Abnormality Detection in Time-Series Bio-Signals using Kolmogorov-Arnold Networks for Resource-Constrained Devices

This study utilizes Kolmogorov-Arnold Networks (KANs) to analyze ECG signals for detecting cardiac rhythm abnormalities. These networks show promise in biosignal analysis, especially ECG, due to their flexibility and low parameter needs, making them ideal for wearable devices. The simple KAN model, with one hidden layer of 64 neurons, was trained on the Telehealth Network of Minas Gerais (TNMG) dataset and tested on the Chinese Physiological Signal Challenge (CPSC) 2018 dataset. It achieved an F1-score of 0.75 and AUROC of 0.95 on TNMG, and an F1-score of 0.62 and AUROC of 0.84 on the CPSC dataset. It has also shown resistance to missing data channels by maintaining a reasonable performance, down to only a single lead left of ECG data instead of the initial 12 leads. Compared with traditional Multi-Layer Perceptron (MLP) and Neural Circuit Policy (NCP, aka. Liquid Time Constant Networks), KANs exhibit superior flexibility, adaptability, interpretability, and efficiency. Their small size and low computational needs make them suitable for hardware use, especially in personalized medical devices. Although the simple KANs model currently underperforms in detecting some cardiac abnormalities, there is potential for improvement in future development as this study only tests the model’s basic capabilities.