Data for Learning Missing Physics from Legacy Simulators with Alternating Neural Integrators – Part 1: Selected Experiments

This record is Part 1 of the data for the manuscript Learning missing physics from legacy simulators with alternating neural integrators. The package is prepared for manuscript review and reproducibility. The source code is available at: https://github.com/shanxue-w/ANI This record mainly contains datasets for selected experiments, including the ODE experiments, the FitzHugh–Nagumo experiment, and the Kolmogorov flow experiment. The dataset for each experiment is placed in the corresponding experiment directory under a folder named `dataset`. For a complete reproduction of all manuscript results, users should refer to all three Zenodo records linked below. This Part 1 record serves as the main entry point to the complete reproducibility package. System requirements:- Linux environment recommended; tested on Ubuntu 22.04 LTS- Python 3.11- MATLAB R2024a for generating some datasets and running selected post-processing scripts- CUDA-capable GPU recommended for training ANI models- At least 16 GB RAM recommended for small to moderate experiments Installation:1. Download and unzip this record.2. Navigate to the root directory of the unzipped folder.3. Create and activate a Python virtual environment, for example using `venv` or `conda`.4. Install dependencies and the package by running: pip install -r requirements.txt pip install -e . Some experiments call MATLAB through commands of the form: matlab -batch "script_name" If `matlab -batch` is not available, please run the corresponding MATLAB scripts manually. Directory structure: Pendulum/ dataset/ Lorenz-stenflo/ dataset/ Glycolytic/ dataset/ CompoundPendulum/ dataset/ Morris/ dataset/ Kan/ dataset/ Euler/ dataset/ FitzHugh-Nagumo/ dataset/ Kolmogorov_Flow/ dataset/ setup.py pyproject.toml requirements.txt README.md Note: The Lorenz-Stenflo and Kolmogorov flow experiments were further extended during the major revision. The final versions used for the main manuscript figures are included in Part 3. Additional FitzHugh-Nagumo revision experiments are also included in Part 3 and are used for supplementary results. Related records: Link to Part 2: https://doi.org/10.5281/zenodo.17412698 Link to Part 3: https://doi.org/10.5281/zenodo.19482734 License This repository contains the processed research data. The source code is available from the accompanying GitHub repository: https://github.com/shanxue-w/ANI Source code: MIT License. Processed datasets, figure/source data, tables, and reproducibility materials: Creative Commons Attribution 4.0 International License (CC BY 4.0), unless otherwise stated. Third-party source datasets retain their original licenses and terms of use.