T-SHIELD regularization

T-SHIELD (Transformation-Selective Hidden Input Evaluation for Learning Dynamics) is a regularization technique that aims to enhance model interpretability while improves model performance. Specifically, T-SHIELD adds a regularization term to the objective function that penalizes if it relies too heavily on a small subset of input features.

You can find the documentation of the package here.

If you use this code or find it useful, please cite our paper:

@article{sevillano2024shield,
  title={SHIELD: A regularization technique for eXplainable Artificial Intelligence},
  author={Sevillano-Garc{\'\i}a, Iv{\'a}n and Luengo, Juli{\'a}n and Herrera, Francisco},
  journal={arXiv preprint arXiv:2404.02611},
  year={2024}
}

Installation

You can install this package by simply use pip:

pip install tshield-xai