Larger Norm More Transferable: An Adaptive Feature Norm Approach for Unsupervised Domain Adaptation

Xu, Ruijia; Li, Guanbin; Yang, Jihan; Li, Lin

doi:10.1109/iccv.2019.00151

Cited by 469 publications

(258 citation statements)

References 38 publications

Supporting

Mentioning

254

Contrasting

Order By: Relevance

“…Zhang et al [14] presented a two-domain-classifier strategy to identify the importance scores of source samples, and match the reweighed source domain and the target domain based on adversarial learning. In addition, Xu et al [15] proposed to adapt feature norms of both domains to achieve equilibrium, which is free from relationship of the label spaces. Recently, some methods apply ensemble learning to improve the discrimination ability of extracted features [37], [38].…”

Section: B Partial Domain Adaptationmentioning

confidence: 99%

Learning Target-Domain-Specific Classifier for Partial Domain Adaptation

Ren

Yang

et al. 2021

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Unsupervised domain adaptation (UDA) aims at reducing the distribution discrepancy when transferring knowledge from a labeled source domain to an unlabeled target domain. Previous UDA methods assume that the source and target domains share an identical label space, which is unrealistic in practice since the label information of the target domain is agnostic. This paper focuses on a more realistic UDA scenario, i.e. partial domain adaptation (PDA), where the target label space is subsumed to the source label space. In the PDA scenario, the source outliers that are absent in the target domain may be wrongly matched to the target domain (technically named negative transfer), leading to performance degradation of UDA methods. This paper proposes a novel Target Domain Specific Classifier Learning-based Domain Adaptation (TSCDA) method. TSCDA presents a soft-weighed maximum mean discrepancy criterion to partially align feature distributions and alleviate negative transfer. Also, it learns a target-specific classifier for the target domain with pseudo-labels and multiple auxiliary classifiers, to further address classifier shift. A module named Peers Assisted Learning is used to minimize the prediction difference between multiple target-specific classifiers, which makes the classifiers more discriminant for the target domain. Extensive experiments conducted on three PDA benchmark datasets show that TSCDA outperforms other state-of-the-art methods with a large margin, e.g. 4% and 5.6% averagely on Office-31 and Office-Home, respectively.

show abstract

Section: B Partial Domain Adaptationmentioning

confidence: 99%

Learning Target-Domain-Specific Classifier for Partial Domain Adaptation

Ren

Yang

et al. 2021

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

show abstract

“…Domain discrepancies commonly exist across different datasets, and variant domain adaptation methods [29,37,46] are intensively proposed to learn domain-invariant feature thus that classier/predictor learned using the source datasets can be generalized to the target [19,26] for visual interaction learning and reasoning. These works propose to explicitly model label dependencies in the form of graphs and adopt the graph to guide feature interaction learning.…”

Section: Adversarial Domain Adaptationmentioning

confidence: 99%

Adversarial Graph Representation Adaptation for Cross-Domain Facial Expression Recognition

Xie¹,

Chen²,

et al. 2020

Proceedings of the 28th ACM International Conference on Multimedia

Self Cite

View full text Add to dashboard Cite

Data inconsistency and bias are inevitable among different facial expression recognition (FER) datasets due to subjective annotating process and different collecting conditions. Recent works resort to adversarial mechanisms that learn domain-invariant features to mitigate domain shift. However, most of these works focus on holistic feature adaptation, and they ignore local features that are more transferable across different datasets. Moreover, local features carry more detailed and discriminative content for expression recognition, and thus integrating local features may enable finegrained adaptation. In this work, we propose a novel Adversarial Graph Representation Adaptation (AGRA) framework that unifies graph representation propagation with adversarial learning for cross-domain holistic-local feature co-adaptation. To achieve this, we first build a graph to correlate holistic and local regions within each domain and another graph to correlate these regions across different domains. Then, we learn the per-class statistical distribution of each domain and extract holistic-local features from the input image to initialize the corresponding graph nodes. Finally, we introduce two stacked graph convolution networks to propagate holistic-local feature within each domain to explore their interaction and across different domains for holistic-local feature co-adaptation. In this way, the AGRA framework can adaptively learn fine-grained domain-invariant features and thus facilitate cross-domain expression recognition. We conduct extensive and fair experiments on several popular benchmarks and show that the proposed AGRA framework achieves superior performance over previous state-of-the-art methods.

show abstract

“…Apart from MMD, Zhang et al [22] define a new divergence, Margin Disparity Discrepancy (MDD), and validate that it has rigorous generalization bounds. SAFN in [39] not only utilizes norm to quantitatively measure domain statistics but also suggests that larger norm features can boost knowledge transfer.…”

Section: Discrepancy Metric Minimizationmentioning

confidence: 99%

Domain Conditioned Adaptation Network

Liu

Lin

et al. 2020

AAAI

View full text Add to dashboard Cite

Tremendous research efforts have been made to thrive deep domain adaptation (DA) by seeking domain-invariant features. Most existing deep DA models only focus on aligning feature representations of task-specific layers across domains while integrating a totally shared convolutional architecture for source and target. However, we argue that such strongly-shared convolutional layers might be harmful for domain-specific feature learning when source and target data distribution differs to a large extent. In this paper, we relax a shared-convnets assumption made by previous DA methods and propose a Domain Conditioned Adaptation Network (DCAN), which aims to excite distinct convolutional channels with a domain conditioned channel attention mechanism. As a result, the critical low-level domain-dependent knowledge could be explored appropriately. As far as we know, this is the first work to explore the domain-wise convolutional channel activation for deep DA networks. Moreover, to effectively align high-level feature distributions across two domains, we further deploy domain conditioned feature correction blocks after task-specific layers, which will explicitly correct the domain discrepancy. Extensive experiments on three cross-domain benchmarks demonstrate the proposed approach outperforms existing methods by a large margin, especially on very tough cross-domain learning tasks.

show abstract

Larger Norm More Transferable: An Adaptive Feature Norm Approach for Unsupervised Domain Adaptation

Cited by 469 publications

References 38 publications

Learning Target-Domain-Specific Classifier for Partial Domain Adaptation

Learning Target-Domain-Specific Classifier for Partial Domain Adaptation

Adversarial Graph Representation Adaptation for Cross-Domain Facial Expression Recognition

Domain Conditioned Adaptation Network

Contact Info

Product

Resources

About