On-device edge-learning for cardiac abnormality detection using a bio-inspired and spiking shallow network

Abstract

This work presents an on-device edge-learning for cardiac abnormality detection by developing a hybrid and spiking form of 2-Dimensional (time-frequency) Convolutional Long-Short-Term Memory (ConvLSTM2D) with Closed-form Continuous-time (CfC) neural network (sCCfC), which is a bio-inspired shallow network. The model achieves an F1 score and AUROC of 0.82 and 0.91 in cardiac abnormalities detection. These results are comparable to the non-spiking ConvLSTM2D-CfC (ConvCfC) model1. Notably, the sCCfC model demonstrates a significantly higher energy efficiency with an estimated power consumption of 4.68µJ/Inf (per inference) on an emulated Loihi’s neuromorphic chip architecture, in contrast to ConvCfC model’s consumption of 450µJ/Inf on a conventional processor. Additionally, as a proof-of-concept, we deployed the sCCfC model on the conventional and relatively resource-constrained Radxa Zero, which is equipped with Amlogic S905Y2 processor foron-device training, which resulted in performance improvements. After initial training of 2 epochs on a conventional GPU, the F1 score and AUROC improved from 0.46 and 0.65 to 0.56 and 0.73 respectively with 5 additional epochs of on-device training. Furthermore, when presented with a new dataset, the sCCfC model showcases strong out-of-sample generalization capabilities that can constitute a pseudo-perspective test, achieving an F1 score and AUROC of 0.71 and 0.86. The spiking sCCfC also outperforms the non-spiking ConvCfC model in robustness regarding effectively handling missing ECG channels during inference. The model’s efficacy extends to single-lead electrocardiogram (ECG) analysis, demonstrating reasonable accuracy in this context, while the focus of our work has been on the computational and memory complexities of the model.