Mother of all BCI Benchmark
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
Short term - what we're working on now
- 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
Medium term - what we're working on next
- 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
Long term - what we're working on in the future
- 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/