koopcore is a Python library designed for learning linear time-invariant (LTI) predictors of dynamical systems. This library provides tools for fitting linear predictors with estimators based on Koopman Kernels.

Please note that koopcore is currently under highly active development and some parts might still be a work in progress as we continuously add new features and improvements.

Create an environment and install koopcore as a local package.

python -m venv koopcore_env
source koopcore_env/bin/activate
pip install --find-links "https://storage.googleapis.com/jax-releases/jax_cuda_releases.html" -e .

Jupyter notebooks containing the experiments are placed in the experiments directory.

This repository contains an implementation of the paper:

[1] Bevanda, P., Beier, M., Lederer, A., Sosnowski, S., Hüllermeier E., & Hirche, S. "Koopman Kernel Regression" in Advances in Neural Information Processing Systems, 2023 [arxiv]

If you found this software useful for your research, consider citing us.

@inproceedings{KKR_neurips2023,
  title = {Koopman Kernel Regression},
  author = {Bevanda, Petar and Beier, Max and Lederer, Armin and Sosnowski, Stefan and H{\"u}llermeier, Eyke and Hirche, Sandra},
  booktitle = {Advances in Neural Information Processing S
62C7
ystems},
  volume = {37},
  year = {2023},
  eprint = {2305.16215},
}