面向通用域适应的免训练标签空间对齐
Training-Free Label Space Alignment for Universal Domain Adaptation
- Department of Artificial Intelligence, Korea University（人工智能系，韩国大学）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对通用域适应中视觉空间对齐易受歧义干扰的问题，提出免训练标签空间对齐方法，借助视觉语言基础模型识别未知类别、细化噪声标签并构建通用分类器，在DomainBed基准上实现显著性能提升。
AI中文摘要：
通用域适应（Universal Domain Adaptation, UniDA）实现从有标注源域到无标注目标域的知识迁移，二者的标签空间可能存在差异，且目标域可能包含私有类别。以往的UniDA方法主要聚焦于视觉空间对齐，但常因内容差异导致的视觉歧义问题陷入困境，限制了方法的鲁棒性与泛化性。为解决该问题，本文提出一种新方法，借助CLIP等近期视觉语言基础模型（Vision-Language Foundation Models, VLMs）强大的零样本能力，仅专注于标签空间对齐以提升迁移稳定性。CLIP仅依据标签名称即可生成任务特定分类器，但由于标签空间无法提前完全获知，将CLIP适配至UniDA任务存在挑战。本研究首先利用生成式视觉语言模型识别目标域中的未知类别。所发现的标签中存在噪声与语义歧义（例如与源标签相似的同义词、上位词、下位词），为标签对齐增加了难度。针对这一问题，本文提出一种面向UniDA的免训练标签空间对齐方法。该方法通过过滤并细化跨域的噪声标签，实现标签空间而非视觉空间的对齐，随后构建融合共享知识与目标域私有类别信息的通用分类器，从而提升域偏移场景下的泛化性。实验结果表明，所提方法在核心DomainBed基准上显著优于现有UniDA技术，H-score平均提升+7.9%，H³-score平均提升+6.1%。此外，引入自训练可进一步提升性能，使H-score与H³-score均额外提升+1.6%。
英文摘要：
Universal domain adaptation (UniDA) transfers knowledge from a labeled source domain to an unlabeled target domain, where label spaces may differ and the target domain may contain private classes. Previous UniDA methods primarily focused on visual space alignment but often struggled with visual ambiguities due to content differences, which limited their robustness and generalizability. To overcome this, we introduce a novel approach that leverages the strong \textit{zero-shot capabilities} of recent vision-language foundation models (VLMs) like CLIP, concentrating solely on label space alignment to enhance adaptation stability. CLIP can generate task-specific classifiers based only on label names. However, adapting CLIP to UniDA is challenging because the label space is not fully known in advance. In this study, we first utilize generative vision-language models to identify unknown categories in the target domain. Noise and semantic ambiguities in the discovered labels -- such as those similar to source labels (e.g., synonyms, hypernyms, hyponyms) -- complicate label alignment. To address this, we propose a training-free label-space alignment method for UniDA (\ours). Our method aligns label spaces instead of visual spaces by filtering and refining noisy labels between the domains. We then construct a \textit{universal classifier} that integrates both shared knowledge and target-private class information, thereby improving generalizability under domain shifts. The results reveal that the proposed method considerably outperforms existing UniDA techniques across key DomainBed benchmarks, delivering an average improvement of \textcolor{blue}{+7.9\%}in H-score and \textcolor{blue}{+6.1\%} in H$^3$-score. Furthermore, incorporating self-training further enhances performance and achieves an additional (\textcolor{blue}{+1.6\%}) increment in both H- and H$^3$-scores.
