This repository contains code for the paper Using Deep Networks for Scientific Discovery in Physiological Signals
Tom Beer, Bar Eini-Porat, Sebastian Goodfellow, Danny Eytan and Uri Shalit
Proceedings of Machine Learning for Healthcare, 2020
Integrate HSICClassifier
from networks.py
and HSICLoss
from hsic.py
in your classification task
-
Download and preprocess the PhysioNet 2017 data by running
python -m ECG.prepare_dataset
-
Train the main task
python -m ECG.train_main_task
-
To evaluate model validity, you may want to run
python -m ECG.train_independence python -m ECG.train_rep2label
-
To visualize the obtained activations:
python -m ECG.visualize_cam
Follow the same steps as in the ECG experiment above, replacing ECG
with EEG
logo created by Atif Arshad from the Noun Project