Multi-Step Online Unsupervised Domain Adaptation
Multi-Step Online Unsupervised Domain Adaptation
复制标题
DOI：
10.1109/icassp40776.2020.9052976
复制
发表时间：
2020-02
期刊：
影响因子：
--
通讯作者：
J. Moon;Debasmit Das;George Lee
中科院分区：
文献类型：
--
作者：
J. Moon;Debasmit Das;George Lee
在本文中，我们解决了在线无监督域适应（OUDA）问题，其中目标数据未标记并按顺序到达。传统的OUDA问题方法主要关注将每个到达的目标数据转换到源域，没有充分考虑到达的目标数据之间的时间一致性和累积统计量。我们针对 OUDA 问题提出了一个多步骤框架，该框架受欧几里得空间几何解释的启发，建立了一种计算平均目标子空间的新方法。该平均目标子空间包含到达的目标数据之间的累积时间信息。此外，作为预处理步骤，将从平均目标子空间计算出的变换矩阵应用于下一个目标数据，使目标数据更接近源域。对四个数据集的实验证明了我们提出的多步骤 OUDA 框架中每个步骤的贡献及其相对于以前方法的性能。
In this paper, we address the Online Unsupervised Domain Adaptation (OUDA) problem, where the target data are unlabelled and arriving sequentially. The traditional methods on the OUDA problem mainly focus on transforming each arriving target data to the source domain, and they do not sufficiently consider the temporal coherency and accumulative statistics among the arriving target data. We propose a multi-step framework for the OUDA problem, which institutes a novel method to compute the mean-target subspace inspired by the geometrical interpretation on the Euclidean space. This mean-target subspace contains accumulative temporal information among the arrived target data. Moreover, the transformation matrix computed from the mean-target subspace is applied to the next target data as a preprocessing step, aligning the target data closer to the source domain. Experiments on four datasets demonstrated the contribution of each step in our proposed multi-step OUDA framework and its performance over previous approaches.