度量特征宇宙：面向源无关域自适应的基于拓扑的伪标签与重力一致性
Measure The Feature Universe: Topology-based Pseudo Labeling and Gravity Consistency for Source-Free Domain Adaptation
把伪标签的生成从 logits 空间搬到目标嵌入流形上，用虚拟特征构建『特征宇宙』并只在统计可靠区域传播标签，同时用基于嵌入间引力的重力一致性按特征相似度调节 CR 强度——这套思路可直接移植到任何需要无标签目标域自训练的分类/检索/分割流程中。
原文依据 existing approaches overlook the geometric structure of the target embedding manifold when assigning pseudo labels
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论源无关域自适应（SFDA）仅利用预训练的源模型本身，将其适配到无标签的目标域，通常依赖伪标签，并辅以辅助知识与一致性正则化（CR）机制来缓解所生成伪标签中的噪声。然而，现有方法在分配伪标签时忽略了目标嵌入流形的几何结构，导致距离度量不可靠，进而产生严重的错误标注。此外，现有 CR 仅作用于输出 logits，因而对特征层面的可靠性不敏感。为解决这些问题，我们提出了一种基于特征宇宙（Feature universe）的新型伪标签方案，该特征宇宙是一个扩展的嵌入空间，用于建模逐类的目标分布；同时提出重力一致性（GV）正则化，它依据特征层面的相似度来调节一致性强度。我们的伪标签策略首先用虚拟特征对嵌入空间进行建模，以构建特征宇宙。在该空间上，伪标签通过特征遍历生成，仅从统计上可靠的区域传播标签。此外，GV 同时促进 logit 层面与特征层面的一致性，在使增强图像的预测相互对齐的同时保持嵌入空间的几何结构。它进一步为每个样本调节 CR 的强度，通过定义在两个输入嵌入之间的基于重力的力，防止噪声伪标签被确认。在 Office-Home、DomainNet-126 和 VisDA-C 上的实验表明，该方法相较先前的 SFDA 方法取得了一致提升，且将 GV 引入基线方法还能带来额外增益。
英文原文摘要 1697 字符
Source-free domain adaptation (SFDA) adapts a pre-trained source model to an unlabeled target domain using only the model itself, typically relying on pseudo labeling augmented with auxiliary knowledge and consistency regularization (CR) mechanisms to alleviate noise in the generated pseudo labels. However, existing approaches overlook the geometric structure of the target embedding manifold when assigning pseudo labels, resulting in unreliable distance measurements and consequently severe mislabeling. Moreover, existing CR is applied solely to output logits, making it insensitive to feature-level reliability. To solve these issues, we propose a novel pseudo labeling scheme based on Feature universe, which is an expanded embedding space that models class-wise target distributions and Gravity consistency (GV) regularization, which modulates consistency strength according to feature-level similarity. Our pseudo labeling strategy first models the embedding space with virtual features to construct a feature universe. On this space, pseudo labels are generated through feature traversal, which propagates labels only from statistically reliable regions. In addition, GV jointly encourages logit- and feature-level consistency, aligning predictions for augmented images while preserving the geometric structure of the embedding space. It further modulates the strength of CR for each sample, preventing the confirmation of noisy pseudo labels through a gravity-based force defined between two input embeddings. Experiments on Office-Home, DomainNet-126, and VisDA-C demonstrate consistent improvements over prior SFDA methods, and incorporating GV into baselines yields additional gains.