Mutual Information-driven Subject-invariant and Class-relevant Deep Representation Learning in BCI

This repository provides the official Pytorch implementation of "Jeon et al., Mutual Information-driven Subject-invariant and Class-relevant Deep Representation Learning in BCI, IEEE-TNNLS, 2021 (DOI: 10.1109/TNNLS.2021.3100583)."

• Contact : E.-J. Jeon ([email protected])
• We propose a novel deep learning framework that learns subject-invariant and class-relevant feature representations in an information-theoretic and end-to-end manner. Our proposed components of feature decomposition and feature enrichment can be naturally plugged into the existing network architectures.

Prerequisites

• Python 3.6+
• PyTorch 1.4.0+

Citation

If you use this code for your research, please cite our paper:

@article{jeon2021mutual,
  title={Mutual information-driven subject-invariant and class-relevant deep representation learning in BCI},
  author={Jeon, Eunjin and Ko, Wonjun and Yoon, Jee Seok and Suk, Heung-Il},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  year={2021},
  publisher={IEEE}
}

Acknowledgements

This work was supported in part by the Institute for Information and Communications Technology Promotion (IITP) grant funded by the Korea Government under Grant 2017-0-00451 (Development of BCI based Brain and Cognitive Computing Technology for Recognizing User’s Intentions using Deep Learning) and in part by the Institute of Information and Communications Technology Planning and Evaluation (IITP) grant funded by the Korea Government (MSIT) under Grant 2019-0-00079 [Department of Artificial Intelligence (Korea University)].