Riemannian representation learning for multi-source domain adaptation

Article Computer Science, Artificial Intelligence

Unsupervised Domain Adaptation via Deep Conditional Adaptation Network

Pengfei Ge et al.

Summary: Unsupervised domain adaptation (UDA) aims to generalize the supervised model trained on a source domain to an unlabeled target domain. To address the conditional dependence between features and labels as well as negative transfer, this paper proposes a Deep Conditional Adaptation Network (DCAN) that aligns the conditional distributions using Conditional Maximum Mean Discrepancy and extracts discriminant information from the target domain using mutual information.

PATTERN RECOGNITION (2023)

添加到收藏夹

Article Computer Science, Artificial Intelligence

BP-triplet net for unsupervised domain adaptation: A Bayesian perspective

Shanshan Wang et al.

Summary: This paper proposes a BP-Triplet Loss method based on Bayesian learning for unsupervised domain adaptation. The method adjusts the weights of sample pairs, self attends to hard sample pairs, and improves the quality of target pseudo labels through adversarial loss. Experimental results demonstrate the effectiveness of the proposed approach for unsupervised domain adaptation.

PATTERN RECOGNITION (2023)

添加到收藏夹

Article Computer Science, Artificial Intelligence

A Two-Way alignment approach for unsupervised multi-Source domain adaptation

Yong-Hui Liu et al.

Summary: The research introduces a Two-Way alignment framework for Multi-Source Domain Adaptation (MDA), which aligns domain-level and category-level information, addressing instance variations.

PATTERN RECOGNITION (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Domain consistency regularization for unsupervised multi-source domain adaptive classification

Zhipeng Luo et al.

Summary: This paper proposes an approach for deep learning-based multi-source unsupervised domain adaptation (MUDA) that focuses on domain consistency regularization and authorization strategy for distribution alignment and decision boundary optimization. Experimental results demonstrate the superior adaptation performance of the approach across multiple MUDA datasets.

PATTERN RECOGNITION (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Dynamic Instance Domain Adaptation

Zhongying Deng et al.

Summary: Most existing studies on unsupervised domain adaptation assume that each domain's training samples come with domain labels, and feature alignment is performed based on these labels. However, this assumption is not valid in finer-grained domains. This paper proposes a dynamic instance domain adaptation method, which adapts deep features to each individual instance by generating instance-adaptive residuals.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2022)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Multi-Source Unsupervised Domain Adaptation via Pseudo Target Domain

Chuan-Xian Ren et al.

Summary: The proposed PTMDA method shows excellent performance in MDA tasks, effectively utilizing information correlation among multiple source domains to map the source domains into the target domain. By using pseudo target domains, PTMDA improves the performance on the real target domain.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2022)

添加到收藏夹

Review Computer Science, Artificial Intelligence

A Review of Domain Adaptation without Target Labels

Wouter M. Kouw et al.

Summary: This review categorizes approaches in domain adaptation into sample-based, feature-based, and inference-based methods, highlighting the importance of conditions for formulating bounds on cross-domain generalization error.

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Deep Subdomain Adaptation Network for Image Classification

Yongchun Zhu et al.

Summary: This study introduces a deep subdomain adaptation network (DSAN) that aligns relevant subdomain distributions across different domains based on the local maximum mean discrepancy (LMMD). DSAN is simple but effective, does not require adversarial training, and converges quickly. It can be easily integrated into feedforward network models to achieve efficient adaptation via backpropagation.

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence

Attention-Based Multi-Source Domain Adaptation

Yukun Zuo et al.

Summary: ABMSDA mitigates the negative effects caused by dissimilar domains by utilizing attention mechanism and domain correlations, leading to improved performance in the target domain.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2021)

添加到收藏夹

Article Computer Science, Artificial Intelligence