustness increasing as client label distributions become more heterogeneous. Implemented with FlashQKAN in PyTorch, the approach demonstrates that quantum-inspired architecture can improve federated biosignal performance while reducing communication overhead, without requiring quantum hardware.
Original abstract
Electrocardiogram (ECG) recordings are sensitive biomedical data, limiting the ability of hospitals and wearable devices to share raw signals for centralized model training. Federated learning addresses this practical privacy constraint by enabling collaborative model training while keeping raw biosignal data at their respective sources. However, federated ECG classification remains challenging due to limited client-side samples, imbalanced arrhythmia labels, and non-independent and identically distributed (non-IID) data across clients. These constraints require classifiers that are both communication-efficient and robust to cross-client distribution shifts. In this work, we evaluate a hybrid quantum-inspired Kolmogorov-Arnold network (HQKAN) against a multilayer perceptron (MLP) for five-class arrhythmia classification on the MIT-BIH dataset and three-class classification on the INCART dataset under federated averaging (FedAvg). Across multiple client configurations, HQKAN improves most aggregate and minority-class metrics while using 37.35% fewer trainable parameters and reducing communication cost by 24.89% on MIT-BIH; on INCART, it achieves corresponding reductions of 44.81% and 36.41%. These results indicate that HQKAN offers a compact, communication-efficient and robust alternative to the MLP baseline for privacy-aware federated learning on biosignal data.
Read the original paperMore in Neural Networks
Browse all 22 papers →End-to-End Hard-Label Cryptanalytic Model Extraction Using Efficient Sign Recovery
Akira Ito, Takayuki Miura, Yosuke Todo
A new query-efficient technique makes it possible to steal the parameters of small black-box neural networks using only their predicted labels.
Retrieving Individual Stems from Music Mixtures with Slot Embeddings
David Braun, Junyi Fan, Pranay Manocha, Donald S. Williamson, Adam Finkelstein
Stembed lets music producers search for individual instrument sounds hidden inside a full song by representing the mixture as multiple searchable stem-like embeddings.
The Linear Representation Hypothesis Needs a Group Action
Louie Hong Yao, Yuhao Li, Shengchao Liu
This paper argues that claims about linear representations only become meaningful once we specify which transformations leave a representation essentially unchanged.