Backprop Induced Feature Weighting for Adversarial Domain Adaptation with Iterative Label Distribution Alignment
Backprop Induced Feature Weighting for Adversarial Domain Adaptation with Iterative Label Distribution Alignment
复制标题
DOI：
10.1109/wacv56688.2023.00047
复制
发表时间：
2023-01
期刊：
影响因子：
--
通讯作者：
Thomas Westfechtel;Hao-Wei Yeh;Qier Meng;Yusuke Mukuta;Tatsuya Harada
中科院分区：
文献类型：
--
作者：
Thomas Westfechtel;Hao-Wei Yeh;Qier Meng;Yusuke Mukuta;Tatsuya Harada
对大型标记数据集的需求是训练准确深度神经网络的限制因素之一。无监督域自适应通过将知识从一个具有许多标记数据的域转移到一个几乎没有标记数据的不同域来解决训练数据有限的问题。一种常见的方法是学习域不变特征，例如使用对抗方法。以前的方法通常分别训练领域分类器和标签分类器网络，其中两个分类网络彼此之间几乎没有交互。在本文中，我们介绍了一种基于分类器的反向传播诱导加权的特征空间。这种方法有两个主要优点。首先，它让领域分类器专注于对分类重要的特征，其次，它将分类和对抗分支更紧密地结合在一起。此外，我们引入了一个迭代的标签分布对齐方法，采用以前运行的结果近似类平衡的数据加载器。我们进行实验和消融研究的三个基准Office-31，家庭，和DomainNet显示我们所提出的算法的有效性。
The requirement for large labeled datasets is one of the limiting factors for training accurate deep neural networks. Unsupervised domain adaptation tackles this problem of limited training data by transferring knowledge from one domain, which has many labeled data, to a different domain for which little to no labeled data is available. One common approach is to learn domain-invariant features for example with an adversarial approach. Previous methods often train the domain classifier and label classifier network separately, where both classification networks have little interaction with each other. In this paper, we introduce a classifier-based backprop-induced weighting of the feature space. This approach has two main advantages. Firstly, it lets the domain classifier focus on features that are important for the classification, and, secondly, it couples the classification and adversarial branch more closely. Furthermore, we introduce an iterative label distribution alignment method, that employs results of previous runs to approximate a class-balanced dataloader. We conduct experiments and ablation studies on three benchmarks Office-31, Office-Home, and DomainNet to show the effectiveness of our proposed algorithm.