This repository contains the Python code to reproduce the results of the paper Neural state-space models: Empirical evaluation of uncertainty quantification by Marco Forgione and Dario Piga, accepted for presentation at the 2023 IFAC World Congress.
- torchid: PyTorch implementation of neural state-space model. Adapted from the library https://github.com/forgi86/pytorch-ident developed by the first author.
- examples: experiments of the paper: Wiener-Hammerstein circuit.
The examples discussed in the paper are:
- Wiener-Hammerstein Benchmark: A circuit with Wiener-Hammerstein behavior. Experimental dataset from http://www.nonlinearbenchmark.org
Experiments were performed on a Python 3.10 conda environment with
- numpy
- scipy
- matplotlib
- pandas
- pytorch (version 1.12)
- functorch (version 0.2.0)
These dependencies may be installed through the commands:
conda install numpy scipy pandas matplotlib
conda install pytorch -c pytorch
pip install functorch
To run the software, please make sure that this repository's root folder is added to your PYTHONPATH.