Euclidean Alignment improves the performance and convergence speed of deep learning models for EEG decoding across multiple subjects.
Scheduled knowledge distillation improves accuracy and efficiency in lightweight BCIs.