论文标题
无源域概括的域unified的及时表示
Domain-Unified Prompt Representations for Source-Free Domain Generalization
论文作者
论文摘要
旨在使模型在看不见的域上起作用的领域概括（DG）是通用人工智能的必然方法。受当前DG数据集的规模和多样性的限制，现有方法很难在开放世界的方案（例如，科幻小说和像素化样式）中扩展到各种域。因此，无源域的概括（SFDG）任务是必要且具有挑战性的。为了解决这个问题，我们提出了一种基于大规模视觉语言预处理模型（例如剪辑）的方法，该方法利用了其中嵌入的广泛域信息。所提出的计划生成了来自域表的各种提示，其中包含比现有DG数据集更多的不同领域。此外，我们的方法从这些提示中得出域ust unifted表示，因此能够应对来自开放世界域的样本。对主流DG数据集的广泛实验，即PACS，VLCS，OfficeHome和Domainnet，表明，与需要训练源域数据的最新方法（SOTA）DG方法相比，所提出的方法可以实现竞争性能。此外，我们收集一个小数据集由两个域组成，以评估所提出方法的开放世界域的概括能力。源代码和数据集将在https://github.com/muse1998/source-free-domain-generalization上公开提供。
Domain generalization (DG), aiming to make models work on unseen domains, is a surefire way toward general artificial intelligence. Limited by the scale and diversity of current DG datasets, it is difficult for existing methods to scale to diverse domains in open-world scenarios (e.g., science fiction and pixelate style). Therefore, the source-free domain generalization (SFDG) task is necessary and challenging. To address this issue, we propose an approach based on large-scale vision-language pretraining models (e.g., CLIP), which exploits the extensive domain information embedded in it. The proposed scheme generates diverse prompts from a domain bank that contains many more diverse domains than existing DG datasets. Furthermore, our method yields domain-unified representations from these prompts, thus being able to cope with samples from open-world domains. Extensive experiments on mainstream DG datasets, namely PACS, VLCS, OfficeHome, and DomainNet, show that the proposed method achieves competitive performance compared to state-of-the-art (SOTA) DG methods that require source domain data for training. Besides, we collect a small datasets consists of two domains to evaluate the open-world domain generalization ability of the proposed method. The source code and the dataset will be made publicly available at https://github.com/muse1998/Source-Free-Domain-Generalization
