LatentDR: Improving Model Generalization Through Sample-Aware Latent Degradation and Restoration
This addresses the challenge of domain shift for machine learning practitioners, offering a versatile method that is incremental but effective in enhancing model robustness.
The paper tackles the problem of poor model generalization to unseen domains, especially with limited training data, by introducing a latent augmentation method that degrades and restores samples in latent space, achieving significant improvements on domain generalization benchmarks and medical imaging datasets.
Despite significant advances in deep learning, models often struggle to generalize well to new, unseen domains, especially when training data is limited. To address this challenge, we propose a novel approach for distribution-aware latent augmentation that leverages the relationships across samples to guide the augmentation procedure. Our approach first degrades the samples stochastically in the latent space, mapping them to augmented labels, and then restores the samples from their corrupted versions during training. This process confuses the classifier in the degradation step and restores the overall class distribution of the original samples, promoting diverse intra-class/cross-domain variability. We extensively evaluate our approach on a diverse set of datasets and tasks, including domain generalization benchmarks and medical imaging datasets with strong domain shift, where we show our approach achieves significant improvements over existing methods for latent space augmentation. We further show that our method can be flexibly adapted to long-tail recognition tasks, demonstrating its versatility in building more generalizable models. Code is available at https://github.com/nerdslab/LatentDR.