A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation

Liang, Jian; Wang, Yunbo; Hu, Dapeng; He, Ran; Feng, Jiashi

doi:10.48550/arxiv.2003.02541

Cited by 5 publications

(14 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To compare PDA methods, we conduct the experiments on origin PDA setting (S65 → T 25) [12], where first 25 classes in alphabetical order are included as shared classes, which is a long-tailed distribution. The comparison results are shown in Table III.…”

Section: B the Analysis Of Experimental Resultsmentioning

confidence: 99%

“…To address the challenge of knowledge sharing from non-shared data without negative transfer effect in NI-UDA, we propose the Source Classifier Filter (SCF) mechanism in big source classifier. The previous work of TAN learned from K-classes in the source domain, ignoring the negative transfer effects brought about by these nonshared classes, that is, error accumulation [11] and uncertainty propagation [12]. In order to solve this problem, we propose a SCF mechanism.…”

Section: B Gada Methodsmentioning

confidence: 99%

“…PDA consider as setting where the label of target domain is a subspace of the source domain with the hard assumption of non-shared classes space is unknown. So these PDA methods are proposed to estimate the non-shared classes space and try to remove the examples in non-shared classes automatically [18], [7], [12]. Different from PDA, Open-Set Domain Adaptation(ODA) [19] assumes that target domain has some unknown non-shared classes.…”

Section: Related Workmentioning

confidence: 99%

“…, n/5 in target domain, set these 5 classes with index i to sparse that contains only 10 samples, 60 samples in each of the other 45 non-sparse classes; then, calculate the accuracy and f1 value of 5 sparse class i and 45 non-sparse classes respectively; at the end, average the test results: Baselines. We compare the proposed GADA with state-of-the-art PDA and adversarial UDA methods, including BA 3 US [12], ResNet [3], CDAN [8], MDD [9], GVB-GD [29], GCAN [26], TAN [10], CDAN+TAN and SGR [25].…”

Section: A Experimental Settingsmentioning

confidence: 99%

“…The previous work of TAN [10] learned from K-classes in the source domain, ignoring the negative transfer effects brought about by these non-shared classes, that is, error accumulation [11] and uncertainty propagation [12]. In order to solve the challenge of knowledge sharing from non-shared data without negative transfer effect, we propose a Source Classifier Filter (SCF) mechanism to filter non-shared data with a fixed confidence threshold of source classifier in feature enhance layer.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

NI-UDA: Graph Adversarial Domain Adaptation from Non-shared-and-Imbalanced Big Data to Small Imbalanced Applications

Xiao¹,

Xiang

Huan

et al. 2021

Preprint

View full text Add to dashboard Cite

We propose a new general Graph Adversarial Domain Adaptation (GADA) based on semantic knowledge reasoning of class structure for solving the problem of unsupervised domain adaptation (UDA) from the big data with non-shared and imbalanced classes to specified small and imbalanced applications (NI-UDA), where non-shared classes mean the label space out of the target domain. Our goal is to leverage priori hierarchy knowledge to enhance domain adversarial aligned feature representation with graph reasoning. In this paper, to address two challenges in NI-UDA, we equip adversarial domain adaptation with Hierarchy Graph Reasoning (HGR) layer and the Source Classifier Filter (SCF). For sparse classes transfer challenge, our HGR layer can aggregate local feature to hierarchy graph nodes by node prediction and enhance domain adversarial aligned feature with hierarchy graph reasoning for sparse classes. Our HGR contributes to learn direct semantic patterns for sparse classes by hierarchy attention in self-attention, non-linear mapping and graph normalization. our SCF is proposed for the challenge of knowledge sharing from non-shared data without negative transfer effect by filtering low-confidence non-shared data in HGR layer. Experiments on two benchmark datasets show our GADA methods consistently improve the state-of-theart adversarial UDA algorithms, e.g. GADA(HGR) can greatly improve f1 of the MDD by 7.19% and GVB-GD by 7.89% respectively on imbalanced source task in Meal300 dataset. The code is available at https://gadatransfer.wixsite.com/gada.

show abstract

Section: B the Analysis Of Experimental Resultsmentioning

confidence: 99%

Section: B Gada Methodsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: A Experimental Settingsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

NI-UDA: Graph Adversarial Domain Adaptation from Non-shared-and-Imbalanced Big Data to Small Imbalanced Applications

Xiao¹,

Xiang

Huan

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

Universal Domain Adaptation without Source Data for Remote Sensing Image Scene Classification

Shi

Zhu

2022

IGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium

View full text Add to dashboard Cite

Existing domain adaptation (DA) approaches are usually not well suited for practical DA scenarios of remote sensing image classification, since these methods (such as unsupervised DA) rely on rich prior knowledge about the relationship between label sets of source and target domains, and source data are usually not accessible in many cases due to the privacy or confidentiality issues. To this end, we propose a novel source data generation-based universal domain adaptation (SDG-UniDA) model, which includes two parts, i.e., the stage of source data generation and the stage of model adaptation. The first stage is to estimate the conditional distribution of source data from the pre-trained model using the knowledge of class-separability in the source domain and then to synthesize the source data. With this synthetic source data in hand, it becomes a universal DA task that requires no prior knowledge on the label sets. A novel transferable weight is proposed to distinguish the shared and private label sets to each domain, thereby promoting the adaptation in the automatically discovered shared label set and recognizing the "unknown" samples successfully. Empirical results show that SDG-UniDA is effective and practical in this challenging setting for remote sensing image scene classification.

show abstract

Partial Domain Adaptation for Scene Classification From Remote Sensing Imagery

Zheng

Zhao

et al. 2023

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

Although domain adaptation approaches have been proposed to tackle cross-regional, multitemporal, and multisensor remote sensing applications since they do not require any human interpretation in the target domain, most current works assume identical label space across the source and the target domains. However, in real-world applications, we often transfer knowledge from a large-scale dataset with rich annotations to a small-scale target dataset with scarcity of labels. In most cases, the label space of the source domain is usually large enough to subsume that of the target domain, which is termed partial domain adaptation. In this article, we propose a new partial domain adaptation algorithm for remote sensing scene classification and our proposed method contains three major parts. First, we employ a progressive auxiliary domain module to alleviate the negative transfer effect caused by outlier classes. Second, we adopt an improved domain adversarial neural network (DANN) with multiweights to better encourage domain confusion. Last but not least, we design an attentive complement entropy regularization to improve the prediction confidence for samples and avoid untransferable samples (such as the samples belonging to outlier classes in the source domain) being mistakenly classified. We collect three common remote sensing datasets to evaluate our proposed method. Our method achieves an average accuracy of 79.36%, which considerably outperforms other state-of-the-art partial domain adaptation methods with an average accuracy improvement of 1.90%-12.45% and attaining a 13.67% gain compared to the straightforward deep learning model (ResNet-50). The experiment results indicate that our approach shows promising prospects for solving more general Manuscript

show abstract

A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation

Cited by 5 publications

References 29 publications

NI-UDA: Graph Adversarial Domain Adaptation from Non-shared-and-Imbalanced Big Data to Small Imbalanced Applications

NI-UDA: Graph Adversarial Domain Adaptation from Non-shared-and-Imbalanced Big Data to Small Imbalanced Applications

Universal Domain Adaptation without Source Data for Remote Sensing Image Scene Classification

Partial Domain Adaptation for Scene Classification From Remote Sensing Imagery

Contact Info

Product

Resources

About