Paper ID: 2303.14537
Deep Augmentation: Self-Supervised Learning with Transformations in Activation Space
Rickard Brüel-Gabrielsson, Tongzhou Wang, Manel Baradad, Justin Solomon
We introduce Deep Augmentation, an approach to implicit data augmentation using dropout or PCA to transform a targeted layer within a neural network to improve performance and generalization. We demonstrate Deep Augmentation through extensive experiments on contrastive learning tasks in NLP, computer vision, and graph learning. We observe substantial performance gains with Transformers, ResNets, and Graph Neural Networks as the underlying models in contrastive learning, but observe inverse effects on the corresponding supervised problems. Our analysis suggests that Deep Augmentation alleviates co-adaption between layers, a form of "collapse." We use this observation to formulate a method for selecting which layer to target; in particular, our experimentation reveals that targeting deeper layers with Deep Augmentation outperforms augmenting the input data. The simple network- and modality-agnostic nature of this approach enables its integration into various machine learning pipelines.
Submitted: Mar 25, 2023