在组合鲁棒性方面，模块化优于不变性
Modularity Trumps Invariance for Compositional Robustness
AI总结：
针对视觉场景组合鲁棒性需求，构建组合图像分类任务，发现增强不变性提升有限，提出模块化架构实现更优组合鲁棒性，且分布内损坏表征不变性与分布外组合鲁棒性无相关性。
摘要：
默认情况下，神经网络对数据分布的变化不具备鲁棒性。这一点已通过简单的图像损坏（如模糊或添加噪声）导致图像分类性能下降得到证实。目前已有许多方法被提出以缓解这些问题，但在大多数情况下，模型仅在单一损坏类型上接受评估。实际上，视觉空间本质上是组合性的，也就是说，除了对基本损坏的鲁棒性之外，还需要对多种损坏组合的鲁棒性。在本研究中，我们构建了一个组合图像分类任务，在给定几种基本损坏的情况下，要求模型泛化到这些损坏的组合形式，即实现组合鲁棒性。我们通过实验对比了经验风险最小化与构建不变性的成对对比损失，结果与领域泛化中的普遍直觉相反，通过增强不变性仅在组合鲁棒性上取得了边际提升。为了突破不变性的局限，遵循此前提出的“模型架构应反映数据结构”的归纳偏置，我们提出了一种模块化架构，其结构复刻了任务的组合特性。随后我们证明，这种模块化方法始终比非模块化方法实现更优的组合鲁棒性。我们还通过实验发现，“分布内”基本损坏的表征之间的不变性程度，与“分布外”损坏组合的鲁棒性之间不存在相关性。
英文摘要：
By default neural networks are not robust to changes in data distribution. This has been demonstrated with simple image corruptions, such as blurring or adding noise, degrading image classification performance. Many methods have been proposed to mitigate these issues but for the most part models are evaluated on single corruptions. In reality, visual space is compositional in nature, that is, that as well as robustness to elemental corruptions, robustness to compositions of corruptions is also needed. In this work we develop a compositional image classification task where, given a few elemental corruptions, models are asked to generalize to compositions of these corruptions. That is, to achieve compositional robustness. We experimentally compare empirical risk minimization with an invariance building pairwise contrastive loss and, counter to common intuitions in domain generalization, achieve only marginal improvements in compositional robustness by encouraging invariance. To move beyond invariance, following previously proposed inductive biases that model architectures should reflect data structure, we introduce a modular architecture whose structure replicates the compositional nature of the task. We then show that this modular approach consistently achieves better compositional robustness than non-modular approaches. We additionally find empirical evidence that the degree of invariance between representations of 'in-distribution' elemental corruptions fails to correlate with robustness to 'out-of-distribution' compositions of corruptions.
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。
