Learning to Optimize Domain Specific Normalization for Domain Generalization

Seo, Seonguk; Suh, Young-Ger; Kim, Dong-Wan; Kim, Geeho; Han, Jong-Woo; Han, Bohyung

doi:10.1007/978-3-030-58542-6_5

Cited by 171 publications

(110 citation statements)

References 20 publications

Supporting

Mentioning

107

Contrasting

Order By: Relevance

“…However, as pointed out by [55], interpretable domain generalization is still a challenge in future research. Similar to our work, methods in [7,45,46] argue that domainspecific factors also matter in DG. Specifically, [45] measures domain similarities in a transferrable domain space that is implicitly learned by domain-specific batch normalization layers.…”

Section: Related Worksupporting

confidence: 85%

“…The results of ERM [49], GroupDRO [44], Mixup [58], MLDG [22], DANN [12], C-DANN [32], CORAL [47], MMD [24], IRM [3], ARM [60], MTL [5], SagNet [37], and RSC [17] are cited from [14] except for Office-31, on which we conduct experiments and report the results reproduced by us. For PACS, we also compare our method with EisNet [57], Entropy Regularization (ER) [62], DMG [7], DSON [46], MASF [9] and MetaReg [4] which also use ResNet-50 as the backbone, the results of these methods are directly cited from original papers. It is worth noting that from the results reported in previous DG research over the years, one can easily find that even a marginal improvement (e.g., 1%) on average accuracy is challenging in the DG community.…”

Section: Comparison With Other Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Learning Transferrable and Interpretable Representations for Domain Generalization

Lü

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

Conventional machine learning models are often vulnerable to samples with different distributions from the ones of training samples, which is known as domain shift. Domain Generalization (DG) challenges this issue by training a model based on multiple source domains and generalizing it to arbitrary unseen target domains. In spite of remarkable results made in DG, a majority of existing works lack a deep understanding of the feature representations learned in DG models, resulting in limited generalization ability when facing domains out-of-distribution. In this paper, we aim to learn a domain transformation space via a domain transformer network (DTN) which explicitly mines the relationship among multiple domains and constructs transferrable feature representations for down-stream tasks by interpreting each feature as a semantically weighted combination of multiple domain-specific features. Our DTN is encouraged to meta-learn the properties and characteristics of domains during the training process based on multiple seen domains, making transformed feature representations more semantical, thus generalizing better to unseen domains. Once the model is constructed, the feature representations of unseen target domains can also be inferred adaptively by selectively combining the feature representations from the diverse set of seen domains. We conduct extensive experiments on five DG benchmarks and the results strongly demonstrate the effectiveness of our approach. CCS CONCEPTS• Computing methodologies → Computer vision; Learning paradigms; Image representations.

show abstract

Section: Related Worksupporting

confidence: 85%

Section: Comparison With Other Methodsmentioning

confidence: 99%

Learning Transferrable and Interpretable Representations for Domain Generalization

Lü

et al. 2021

Proceedings of the 29th ACM International Conference on Multimedia

View full text Add to dashboard Cite

show abstract

“…Recently, some methods are also proposed to address the domain generalization problem in the classification and semantic segmentation tasks [27], [28], [29], [30], [31], [32], [33], [34]. Particularly, the data augmentation based methods have obtained significant advance in these task, which can be divided into image-level and feature-level data augmentation.…”

Section: B Domain Generalizationmentioning

confidence: 99%

A Novel Mix-normalization Method for Generalizable Multi-source Person Re-identification

Liu¹,

Wang²,

Shi³

et al. 2022

Preprint

View full text Add to dashboard Cite

Person re-identification (Re-ID) has achieved great success in the supervised scenario. However, it is difficult to directly transfer the supervised model to arbitrary unseen domains due to the model overfitting to the seen source domains. In this paper, we aim to tackle the generalizable multi-source person Re-ID task (i.e., there are multiple available source domains, and the testing domain is unseen during training) from the data augmentation perspective, thus we put forward a novel method, termed MixNorm, which consists of domain-aware mixnormalization (DMN) and domain-ware center regularization (DCR). Different from the conventional data augmentation, the proposed domain-aware mix-normalization to enhance the diversity of features during training from the normalization view of the neural network, which can effectively alleviate the model overfitting to the source domains, so as to boost the generalization capability of the model in the unseen domain. To better learn the domain-invariant model, we further develop the domainaware center regularization to better map the produced diverse features into the same space. Extensive experiments on multiple benchmark datasets validate the effectiveness of the proposed method and show that the proposed method can outperform the state-of-the-art methods. Besides, further analysis also reveals the superiority of the proposed method.

show abstract

“…Domain generalization. Unlike domain adaptation, domain generalization [3,4,5] cannot use any sample of the target domain, but it still has to capture transferable information across domains. To complete the classification task without the target domain available at the time of training, labelled data from several related classification tasks can be used.…”

Section: Related Workmentioning

confidence: 99%

A Framework for Self-Supervised Federated Domain Adaptation

Wang

Wu³

et al. 2021

Preprint

View full text Add to dashboard Cite

Unsupervised federated domain adaptation uses the knowledge from several distributed unlabelled source domains to complete the learning on the unlabelled target domain. Some of the existing methods have limited effectiveness and involve frequent communication. This paper proposes a framework to solve the distributed multi-source domain adaptation problem, referred as self-supervised federated domain adaptation (SFDA). Specifically, a multi-domain model generalization balance (MDMGB) is proposed to aggregate the models from multiple source domains in each round of communication. A weighted strategy based on centroid similarity is also designed for SFDA. SFDA conducts self-supervised training on the target domain to tackle domain shift. Compared with the classical federated adversarial domain adaptation algorithm, SFDA is not only strong in communication cost and privacy protection but also improves in the accuracy of the model.

show abstract

Learning to Optimize Domain Specific Normalization for Domain Generalization

Cited by 171 publications

References 20 publications

Learning Transferrable and Interpretable Representations for Domain Generalization

Learning Transferrable and Interpretable Representations for Domain Generalization

A Novel Mix-normalization Method for Generalizable Multi-source Person Re-identification

A Framework for Self-Supervised Federated Domain Adaptation

Contact Info

Product

Resources

About