InVAErt networks for amortized inference and identifiability analysis of lumped parameter hemodynamic models

This repository contains pre-trained neural network models, training dataset and all related scripts. For additional information, please refer to the publications below:

1. InVAErt networks for amortized inference and identifiability analysis of lumped parameter hemodynamic models, Guoxiang Grayson Tong, Carlos A. Sing-Long Collao, and Daniele E. Schiavazzi.

2. InVAErt networks: A data-driven framework for model synthesis and identifiability analysis, Guoxiang Grayson Tong, Carlos A. Sing-Long Collao, and Daniele E. Schiavazzi.

Description of the Tools folder

1. DNN_tools.py: common functions for deep neural network modeling.
2. Model.py: neural network modules.
3. NF_tools.py: functions used by the Real-NVP based normalizing flow model.
4. Training_tools.py: training and testing functions of the inVAErt networks.
5. EHR_tools.py: functions used in the study of the EHR dataset.
6. cvsim6_scripts.py: functions used in the study of the CVSim-6 model.
7. plotter.py: common plotting functions.

Description of the Solver folder

1. CVSIM6-Training-Data: dataset used in this paper.
2. CVSim6_stiffness_pictures: results of the CVSim-6 stiffness analysis.
3. ExternalData: the EHR dataset.
4. cvsim6-explicit-RK4.py: explicit RK4 solver for the CVSim-6 system.
5. cvsim6-stiffness_analysis.py : functions for the stiffness analysis of the CVSim-6 system.
6. cvsim6-system-test.py: testing CVSim-6 solver with the reference parameter set.
7. cvsim6-training-data-generator.py: parallel training data generator.
8. cvsim6_plotting_functions.py: plotting functions for the stiffness analysis.
9. cvsim6_simulator.py: an implicit, adaptive solver for the CVSim-6 system.

Description of the Model folder

1. Structural_id_study: pre-trained models for the structural identifiability analysis.
2. EHR: pre-trained models for the study of the EHR dataset.

Recommended dependencies

• Pytorch: 2.4.1
• CUDA: 11.8
• Python: 3.10.12
• numpy: 1.26.4
• scipy: 1.12.0
• matplotlib: 3.9.2
• mpi4py: 4.0.0
• sklearn: 1.5.2

Please stay tuned for the Jupyter Notebook tutorials!