Introduction

This repository provides a unified framework for Dynamical Systems Reconstruction (DSR). It integrates established generative recurrent DSR models (Durstewitz (2017), Hess et al. (2023), Brenner et al. (2024)) within the Multimodal Teacher Forcing (MTF) framework (Brenner et al. (2024)) for heterogeneous data. Furthermore, it uses parameter hierarchies designed to enhance model interpretability and enable scaling across extensive datasets (Brenner et al. (2025)). Built with modularity in mind, the framework ensures high extensibility, allowing for the seamless integration of custom models.

Citation

Please cite the following papers when using this framework in your own work:

BibTeX here