CADDA: Class-wise Automatic Differentiable Data Augmentation for EEG Signals
This work addresses the problem of cumbersome manual augmentation design for neuroscience signals like EEG, offering a novel automated approach that is incremental but improves efficiency and performance in supervised learning applications such as sleep stage classification.
The paper tackles the challenge of automatically finding effective data augmentation policies for complex data like EEG signals, introducing a differentiable relaxation that achieves optimal performance with faster training than competing methods in class-agnostic settings and outperforms gradient-free methods in class-wise settings.
Data augmentation is a key element of deep learning pipelines, as it informs the network during training about transformations of the input data that keep the label unchanged. Manually finding adequate augmentation methods and parameters for a given pipeline is however rapidly cumbersome. In particular, while intuition can guide this decision for images, the design and choice of augmentation policies remains unclear for more complex types of data, such as neuroscience signals. Besides, class-dependent augmentation strategies have been surprisingly unexplored in the literature, although it is quite intuitive: changing the color of a car image does not change the object class to be predicted, but doing the same to the picture of an orange does. This paper investigates gradient-based automatic data augmentation algorithms amenable to class-wise policies with exponentially larger search spaces. Motivated by supervised learning applications using EEG signals for which good augmentation policies are mostly unknown, we propose a new differentiable relaxation of the problem. In the class-agnostic setting, results show that our new relaxation leads to optimal performance with faster training than competing gradient-based methods, while also outperforming gradient-free methods in the class-wise setting. This work proposes also novel differentiable augmentation operations relevant for sleep stage classification.