Attention Based Joint Learning for Supervised Electrocardiogram Arrhythmia Differentiation with Unsupervised Abnormal Beat Segmentation

Xinrong Hu, long wen, shushui wang, Dongpo Liang, Jian Zhuang, Yiyu Shi

Keywords: interpretability, multitask learning, attention mechanism, electrocardiography

Abstract: Deep learning has shown great promise in arrhythmia classification in electrocar- diogram (ECG). Existing works, when classifying an ECG segment with multiple beats, do not identify the locations of the anomalies, which reduces clinical inter- pretability. On the other hand, segmenting abnormal beats by deep learning usu- ally requires annotation for a large number of regular and irregular beats, which can be laborious, sometimes even challenging, with strong inter-observer variabil- ity between experts. In this work, we propose a method capable of not only dif- ferentiating arrhythmia but also segmenting the associated abnormal beats in the ECG segment. The only annotation used in the training is the type of abnormal beats and no segmentation labels are needed. Imitating human’s perception of an ECG signal, the framework consists of a segmenter and classifier. The segmenter outputs an attention map, which aims to highlight the abnormal sections in the ECG by element-wise modulation. Afterwards, the signals are sent to a classifier for arrhythmia differentiation. Though the training data is only labeled to super- vise the classifier, the segmenter and the classifier are trained in an end-to-end manner so that optimizing classification performance also adjusts how the abnor- mal beats are segmented. Validation of our method is conducted on two dataset. We observe that involving the unsupervised segmentation in fact boosts the clas- sification performance. Meanwhile, a grade study performed by experts suggests that the segmenter also achieves satisfactory quality in identifying abnormal beats, which significantly enhances the interpretability of the classification results.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: This paper presents a joint learning framework for supervised arrhythmia differentiation with unsupervised abnormal heart beat seg mentation on ECG, where the two tasks can benefit from each other.

Reviewed Version (pdf): https://openreview.net/references/pdf?id=HOFhU3Uy5x