Correlation-aware adversarial domain adaptation and generalization

Rahman, Mohammad Mahfujur; Fookes, Clinton; Baktashmotlagh, Mahsa; Sridharan, Sridha

doi:10.1016/j.patcog.2019.107124

Cited by 110 publications

(38 citation statements)

References 31 publications

(100 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Kernel methods [20,47,48,21,49,50,51,52,53,22] Explicit feature alignment [54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71] Domain adversarial learning [72,73,59,74,75,76,77,41,78,79,23,80,81,82] Invariant risk minimization [83] Feature disentanglement…”

Section: Domain-invariant Representation Learningmentioning

confidence: 99%

Generalizing to Unseen Domains: A Survey on Domain Generalization

Wang

Lan

Liu

et al. 2021

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence

175

View full text Add to dashboard Cite

Domain generalization (DG), i.e., out-of-distribution generalization, has attracted increased interests in recent years. Domain generalization deals with a challenging setting where one or several different but related domain(s) are given, and the goal is to learn a model that can generalize to an unseen test domain. For years, great progress has been achieved. This paper presents the first review for recent advances in domain generalization. First, we provide a formal definition of domain generalization and discuss several related fields. Then, we categorize recent algorithms into three classes and present them in detail: data manipulation, representation learning, and learning strategy, each of which contains several popular algorithms. Third, we introduce the commonly used datasets and applications. Finally, we summarize existing literature and present some potential research topics for the future.

show abstract

Section: Domain-invariant Representation Learningmentioning

confidence: 99%

Generalizing to Unseen Domains: A Survey on Domain Generalization

Wang

Lan

Liu

et al. 2021

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence

175

View full text Add to dashboard Cite

show abstract

“…Another line of works optimizes for features that confuse a domain discriminator model (Albuquerque et al 2019;Shao et al 2019;Rahman et al 2020;Deng et al 2020), and includes DANN (Ganin et al 2016) and its class-conditional extension C-DANN (Li et al 2018c). Other works additionally involve the classifier in the representation alignment, either by optimizing for an embedding space such that the optimal linear classifier on top of it is the same across different domains (IRM) (Arjovsky et al 2019), or by passing a domain-specific mean embedding to the classifier as a second argument (MTL) (Blanchard et al 2017).…”

Section: Domain Generalizationmentioning

confidence: 99%

COLUMBUS: Automated Discovery of New Multi-Level Features for Domain Generalization via Knowledge Corruption

Frikha¹,

Krompaß²,

Tresp³

2021

Preprint

View full text Add to dashboard Cite

Machine learning models that can generalize to unseen domains are essential when applied in real-world scenarios involving strong domain shifts. We address the challenging domain generalization (DG) problem, where a model trained on a set of source domains is expected to generalize well in unseen domains without any exposure to their data. The main challenge of DG is that the features learned from the source domains are not necessarily present in the unseen target domains, leading to performance deterioration. We assume that learning a richer set of features is crucial to improve the transfer to a wider set of unknown domains. For this reason, we propose COLUMBUS, a method that enforces new feature discovery via a targeted corruption of the most relevant input and multi-level representations of the data. We conduct an extensive empirical evaluation to demonstrate the effectiveness of the proposed approach which achieves new state-of-the-art results by outperforming 18 DG algorithms on multiple DG benchmark datasets in the DOMAINBED framework.

show abstract

“…SUDA aims to learn a model well-performing on the target domain given a labeled source domain and an unlabeled target domain. It is usually achieved by discrepancy-based methods [6,19,20], adversarial learning [8,21,22,23,24], and self-training methods [25]. Tzeng et al [6] first proposed Maximum Mean Discrepancy (MMD) to measure the distance between domain distributions.…”

Section: Related Workmentioning

confidence: 99%

“…Zuo et al [21] applied different strategies to easy and tough examples to improve the domain adaptation performance. Rahman et al [22] combined correlation alignment with adversarial learning to tackle the domain adaptation and domain generalization problems. Liang et al [25] proposed to leverage the uncertainty of pseudo labels to achieve optimal feature transformation.…”

Section: Related Workmentioning

confidence: 99%

Domain Consistency Regularization for Unsupervised Multi-source Domain Adaptive Classification

Luo

Zhang²,

et al. 2021

Preprint

View full text Add to dashboard Cite

Deep learning-based multi-source unsupervised domain adaptation (MUDA) has been actively studied in recent years. Compared with single-source unsupervised domain adaptation (SUDA), domain shift in MUDA exists not only between the source and target domains but also among multiple source domains. Most existing MUDA algorithms focus on extracting domain-invariant representations among all domains whereas the task-specific decision boundaries among classes are largely neglected. In this paper, we propose an end-to-end trainable network that exploits domain Consistency Regularization for unsupervised Multi-source domain Adaptive classification (CRMA). CRMA aligns not only the distributions of each pair of source and target domains but also that of all domains. For each pair of source and target domains, we employ an intra-domain consistency to regularize a pair of domain-specific classifiers to achieve intra-domain alignment. In addition, we design an inter-domain consistency that targets joint inter-domain alignment among all domains. To address different similarities between multiple source domains and the target domain, we design an authorization strategy that assigns different authorities to domain-specific classifiers adaptively for optimal pseudo label prediction and self-training. Extensive experiments show that CRMA tackles unsupervised domain adaptation effectively

show abstract

Correlation-aware adversarial domain adaptation and generalization

Cited by 110 publications

References 31 publications

Generalizing to Unseen Domains: A Survey on Domain Generalization

Generalizing to Unseen Domains: A Survey on Domain Generalization

COLUMBUS: Automated Discovery of New Multi-Level Features for Domain Generalization via Knowledge Corruption

Domain Consistency Regularization for Unsupervised Multi-source Domain Adaptive Classification

Contact Info

Product

Resources

About