Existing unsupervised domain adaptation approaches primarily focus on reducing the data distribution gap between the source and target domains,often neglecting the influence of class information,leading to inaccurate ...Existing unsupervised domain adaptation approaches primarily focus on reducing the data distribution gap between the source and target domains,often neglecting the influence of class information,leading to inaccurate alignment outcomes.Guided by this observation,this paper proposes an adaptive inter-intra-domain discrepancy method to quantify the intra-class and inter-class discrepancies between the source and target domains.Furthermore,an adaptive factor is introduced to dynamically assess their relative importance.Building upon the proposed adaptive inter-intradomain discrepancy approach,we develop an inter-intradomain alignment network with a class-aware sampling strategy(IDAN-CSS)to distill the feature representations.The classaware sampling strategy,integrated within IDAN-CSS,facilitates more efficient training.Through multiple transfer diagnosis cases,we comprehensively demonstrate the feasibility and effectiveness of the proposed IDAN-CSS model.展开更多
Deep neural networks have been successfully applied to numerous machine learning tasks because of their impressive feature abstraction capabilities.However,conventional deep networks assume that the training and test ...Deep neural networks have been successfully applied to numerous machine learning tasks because of their impressive feature abstraction capabilities.However,conventional deep networks assume that the training and test data are sampled from the same distribution,and this assumption is often violated in real-world scenarios.To address the domain shift or data bias problems,we introduce layer-wise domain correction(LDC),a new unsupervised domain adaptation algorithm which adapts an existing deep network through additive correction layers spaced throughout the network.Through the additive layers,the representations of source and target domains can be perfectly aligned.The corrections that are trained via maximum mean discrepancy,adapt to the target domain while increasing the representational capacity of the network.LDC requires no target labels,achieves state-of-the-art performance across several adaptation benchmarks,and requires significantly less training time than existing adaptation methods.展开更多
基金the National Natural Science Foundation of China(Grant Nos.52275104,51905160)the Natural Science Fund for Excellent Young Scholars of Hunan Province(Grant No.2021JJ20017)。
文摘Existing unsupervised domain adaptation approaches primarily focus on reducing the data distribution gap between the source and target domains,often neglecting the influence of class information,leading to inaccurate alignment outcomes.Guided by this observation,this paper proposes an adaptive inter-intra-domain discrepancy method to quantify the intra-class and inter-class discrepancies between the source and target domains.Furthermore,an adaptive factor is introduced to dynamically assess their relative importance.Building upon the proposed adaptive inter-intradomain discrepancy approach,we develop an inter-intradomain alignment network with a class-aware sampling strategy(IDAN-CSS)to distill the feature representations.The classaware sampling strategy,integrated within IDAN-CSS,facilitates more efficient training.Through multiple transfer diagnosis cases,we comprehensively demonstrate the feasibility and effectiveness of the proposed IDAN-CSS model.
基金supported by the National Key R&D Program of China(No.2016YFB1200203)the National Natural Science Foundation of China(Nos.41427806 and 61273233)
文摘Deep neural networks have been successfully applied to numerous machine learning tasks because of their impressive feature abstraction capabilities.However,conventional deep networks assume that the training and test data are sampled from the same distribution,and this assumption is often violated in real-world scenarios.To address the domain shift or data bias problems,we introduce layer-wise domain correction(LDC),a new unsupervised domain adaptation algorithm which adapts an existing deep network through additive correction layers spaced throughout the network.Through the additive layers,the representations of source and target domains can be perfectly aligned.The corrections that are trained via maximum mean discrepancy,adapt to the target domain while increasing the representational capacity of the network.LDC requires no target labels,achieves state-of-the-art performance across several adaptation benchmarks,and requires significantly less training time than existing adaptation methods.
