Journal of Computer Applications ›› 2025, Vol. 45 ›› Issue (9): 2727-2736.DOI: 10.11772/j.issn.1001-9081.2024091277

• Artificial intelligence •    

Review of open set domain adaptation

Chuang WANG, Lu YU(), Jianwei CHEN, Cheng PAN, Wenbo DU   

  1. School of Communications Engineering,Army Engineering University,Nanjing Jiangsu 210007,China
  • Received:2024-09-09 Revised:2025-02-25 Accepted:2025-03-03 Online:2025-03-26 Published:2025-09-10
  • Contact: Lu YU
  • About author:WANG Chuang, born in 1995, M. S. candidate. His research interests include transfer learning, pattern recognition.
    CHEN Jianwei, born in 2000, M. S. candidate. His research interests include transfer learning, pattern recognition.
    PAN Cheng, born in 2000, M. S. candidate. His research interests include adversarial attacks, deep learning.
    DU Wenbo, born in 2002, M. S. candidate. His research interests include wireless networks, machine learning.
  • Supported by:
    National Natural Science Foundation of China(62471486)

开集域适应综述

王闯, 俞璐(), 陈健威, 潘成, 杜文博   

  1. 陆军工程大学 通信工程学院,南京 210007
  • 通讯作者: 俞璐
  • 作者简介:王闯(1995—),男,安徽滁州人,硕士研究生,主要研究方向:迁移学习、模式识别
    陈健威(2000—),男,广东江门人,硕士研究生,主要研究方向:迁移学习、模式识别
    潘成(2000—),男,山东烟台人,硕士研究生,主要研究方向:对抗攻击、深度学习
    杜文博(2002—),男,河南驻马店人,硕士研究生,主要研究方向:无线网络、机器学习。
  • 基金资助:
    国家自然科学基金资助项目(62471486)

Abstract:

As a critical technique in transfer learning, domain adaptation addresses the issue of different distributions in training and test datasets well. However, traditional domain adaptation methods are typically limited to scenarios where the target-domain and source-domain datasets are with same number and types of categories. In practical applications, these scenarios are often difficult to meet. Open Set Domain Adaptation (OSDA) emerges to address this challenge. In order to fill the gap in this field and provide a reference for related research, a summary and analysis of OSDA methods emerged in recent years were conducted. Firstly, the related concepts and basic structure were introduced. Secondly, the related typical methods were sorted out and analyzed from three stages: data augmentation-oriented, feature extraction-oriented, and classifier-oriented. Finally, future development directions of OSDA were prospected.

Key words: transfer learning, Open Set Domain Adaptation (OSDA), data augmentation, feature extraction, classifier

摘要:

作为迁移学习的关键技术,域适应能很好地解决训练和测试数据集分布不同的问题。然而,传统的域适应方法通常只适用于目标域和源域数据集所含类别的数量和种类相同的情况,在实际场景中该条件通常很难满足。开集域适应(OSDA)正是为了解决此问题而出现的。为了填补该领域的空白,并为相关研究提供借鉴参考,对近年来出现的OSDA方法进行归纳分析。首先,介绍相关概念与基本结构;其次,分别从针对数据增强、针对特征提取以及针对分类器3个阶段梳理分析相关的典型方法;最后,对OSDA的未来发展方向进行展望。

关键词: 迁移学习, 开集域适应, 数据增强, 特征提取, 分类器

CLC Number: