Domain Generalization via Adversarially Learned Novel Domains
Domain Generalization via Adversarially Learned Novel Domains
复制标题
通过对抗性学习的新领域进行领域泛化
DOI：
10.1109/access.2022.3209815
复制
发表时间：
2023
期刊：
影响因子：
3.9
通讯作者：
Yu Zhe; Kazuto Fukuchi; Youhei Akimoto; Jun Sakuma
中科院分区：
文献类型：
--
作者：
Kakizaki Kazuya;Fukuchi Kazuto;Sakuma Jun;Yu Zhe; Kazuto Fukuchi; Youhei Akimoto; Jun Sakuma
本研究重点关注领域泛化任务，旨在通过利用多个训练领域来学习泛化到未见领域的模型。更具体地说，我们遵循对抗性数据增强的思想，旨在合成和增强具有“硬”域的训练数据，以提高模型的域泛化能力。然而，之前的研究仅使用与训练数据相似的样本来增强训练数据，导致泛化能力有限。为了缓解这个问题，我们提出了一种新颖的对抗性数据增强方法，称为 GADA（生成对抗性域增强），该方法采用图像到图像转换模型来获得新域的分布，这些新域在语义上与训练域不同，同时难以分类。评估和进一步分析表明GADA符合我们的预期；使用语义不同的样本进行对抗性数据增强可以带来更好的领域泛化性能。
This study focuses on the domain generalization task, which aims to learn a model that generalizes to unseen domains by utilizing multiple training domains. More specifically, we follow the idea of adversarial data augmentation, which aims to synthesize and augment training data with “hard” domains to improve the model’s domain generalization ability. However, previous studies augmented training data only with samples similar to the training data, resulting in limited generalization ability. To alleviate this issue, we propose a novel adversarial data augmentation method, termed GADA (generative adversarial domain augmentation), which employs an image-to-image translation model to obtain a distribution of novel domains that are semantically different from the training domains, and, at the same time, hard to classify. Evaluation and further analysis suggest that GADA fits our expectation; adversarial data augmentation with semantically different samples leads to better domain generalization performance.