Unsupervised Domain Adaptation using Deep Networks with Cross-Grafted Stacks

Hou, Jinyong; Ding, Xuejie; Deng, Jeremiah D.; Cranefield, Stephen

doi:10.1109/iccvw.2019.00407

Cited by 7 publications

(7 citation statements)

References 28 publications

(44 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

Section: Uda With Data Augmentationmentioning

confidence: 99%

“…Variational Autoencoder (VAE) has been a prevalent generative model for data generation and it has been used for UDA in literature (Chen, Chen, Jin, Liu, & Cheng, 2019;Hou, Ding, Deng, & Cranefield, 2019;Hsu et al, 2017;Ilse, Tomczak, Louizos, & Welling, 2020;Xu et al, 2020). Hou et al (2019) aim to generate synthetic target-domain data with VAEs trained domain-wisely. Subsequently, the higher-level and lowerlevel layers of the decoders for source and target domains are cross-stacked to form new VAEs which can be used to transform images from one domain to the other.…”

Section: Uda With Data Augmentationmentioning

confidence: 99%

“…Subsequently, the higher-level and lowerlevel layers of the decoders for source and target domains are cross-stacked to form new VAEs which can be used to transform images from one domain to the other. However, the effectiveness of the idea was only validated on digits data in Hou et al (2019) and is questionable for more complicated image classification tasks. In contrast to pixel-level image generation, a more reliable alternative is employed in our work which aims to generate image features with a simplified VAE model.…”

Section: Uda With Data Augmentationmentioning

confidence: 99%

See 2 more Smart Citations

Data augmentation with norm-AE and selective pseudo-labelling for unsupervised domain adaptation

2023

View full text Add to dashboard Cite

Section: Uda With Data Augmentationmentioning

confidence: 99%

Section: Uda With Data Augmentationmentioning

confidence: 99%

Section: Uda With Data Augmentationmentioning

confidence: 99%

See 1 more Smart Citation

Data augmentation with norm-AE and selective pseudo-labelling for unsupervised domain adaptation

2023

View full text Add to dashboard Cite

“…There has been a large body of research showing that approaches using generative models are useful for improving a model's generalization performance across domains [7]. Generative models mimic real-world data distributions to generate data in the target domain for adaptation, especially using generative adversarial neural networks (GANs) or autoencoder structures [7][8][9][10][11][12][13].…”

Section: Introductionmentioning

confidence: 99%

Bidirectional-Feature-Learning-Based Adversarial Domain Adaptation with Generative Network

Han,

Choo,

Jeong

2023

Applied Sciences

View full text Add to dashboard Cite

Studying domain adaptation is a recent research trend. Generally, many generative models that researchers have studied perform well on training data from a specific domain. However, their ability to be generalized to other domains might be limited. Therefore, a growing body of research has utilized domain adaptation techniques to address the problem of generative models being vulnerable to input from other domains. In this paper, we focused on generative models and representation learning. Generative models have received a lot of attention for their ability to generate various types of data such as images, music, and text. In particular, studies utilizing generative adversarial neural networks (GANs) and autoencoder structures have received a lot of attention. In this paper, we solved the domain adaptation problem by reconstructing real image data using an autoencoder structure. In particular, reconstructed image data, considered a type of noisy image data, are used as input data. How to reconstruct data by extracting features and selectively transforming them in order to reduce differences in characteristics between domains entails representative learning. Considering these research trends, this paper proposed a novel methodology combining bidirectional feature learning and generative networks to innovatively approach the domain adaptation problem. It could improve the adaptation ability by accurately simulating the real data distribution. The experimental results show that the proposed model outperforms the traditional DANN and ADDA. This demonstrates that combining bidirectional feature learning and generative networks is an effective solution in the field of domain adaptation. These results break new ground in the field of domain adaptation. They are expected to provide great inspiration for future research and applications. Finally, through various experiments and evaluations, we verify that the proposed approach outperforms the existing works. We conducted experiments for representative generative models and domain adaptation techniques and found that the proposed approach was effective in improving data and domain robustness. We hope to contribute to the development of domain-adaptive models that are robust to the domain.

show abstract

“…This paper is extended from a conference publication [15], where the CGGS-based framework was first presented and some preliminary experiment results were provided. Here in this work we give a detailed presentation of the entire framework, with the following additional technical and experimental contents introduced: (1) A theoretical explanation of the generation of transition spaces as obtained from a probabilistic weights perturbation perspective; (2) new experimental results on more benchmark datasets to expand the empirical evaluation of DATL against the state of the art, plus the new results on cross-task generalization; and (3) further analysis and visual evaluation of the DATL framework.…”

Section: Introductionmentioning

confidence: 99%

Deep Adversarial Transition Learning using Cross-Grafted Generative Stacks

Hou,

Ding,

Cranefield

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Current deep domain adaptation methods used in computer vision have mainly focused on learning discriminative and domain-invariant features across different domains. In this paper, we present a novel "deep adversarial transition learning" (DATL) framework that bridges the domain gap by projecting the source and target domains into intermediate, transitional spaces through the employment of adjustable, cross-grafted generative network stacks and effective adversarial learning between transitions. Specifically, we construct variational auto-encoders (VAE) for the two domains, and form bidirectional transitions by cross-grafting the VAEs' decoder stacks. Furthermore, generative adversarial networks (GAN) are employed for domain adaptation, mapping the target domain data to the known label space of the source domain. The overall adaptation process hence consists of three phases: feature representation learning by VAEs, transitions generation, and transitions alignment by GANs. Experimental results demonstrate that our method outperforms the state-of-theart on a number of unsupervised domain adaptation benchmarks.

show abstract

Unsupervised Domain Adaptation using Deep Networks with Cross-Grafted Stacks

Cited by 7 publications

References 28 publications

Data augmentation with norm-AE and selective pseudo-labelling for unsupervised domain adaptation

Data augmentation with norm-AE and selective pseudo-labelling for unsupervised domain adaptation

Bidirectional-Feature-Learning-Based Adversarial Domain Adaptation with Generative Network

Deep Adversarial Transition Learning using Cross-Grafted Generative Stacks

Contact Info

Product

Resources

About