AugLoss: A Robust Augmentation-based Fine Tuning Methodology
This work addresses robustness concerns for deep learning models under real-world corruptions, representing an incremental step by combining existing techniques to handle multiple issues at once.
The paper tackles the simultaneous problems of train-time noisy labeling and test-time feature distribution shifts in deep learning models by introducing AugLoss, a methodology that unifies data augmentation and robust loss functions, achieving gains in robustness across varied real-world dataset corruption settings.
Deep Learning (DL) models achieve great successes in many domains. However, DL models increasingly face safety and robustness concerns, including noisy labeling in the training stage and feature distribution shifts in the testing stage. Previous works made significant progress in addressing these problems, but the focus has largely been on developing solutions for only one problem at a time. For example, recent work has argued for the use of tunable robust loss functions to mitigate label noise, and data augmentation (e.g., AugMix) to combat distribution shifts. As a step towards addressing both problems simultaneously, we introduce AugLoss, a simple but effective methodology that achieves robustness against both train-time noisy labeling and test-time feature distribution shifts by unifying data augmentation and robust loss functions. We conduct comprehensive experiments in varied settings of real-world dataset corruption to showcase the gains achieved by AugLoss compared to previous state-of-the-art methods. Lastly, we hope this work will open new directions for designing more robust and reliable DL models under real-world corruptions.