Source Hypothesis Transfer for Unsupervised Domain Adaptation ICML 2020 链接:https://arxiv.org/pdf/2002.08546.pdf 这篇文章算是比较早期做source free的工作了,后面的工作基本上都会以这篇作为baseline。它的方法也是非常简单:源模型(source model)由一个feature extractor和一个classifier构成,它的方法就是固定住c...
Unsupervised domain adaptation (UDA) has achieved great success in handling cross-domain machine learning applications.It typically benefits the model training of unlabeled target domain by leveraging knowledge from labeled source domain.For this purpose,the minimization of the marginal distribution ...
Source Hypothesis Transfer for Unsupervised Domain Adaptation。 既然不需要访问源域数据集,那么如何获取源域的信息呢,作者提出可以将在源域数据集上训练好的模型拿来直接用,使用无标记的目标域数据集对模型进行调整,这也是Source-Free DA问题的惯用解决方案。相比于源域数据集,源域模型的体量就要小很多了: Digits和...
模型良好初始化）的unsupervised clustering问题。没有source data其实就不用直接解决domain shift因为只关注...
下面是个人分类整理的Source-Free的论文列表,非完全,后续将跟踪补充,欢迎各位大佬指导补充! 1.冻结分类器 BAIT(Unsupervised Domain Adaptation without Source Data by Casting a BAIT) GITHUB:https://github.com/BIT-DA/DCAN SHOT(Do We Really Need to Access the Source Data? Source Hypothesis Transfer for...
GITHUB:https://github.com/youngryan1993/PrDA-Progressive-Domain-Adaptation-from-a-Source-Pre-trained-Model Noisy Label(Self-Supervised Noisy Label Learning for Source-Free Unsupervised Domain Adaptation) GITHUB: 4.加权机制 usfda(Universal Source-Free Domain Adaptation) ...
上图展示了source-free domain adaptation和一般的DA的区别。在之前的两篇source-free的论文中已经反复讲解,不再赘述。 1 方法 这文章也是使用Positive learning和Negative Learning的方法。 方法名称:Source-Free domain adaptive Semantic Segmentation (SFSS) ...
Paper tables with annotated results for Enhancing and Adapting in the Clinic: Source-free Unsupervised Domain Adaptation for Medical Image Enhancement
Source-Free Unsupervised Domain Adaptation (SF-UDA) aims to transfer a model's performance from a labeled source domain to an unlabeled target domain without direct access to source samples, addressing data privacy issues. However, most existing SF-UDA approaches assume the availability of abundant ...
[CVPR'24]Source-Free Domain Adaptation with Frozen Multimodal Foundation Model (DIFO) [IJCV'23]Source-Free Domain Adaptation via Target Prediction Distribution Searching (TPDS) [NN'22]Semantic consistency learning on manifold for source data-free unsupervised domain adaptation (SCLM) ...