解释你的决策：视觉分类中用于泛化的逻辑推理正则化
Interpret Your Decision: Logical Reasoning Regularization for Generalization in Visual Classification
AI总结：
本文提出逻辑推理正则化L-Reg，通过将逻辑分析框架与图像分类结合，降低模型复杂度并提取显著特征，从而在多域泛化和广义类别发现等未见数据场景中显著提升视觉分类的泛化能力。
摘要：
视觉模型在图像分类中表现出色，但难以泛化到未见数据，例如对来自未见域的图像进行分类或发现新类别。在本文中，我们探讨了视觉分类中逻辑推理与深度学习泛化之间的关系。我们推导出一种称为 L-Reg 的逻辑正则化，它将逻辑分析框架与图像分类联系起来。我们的工作表明，L-Reg 在特征分布和分类器权重方面降低了模型的复杂度。具体而言，我们揭示了 L-Reg 带来的可解释性，因为它使模型能够提取显著特征（例如人脸对应于人物）以进行分类。理论分析和实验表明，L-Reg 增强了在各种场景下的泛化能力，包括多域泛化和广义类别发现。在图像跨越未知类别和未见域的复杂现实场景中，L-Reg 持续提升泛化能力，凸显了其实际有效性。
英文摘要：
Vision models excel in image classification but struggle to generalize to unseen data, such as classifying images from unseen domains or discovering novel categories. In this paper, we explore the relationship between logical reasoning and deep learning generalization in visual classification. A logical regularization termed L-Reg is derived which bridges a logical analysis framework to image classification. Our work reveals that L-Reg reduces the complexity of the model in terms of the feature distribution and classifier weights. Specifically, we unveil the interpretability brought by L-Reg, as it enables the model to extract the salient features, such as faces to persons, for classification. Theoretical analysis and experiments demonstrate that L-Reg enhances generalization across various scenarios, including multi-domain generalization and generalized category discovery. In complex real-world scenarios where images span unknown classes and unseen domains, L-Reg consistently improves generalization, highlighting its practical efficacy.
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。
