A Gradient Boosting Approach for Training Convolutional and Deep Neural Networks

GB-CNN is the python library for working with Gradient Boosted - Convolutional Neural Networks (GB-CNN) and Gradient Boosted Deep Neural Networks (GB-DNN).

• GB-CNN is designed to implement the Gradient Boosted Convolutional layers for the image datasets.

• GB-DNN is designed to implement the Gradient Boosted Dense layers for the big dimension tabular datasets.

Citing

Please use the following BibTeX or download the bib file for citing the paper in your research.

Or use the CITATION to cite the package and codes.

@ARTICLE{10130606,
  author={Emami, Seyedsaman and Martínez-Muñoz, Gonzalo},
  journal={IEEE Open Journal of Signal Processing}, 
  title={A Gradient Boosting Approach for Training Convolutional and Deep Neural Networks}, 
  year={2023},
  volume={4},
  pages={313-321},
  doi={10.1109/OJSP.2023.3279011}}

License

The package is licensed under the GNU Lesser General Public License v2.1.

Documentation

For more information, please refer to the Wiki.

Development

Our latest algorithm is available on the main branch of the repository.

Related released versions are stored on Releases.

Date-released

03.Feb.2023

Date-updated

27.Feb.2023

Version

0.0.1