computer-vision-5

Universal Domain Adaptation

3260 papers • 126 benchmarks • 313 datasets

This task has no description! Would you like to contribute one?

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in universal-domain-adaptation-5

Trend

Dataset

Best Model

Actions

Office-Home

Office-31

VisDA2017

Libraries

i

Use these libraries to find universal-domain-adaptation-5 models and implementations

ispc-lab/glc

3 papers 27

Datasets

Subtasks

No subtasks available.

Most implemented papers

Upcycling Models Under Domain and Category Shift

Changjun Jiang, Guang-Sheng Chen, Sanqing Qu, Tianpei Zou, Florian Roehrbein, Cewu Lu, Dacheng Tao•Sun Mar 12 2023

This paper designs a novel, adaptive one-vs-all global clustering algorithm to achieve the distinction across different target classes and introduces a local k-NN clustering strategy to alleviate negative transfer.

61

Content

DomainNet

0

Paper Graph

Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation

Jiashi Feng, Jian Liang, Dapeng Hu•Wed Feb 19 2020

This work proposes a simple yet generic representation learning framework, named SHOT, which freezes the classifier module (hypothesis) of the source model and learns the target-specific feature extraction module by exploiting both information maximization and self-supervised pseudo-labeling to implicitly align representations from the target domains to the source hypothesis.

1539 0

Paper Graph

OVANet: One-vs-All Network for Universal Domain Adaptation

Kate Saenko, Kuniaki Saito•Tue Apr 06 2021

A method to learn the thresh-old using source samples and to adapt it to the target domain by minimizing class entropy and the resulting framework is the simplest of all baselines of UNDA and is insensitive to the value of a hyper-parameter, yet outperforms baselines with a large margin.

181 0

Paper Graph

LEAD: Learning Decomposition for Source-free Universal Domain Adaptation

Florian Rohrbein, Changjun Jiang, Sanqing Qu, Tianpei Zou, Lianghua He, Alois Knoll, Guang Chen•Tue Mar 05 2024

This paper proposes a new idea of LEArning Decomposition (LEAD), which decouples features into source-known and-unknown components to identify target-private data and is complementary to most existing methods.

45 0

Paper Graph

GLC++: Source-Free Universal Domain Adaptation Through Global-Local Clustering and Contrastive Affinity Learning

Florian Rohrbein, Changjun Jiang, Sanqing Qu, Tianpei Zou, Cewu Lu, Guang Chen, Dacheng Tao•Wed Mar 20 2024

This paper proposes a novel Global and Local Clustering (GLC) technique, which comprises an adaptive one-vs-all global clustering algorithm to discern between target classes, complemented by a local k-NN clustering strategy to mitigate negative transfer.

7 0

Paper Graph

Universal Domain Adaptation through Self Supervision

Kate Saenko, S. Sclaroff, Kuniaki Saito•Tue Feb 18 2020

This work proposes a more universally applicable domain adaptation approach that can handle arbitrary category shift, called Domain Adaptative Neighborhood Clustering via Entropy optimization (DANCE), and uses entropy-based feature alignment and rejection to align target features with the source, or reject them as unknown categories based on their entropy.

371 0

Paper Graph

Divergence Optimization for Noisy Universal Domain Adaptation

Y. Ushiku, Atsushi Hashimoto, Qing Yu•Wed Mar 31 2021

A two-head convolutional neural network framework to solve all problems simultaneously that can detect noisy source samples, find "unknown" classes in the target domain, and align the distribution of the source and target domains.

30 0

Paper Graph

On Universal Black-Box Domain Adaptation

K. Jia, Yabin Zhang, Bin Deng, Hui Tang, Changxing Ding•Fri Apr 09 2021

By accessing the interface of source model only, this framework outperforms existing methods of universal domain adaptation that make use of source data and/or source models, with a newly proposed (and arguably more reasonable) metric of H-score, and performs on par with them with the metric of averaged class accuracy.

11 0

Paper Graph

Domain Consensus Clustering for Universal Domain Adaptation

Yunchao Wei, Yi Yang, Guangrui Li, Guoliang Kang, Yi Zhu•Mon May 31 2021

In this paper, we investigate Universal Domain Adaptation (UniDA) problem, which aims to transfer the knowledge from source to target under unaligned label space. The main challenge of UniDA lies in how to separate common classes (i.e., classes shared across domains), from private classes (i.e., classes only exist in one domain). Previous works treat the private samples in the target as one generic class but ignore their intrinsic structure. Consequently, the resulting representations are not compact enough in the latent space and can be easily confused with common samples. To better exploit the intrinsic structure of the target domain, we propose Domain Consensus Clustering (DCC), which exploits the domain consensus knowledge to discover discriminative clusters on both common samples and private ones. Specifically, we draw the domain consensus knowledge from two aspects to facilitate the clustering and the private class discovery, i.e., the semantic-level consensus, which identifies the cycle-consistent clusters as the common classes, and the sample-level consensus, which utilizes the cross-domain classification agreement to determine the number of clusters and discover the private classes. Based on DCC, we are able to separate the private classes from the common ones, and differentiate the private classes themselves. Finally, we apply a class-aware alignment technique on identified common samples to minimize the distribution shift, and a prototypical regularizer to inspire discriminative target clusters. Experiments on four benchmarks demonstrate DCC significantly outperforms previous state-of-the-arts.

142 0

Paper Graph

Distance-based Hyperspherical Classification for Multi-source Open-Set Domain Adaptation

T. Tommasi, Barbara Caputo, S. Bucci, Francesco Cappio Borlino•Sun Jul 04 2021

This work tackles multi-source Open-Set domain adaptation by introducing HyMOS: a straightforward model that exploits the power of contrastive learning and the properties of its hyperspherical feature space to correctly predict known labels on the target, while rejecting samples belonging to any unknown class.

20 0

Paper Graph

Adding a benchmark result helps the community track progress.