Pseudo-margin-based universal domain adaptation

Universal Domain Adaptation

Domain Adaptation in Computer Vision with Deep Learning ◽

10.1007/978-3-030-45529-3_11 ◽

2020 ◽

pp. 195-211 ◽

Cited By ~ 1

Author(s):

Kaichao You ◽

Mingsheng Long ◽

Zhangjie Cao ◽

Jianmin Wang ◽

Michael I. Jordan

Keyword(s):

Domain Adaptation ◽

Universal Domain

Download Full-text

Domain Consensus Clustering for Universal Domain Adaptation

10.1109/cvpr46437.2021.00963 ◽

2021 ◽

Author(s):

Guangrui Li ◽

Guoliang Kang ◽

Yi Zhu ◽

Yunchao Wei ◽

Yi Yang

Keyword(s):

Domain Adaptation ◽

Consensus Clustering ◽

Universal Domain

Download Full-text

Synthetic Source Universal Domain Adaptation through Contrastive Learning

Sensors ◽

10.3390/s21227539 ◽

2021 ◽

Vol 21 (22) ◽

pp. 7539

Author(s):

Jungchan Cho

Keyword(s):

Domain Adaptation ◽

Synthetic Data ◽

Real Data ◽

Target Domain ◽

Source Domain ◽

Universal Domain ◽

Model Training ◽

Using Data ◽

Imaging Sensors ◽

Target Data

Universal domain adaptation (UDA) is a crucial research topic for efficient deep learning model training using data from various imaging sensors. However, its development is affected by unlabeled target data. Moreover, the nonexistence of prior knowledge of the source and target domain makes it more challenging for UDA to train models. I hypothesize that the degradation of trained models in the target domain is caused by the lack of direct training loss to improve the discriminative power of the target domain data. As a result, the target data adapted to the source representations is biased toward the source domain. I found that the degradation was more pronounced when I used synthetic data for the source domain and real data for the target domain. In this paper, I propose a UDA method with target domain contrastive learning. The proposed method enables models to leverage synthetic data for the source domain and train the discriminativeness of target features in an unsupervised manner. In addition, the target domain feature extraction network is shared with the source domain classification task, preventing unnecessary computational growth. Extensive experimental results on VisDa-2017 and MNIST to SVHN demonstrated that the proposed method significantly outperforms the baseline by 2.7% and 5.1%, respectively.

Download Full-text

Divergence Optimization for Noisy Universal Domain Adaptation

10.1109/cvpr46437.2021.00254 ◽

2021 ◽

Author(s):

Qing Yu ◽

Atsushi Hashimoto ◽

Yoshitaka Ushiku

Keyword(s):

Domain Adaptation ◽

Universal Domain

Download Full-text

Universal Domain Adaptation

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) ◽

10.1109/cvpr.2019.00283 ◽

2019 ◽

Cited By ~ 27

Author(s):

Kaichao You ◽

Mingsheng Long ◽

Zhangjie Cao ◽

Jianmin Wang ◽

Michael I. Jordan

Keyword(s):

Domain Adaptation ◽

Universal Domain

Download Full-text

A New Universal Domain Adaptive Method for Diagnosing Unknown Bearing Faults

Entropy ◽

10.3390/e23081052 ◽

2021 ◽

Vol 23 (8) ◽

pp. 1052

Author(s):

Zhenhao Yan ◽

Guifang Liu ◽

Jinrui Wang ◽

Huaiqian Bao ◽

Zongzhen Zhang ◽

...

Keyword(s):

Fault Diagnosis ◽

Domain Adaptation ◽

Transfer Problem ◽

Adaptive Method ◽

Transfer Model ◽

Target Domain ◽

Network Failure ◽

Multiple Datasets ◽

Mechanical Faults ◽

Universal Domain

The domain adaptation problem in transfer learning has received extensive attention in recent years. The existing transfer model for solving domain alignment always assumes that the label space is completely shared between domains. However, this assumption is untrue in the actual industry and limits the application scope of the transfer model. Therefore, a universal domain method is proposed, which not only effectively reduces the problem of network failure caused by unknown fault types in the target domain but also breaks the premise of sharing the label space. The proposed framework takes into account the discrepancy of the fault features shown by different fault types and forms the feature center for fault diagnosis by extracting the features of samples of each fault type. Three optimization functions are added to solve the negative transfer problem when the model solves samples of unknown fault types. This study verifies the performance advantages of the framework for variable speed through experiments of multiple datasets. It can be seen from the experimental results that the proposed method has better fault diagnosis performance than related transfer methods for solving unknown mechanical faults.

Download Full-text