Build a comprehensive benchmark of popular BCI algorithms applied on an extensive list of freely available EEG datasets.
Most of the issues and ideas are discussed on Gitter and during office hours (visio meeting every Thursday at 18:30 GMT+1, you could ask for a link on Gitter channel). The discussion are reported in the dedicated wiki page
- Backend features for dev: pre-commit using black, isort and prettier, with an updating CONTRIBUTING section.
- Including support for more datasets
- Add more classification pipelines
- Up to date and automatically built documentation
- Having a leaderboard that display the score of all classification pipelines on all datasets
- Transfer learning support, we already have learning curve support.
- BIDS compliant formatting
- Pytorch support or other backend, different from vanilla scikit-learn.
- Connect the leaderboard with Papers with code
- Organize code sprint and ML competition
- Support different paradigm (affective BCI)
- Support more than fNIRS and EEG signal
- Organize a resilient and decentralized data sharing, using dvc or [Datalad](https://www.datalad.org/