How to Cite

If you use fusionlab-learn in your research, projects, or publications, we appreciate citations! Proper acknowledgment helps support the project’s development and allows others to find the work.

Citing the Software

To cite the fusionlab-learn software package itself, please include the version number you used, the authors/contributors, the title, year, and the project’s URL. You can find the installed version using:

import fusionlab
print(fusionlab.__version__)

BibTeX Entry for Software:

@software{fusionlab_software,
  author       = {Kouadio, Kouao Laurent and {FusionLab Contributors}},
  title        = {{FusionLab-Learn: Igniting Next-Gen Fusion Models}},
  url          = {https://github.com/earthai-tech/fusionlab-learn/},
  version      = {X.Y.Z},
  year         = {YYYY}
}

Citing the XTFT Research Paper

The core concepts, design, and evaluation of the Extreme Temporal Fusion Transformer (XTFT) model, a key component of fusionlab-learn, are detailed in the following research paper:

Note

The following citation details are based on a manuscript currently submitted for publication. Please check for updated information (DOI, volume, pages) if citing after publication.

BibTeX Entry for Paper:

@article{kouadio2025xtft,
  author       = {Kouadio, Kouao Laurent and Liu, Zhuo and Liu, Jianxin and Bizimana, Pierre and Liu, Wenxiang},
  title        = {{XTFT: A Next-Generation Temporal Fusion Transformer for Uncertainty-Rich Time Series Forecasting}},
  journal      = {Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year         = {2025},
  note         = {Preprint/Submitted Version}
}

Recommendation

If your work heavily relies on the specific methodologies or findings presented in the XTFT paper, we recommend citing both the software package and the research paper.

Thank you for acknowledging fusionlab-learn!