Make it SING：分析分类器中的语义不变性
Make it SING: Analyzing Semantic Invariants in Classifiers
把分类器线性映射的零空间方向显式构造为等价图像，并借助多模态视觉语言模型为这些方向赋予自然语言语义，为分析模型不变性、比较不同骨干（ResNet 与 DINO-ViT）的语义保持能力提供了可复用的探测工具。
原文依据 SING can be applied to a single image, uncovering local invariants, or to sets of images
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论所有分类器，包括最优的视觉模型，都具有不变性，其部分根源在于其线性映射的几何结构。这些不变性位于分类器的零空间中，会诱导出映射到相同输出的等价输入集合。这些不变性的语义内容仍然模糊，因为现有方法难以提供人类可解释的信息。为填补这一空白，我们提出零空间几何的语义解释（SING），一种构造相对于网络等价的图像、并为可用变化赋予语义解释的方法。我们使用从网络特征到多模态视觉语言模型的映射，从而获得所诱导语义偏移的自然语言描述和视觉示例。SING 可应用于单张图像，揭示局部不变性；也可应用于图像集合，从而在类别和模型层面开展广泛的统计分析。例如，我们的方法揭示 ResNet50 会将相关语义属性泄漏到零空间，而 DINO-ViT（一种使用自监督 DINO 预训练的 ViT）在不变空间中维持类别语义方面更优。代码见 https://tinyurl.com/github-SING。
英文原文摘要 1267 字符
All classifiers, including state-of-the-art vision models, possess invariants, partially rooted in the geometry of their linear mappings. These invariants, which reside in the null-space of the classifier, induce equivalent sets of inputs that map to identical outputs. The semantic content of these invariants remains vague, as existing approaches struggle to provide human-interpretable information. To address this gap, we present Semantic Interpretation of the Null-space Geometry (SING), a method that constructs equivalent images, with respect to the network, and assigns semantic interpretations to the available variations. We use a mapping from network features to multi-modal vision language models. This allows us to obtain natural language descriptions and visual examples of the induced semantic shifts. SING can be applied to a single image, uncovering local invariants, or to sets of images, allowing a breadth of statistical analysis at the class and model levels. For example, our method reveals that ResNet50 leaks relevant semantic attributes to the null space, whereas DINO-ViT, a ViT pretrained with self-supervised DINO, is superior in maintaining class semantics across the invariant space. Code is available at https://tinyurl.com/github-SING.