Dealing with the Evil Twins: Improving Random Augmentation by Addressing Catastrophic Forgetting of Diverse Augmentations
This addresses the challenge of improving out-of-distribution generalization in machine learning by enhancing random augmentation, which is incremental as it builds on existing augmentation methods.
The paper tackles the problem of random augmentation's suboptimal generalization by identifying that its stochastic nature causes colliding augmentations that distort learned features, similar to catastrophic forgetting, and proposes a simple solution that improves generalization performance across various single source domain generalization benchmarks.
Data augmentation is a promising tool for enhancing out-of-distribution generalization, where the key is to produce diverse, challenging variations of the source domain via costly targeted augmentations that maximize its generalization effect. Conversely, random augmentation is inexpensive but is deemed suboptimal due to its limited effect. In this paper, we revisit random augmentation and explore methods to address its shortcomings. We show that the stochastic nature of random augmentation can produce a set of colliding augmentations that distorts the learned features, similar to catastrophic forgetting. We propose a simple solution that improves the generalization effect of random augmentation by addressing forgetting, which displays strong generalization performance across various single source domain generalization (sDG) benchmarks.