Eliminating the covariate shift cross domains is one of the common methods to deal with the issue of domain shift in visual unsupervised domain adaptation. However, current alignment methods, especially the prototype based or sample-level based methods neglect the structural properties of the underlying distribution and even break the condition of covariate shift. To relieve the limitations and conflicts, we introduce a novel concept named (virtual) mirror, which represents the equivalent sample in another domain. The equivalent sample pairs, named mirror pairs reflect the natural correspondence of the empirical distributions. Then a mirror loss, which aligns the mirror pairs cross domains, is constructed to enhance the alignment of the domains. The proposed method does not distort the internal structure of the underlying distribution. We also provide theoretical proof that the mirror samples and mirror loss have better asymptotic properties in reducing the domain shift. By applying the virtual mirror and mirror loss to the generic unsupervised domain adaptation model, we achieved consistent superior performance on several mainstream benchmarks.
翻译:消除共变转移跨域是处理视觉上不受监督的域适应域域变化问题的常见方法之一。 但是,当前的调整方法,特别是原型或样本级方法,忽视了基础分布的结构属性,甚至打破了共变转移的条件。 为了缓解限制和冲突,我们引入了名为(虚拟)镜像的新概念,它代表了另一个域的同等样本。对应的样本对子,名为镜对子,反映了经验分布的自然对应。然后,对镜对子跨域进行对齐的镜损,用来加强域的对齐。拟议方法并不扭曲基础分布的内部结构。我们还提供理论证据,证明镜样和镜损在减少域转移方面有更好的非约束性属性。通过将虚拟镜像和镜损应用到通用的无监督域适应模型中,我们在若干主流基准上取得了一致的优异性表现。