应对邪恶双胞胎：通过解决多样化增强的灾难性遗忘问题来改进随机增强
Dealing with the Evil Twins: Improving Random Augmentation by Addressing Catastrophic Forgetting of Diverse Augmentations
AI总结：
本文重新审视随机增强，发现其随机性会产生类似灾难性遗忘的冲突增强并扭曲特征，提出通过解决遗忘问题来提升其泛化效果，在多个单源域泛化基准上表现优异。
摘要：
数据增强是提升分布外泛化能力的有力工具，其关键在于通过代价高昂的定向增强产生源域的多样化、高难度变体，从而最大化其泛化效果。相反，随机增强成本低廉，但因效果有限而被认为是次优的。在本文中，我们重新审视随机增强，并探索解决其缺陷的方法。我们表明，随机增强的随机特性会产生一组冲突的增强，从而扭曲已学习的特征，这类似于灾难性遗忘。我们提出了一种简单的解决方案，通过解决遗忘问题来提升随机增强的泛化效果，该方法在多个单源域泛化（sDG）基准上展现出强大的泛化性能。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
