Techniques for visualizing LSTMs applied to electrocardiograms

2017/05/23 by Jos van der Westhuizen, van der Westhuizen, Jos, Joan Lasenby +1 · 1 citation
Computer Science · #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning in Healthcare #Neural Networks and Applications #Time Series Analysis and Forecasting

paper · pdf · doi:10.48550/arxiv.1705.08153

openalex publication_date 2017/05/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper explores four different visualization techniques for long short-term memory (LSTM) networks applied to continuous-valued time series. On the datasets analysed, we find that the best visualization technique is to learn an input deletion mask that optimally reduces the true class score. With a specific focus on single-lead electrocardiograms from the MIT-BIH arrhythmia dataset, we show that salient input features for the LSTM classifier align well with medical theory.

Citations

• PhysioBank, PhysioToolkit, and PhysioNet
• ImageNet classification with deep convolutional neural networks
• Architectural Complexity Measures of Recurrent Neural Networks
• Genetic Architect: Discovering Genomic Structure with Learned Neural\n Architectures
• Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
• RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism

Cited by

• Understanding Memory Modules on Learning Simple Algorithms

Related

• CNNs, LSTMs, and Attention Networks for Pathology Detection in Medical Data
• A Compact LSTM-SVM Fusion Model for Long-Duration Cardiovascular Diseases Detection
• 3-Lead to 12-Lead ECG Reconstruction: A Novel AI-based Spatio-Temporal Method
• xLSTM-ECG: Multi-label ECG Classification via Feature Fusion with xLSTM
• Development of Interpretable Machine Learning Models to Detect Arrhythmia based on ECG Data