zhaoxin94 / awesome-domain-adaptation Goto Github PK

Hello @zhaoxin94,

Thanks for the great DA paper list. Could you add the ICML 2021 paper : "Unbalanced minibatch Optimal Transport; applications to Domain Adaptation" to the list please ? It is an optimal transport method which was on both DA and partial DA.

The link to the paper is https://arxiv.org/abs/2103.03606 and to the code is https://github.com/kilianFatras/JUMBOT

Thank you very much.

Hi Xin,
Thanks for your effort to this repo. We now have a new ACCV2022 paper for Unsupervised DA.
You can put this paper in Unsupervised DA - Other methods category.

Paper link: https://openaccess.thecvf.com/content/ACCV2022/html/Wang_Revisiting_Unsupervised_Domain_Adaptation_Models_a_Smoothness_Perspective_ACCV_2022_paper.html

Code link: (Pytorch)
https://github.com/Wang-Xiaodong1899/LeCo_UDA

Thanks very much!
Xiaodong

Hi, thank you for your amazing work. Could you add this paper to the repo?
Paper is accepted to Image and Vision Computing journal.
Title: An Unsupervised Domain Adaptation Scheme for Single-Stage Artwork Recognition in Cultural Sites
Paper: https://arxiv.org/abs/2008.01882v3
Dataset: https://iplab.dmi.unict.it/EGO-CH-OBJ-UDA/
Code: https://github.com/fpv-iplab/DA-RetinaNet

Thank you :D

Greetings!

Could you please include the following two papers on your list?
https://github.com/donglao/PCEDA
https://github.com/YanchaoYang/FDA

Cheers

Hi, thank you for organizing such a great repo.
I would like to suggest our CVPR 2021 work to this collection.

Visualizing Adapted Knowledge in Domain Transfer

arxiv: https://arxiv.org/abs/2104.10602
code: https://github.com/hou-yz/DA_visualization
section: Explainable (new); Image-to-Image Translation; Source-Free Domain Adaptation

This is the first attempt at explaining and visualizing the transferred knowledge during domain adaptation. We believe it would make a fine addition to a new Explainable section in the collection. In addition, it can also be deemed as a new approach under the Image-to-Image Translation section, and also helps to improve Source-Free Domain Adaptation performance.

Thank you for your effort!

Thanks for this awesome repo. Please add the following ICRA 2022 paper on UDA:

Unsupervised Domain Adaptation in LiDAR Semantic Segmentation with Self-Supervision and Gated Adapters, ICRA 2022 (accepted).
arxiv: https://arxiv.org/abs/2107.09783

Hi：
Thanks for your contributions. We now have a SOTA MSDA paper that was published in TPAMI:
[Graphical Modeling for Multi-Source Domain Adaptation] [pytorch]. This paper is the extension of our ECCV2020 paper, Learning to Combine: Knowledge Aggregation for Multi-Source Domain Adaptation.

We wish the work could be added to MSDA.
Thanks very much!

Hi,

Please, add the paper Attending to Discriminative Certainty for Domain Adaptation which is accepted in CVPR 2019.

Thanks,
Shanu

Hello, I have a preprint on SSDA here - https://arxiv.org/abs/2211.11975 - I was wondering how can I add it to this page?
Thanks for your time and help,
Megh

Some papers on heterogeneous domain adaptation mention this word, but no good explanation is given. Does anyone understand the meaning of this word.

Paper: Unsupervised Heterogeneous Domain Adaptation with Sparse Feature Transformation

context:

1. Some semi-supervised HDA methods even utilize parallel unlabeled instances to learn cross-domain representations
2. A few unsupervised HDA approaches overcome this dependence limitation on labeled target data by learning a common latent correlation subspace based only on parallel instances
3. he method uses a linear function to transform the source domain features into the target domain features to match the parallel instances, while minimizing the cross domain distribution divergence by aligning the transformed source domain covariance matrix with the target domain covariance matrix.

Thanks for creating this repository! It's a very comprehensive source of information.

Could you please add our ICML 2023 paper, POUF: Prompt-oriented unsupervised fine-tuning for large pre-trained models
? The code is provided in this link?

We appreciate your help. Thanks!

Hello, could you add the github source code link, "Domain Specific Batch Normalization for Unsupervised Domain Adaptation" accepted in CVPR 2019 ?
URL: https://github.com/wgchang/DSBN

Many thanks.
WG. Chang

Thank you for the continuous updates of your amazing repository!
I would like to suggest some new papers here:

Semantic Segmentation:

Plugging Self-Supervised Monocular Depth into Unsupervised Domain Adaptation for Semantic Segmentation [WACV 2022]

Shallow Features Guide Unsupervised Domain Adaptation for Semantic Segmentation at Class Boundaries [WACV 2022]

Moreover, I suggest also this paper which focus on UDA for Point Cloud Classification:

RefRec: Pseudo-labels Refinement via Shape Reconstruction for Unsupervised 3D Domain Adaptation [3DV 2021 Oral]

Thank you again for your effort!

Thank you for maintaining the paper lists!

We woulid like to request to add our NeurIPS2022 paper: "Meta-DMoE: Adapting to Domain Shift by Meta-Distillation from Mixture-of-Experts"

We utilize the multi-source paradigm and train multiple model experts for each of the source domian. At test-time, for each target domain, some unlabeled data are sampled to query the knowledge from the experts to distill the domain knowledge to a student network. The student network is then used for downstream inference.

We think it might be suitable for multi-source DA, source-free DA, few-shot UDA or domain generalization.

Thank you!

Thank you for your awesome repository for the researcher's references.
I would like to suggest some new papers here.

Optimal Transport for single UDA:

• TIDOT: A Teacher Imitation Learning Approach for Domain Adaptation with Optimal Transport [IJCAI 2021]
• LAMDA: Label Matching Deep Domain Adaptation [ICML 2021]

Optimal Transport for Multi-source DA:

• MOST: Multi-Source Domain Adaptation via Optimal Transport for Student-Teacher Learning [UAI 2021]

Multi-source DA:

• STEM: An approach to Multi-source Domain Adaptation with Guarantees [ICCV 2021]

Many thanks for your effort!

Hi Xin, could you please add the new ICCV-21 oral paper on video domain adaptation:
Partial Video Domain Adaptation with Partial Adversarial Temporal Attentive Network

with source code:
https://github.com/xuyu0010/PATAN

Thanks for creating this repository! It's a very comprehensive source of information.

Could you please add our NeurIPS 2021 paper, a prototype-oriented framework for unsupervised domain adaptation? The code is provided in this link?

We appreciate your help. Thanks!

Hi,
Thanks for your efforts in compiling and actively maintaining this list. Could you please add three of our recent works to this list:

Balancing Discriminability and Transferability for Source-Free Domain Adaptation [ICML2022] [Project Page]

Concurrent Subsidiary Supervision for Unsupervised Source-Free Domain Adaptation [ECCV2022] [Project Page]

Subsidiary Prototype Alignment for Universal Domain Adaptation [NeurIPS2022] [Project Page]

The first two can go in Source-Free DA section and third one in Universal DA.

Thanks again

Hi @zhaoxin94
We have a new UDA paper on ECCV-2020. The paper proposes a simple yet effective adversarial scheme to improve the original DANN (RevGrad) for domain adaptation.
Could you please help add it in this repo?

Name: Mind the Discriminability: Asymmetric Adversarial Domain Adaptation
Url: https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123690579.pdf

Hi：
Thanks for your contributions to this field. We now have a SOTA semi-DA paper that was published in Neural Networks Journal:
[Context-guided entropy minimization for semi-supervised domain adaptation] [pytorch]

We wish the work could be added to Semi-supervised DA or source-free DA.
Thanks very much!

Hi,

Thanks for maintaining this repository. I have taken lots of inspiration from the papers you've collated!

I was wondering if you could add our work on multi-source DA that achieves SOTA in UDA for 3D object detection. We are the first approach in UDA for 3D object detection to leverage multiple pre-trained detectors and is able to generalize to lidars of any point cloud density. Our work shows significant improvements over state of the art.

Title: MS3D++: Ensemble of Experts for Multi-Source Unsupervised Domain Adaptation in 3D Object Detection
Paper link: https://arxiv.org/abs/2308.05988
Code: https://github.com/darrenjkt/ms3d

Hi Xin,

The repository is very impressive. Can you please add my IEEE Transactions on Multimedia article on open-set DA in your repository?

Title: Adversarial Network with Multiple Classifiers for Open Set Domain Adaptation
Authors: Tasfia Shermin, Guojun Lu, Shyh Wei Teng, Manzur Murshed, Ferdous Sohel
Publisher: IEEE TMM
Year: 2020
Code: https://github.com/tasfia/DAMC

Regards,
Tasfia

Thank you for maintaining this repository!
I would like to suggest a new conference paper belonging to Adversarial-based UDA (CVPR2022).
Title: Reusing the Task-specific Classifier as a Discriminator: Discriminator-free Adversarial Domain Adaptation
Link: https://arxiv.org/abs/2204.03838
Code: https://github.com/xiaoachen98/DALN

Hi, this project is a nice job for newbee.

There exist some code repo that is missing in this project.

Diversify and Match: A Domain Adaptive Representation Learning Paradigm for Object Detection; Code at https://github.com/TKKim93/DivMatch

Adapting Object Detectors via Selective Cross-Domain Alignment; Code at https://github.com/xinge008/SCDA

Hi!

thx for your helpfull and always updated work!
I'd like to point out our new papers accepted at ECCV 2022:

1. GIPSO: we investigate the new adaptation field of Source-Free Online Domain Adaptation (SF-OUDA) for point cloud segmentation. We show that existing works in offline UDA or SF-UDA do not suffice for the online settings and propose the first online approach based on geometrically informad propagation (GIPSO).

• Title: GIPSO: Geometrically Informed Propagation for Online Adaptation in 3D LiDAR Segmentation
• Paper: https://arxiv.org/abs/2207.09763
• Code: https://github.com/saltoricristiano/gipso-sfouda

2. CoSMix: we propose the first mixup based strategy for UDA in point cloud segmentation. Here, adaptation is achieved through a double branch mixing strategy making use of the semantic information and pseudo-labels. CoSMix achieves SOTA results in synthetic to real UDA!

• Title: CoSMix: Compositional Semantic Mix for Domain Adaptation in 3D LiDAR Segmentation
• Paper: https://arxiv.org/abs/2207.09778
• Code: https://github.com/saltoricristiano/cosmix-uda

Thanks in advance and best regards! 🚀

Hi, @zhaoxin94, thx for sharing. Could you please add our new TPAMI 2023 paper on Domain Adaptive Semantic Segmentation: SePiCo: Semantic-Guided Pixel Contrast for Domain Adaptive Semantic Segmentation, the link is https://ieeexplore.ieee.org/document/10018569

with arxiv version:
https://arxiv.org/abs/2204.08808

and source code:
https://github.com/BIT-DA/SePiCo

Thank you very much!

Thanks for this amazingly informative repository.

Kindly add this new paper "Generalize Then Adapt: Source-Free Domain Adaptive Semantic Segmentation" recently published in ICCV2021. Also linking the project page where the PyTorch code link will be updated soon.

Once again, thanks for your efforts in maintaining this awesome repo!

Thank you for maintaining this repository!
I would like to suggest a new conference paper belonging to UDA Semantic Segmentation (NeurIPS2022).
Title: Deliberated Domain Bridging for Domain Adaptive Semantic Segmentation
Link: https://arxiv.org/abs/2209.07695
Code: https://github.com/xiaoachen98/DDB

Thank you very much for maintaining this excellent repo.

Would you mind adding our ECCV 2022 paper "DecoupleNet: Decoupled Network for Domain Adaptive Semantic Segmentation"?
The paper link is https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136930362.pdf
The arXiv link is https://arxiv.org/pdf/2207.09988.pdf
The Code link is https://github.com/dvlab-research/DecoupleNet

Our work aims to solve the UDA Semantic Segmentation problem and achieves strong performance. We also propose a method based on self-training to further boost performance. Moreover, it can generalize well on the UDA classification task.

So I think it should to assigned to both the UDA Semantic Segmentation and Self-training-based categories.

Thank you very much!

Thank you for meticulously maintaining this repository!
I would like to suggest a new papers here.
Title: Probabilistic Contrastive Learning for Domain Adaptation
Link: https://arxiv.org/abs/2111.06021
Code: https://github.com/ljjcoder/Probabilistic-Contrastive-Learning

In this paper, authors point out that feature contrastive learning is inferior to probabilistic contrastive learning in the domain adaptation task and demonstrate the effectiveness of probabilistic contrastive learning on multiple tasks (UDA, SSDA, SSL and UDA detection).

Many thanks for your effort!

Hi Xin,

Thanks for your effort to maintain this repository!

Recently, we have a new paper for unsupervised domain adaptation Global-Local Regularization Via Distributional Robustness that has been accepted at the International Conference on Artificial Intelligence and Statistics (AISTATS 2023). Official release source code (Pytorch) could be found here.

We think it might be suitable for Optimal Transport-based DA or Domain generalization.

Thanks in advance and best regards,
Hoang Phan

Hi,

Please, add the paper "Curriculum based Dropout Discriminator for Domain Adaptation" which is accepted in BMVC 2019.

Project page and code repo is : https://delta-lab-iitk.github.io/CD3A/

Thanks,

Hi. Thank you for your great efforts to maintain such an unbelievable exhaustive list for domain adaptation research.

Our paper, "Partially-Shared Variational Auto-encoders for Unsupervised Domain Adaptation with Target Shift", is currently listed as an arXiv paper, but it appeared in ECCV2020.
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2472_ECCV_2020_paper.php

We are really happy if you could update the list.
Thank you very much again.

Hi Xin,

Thanks for your excellent work! The repository is very impressive and helpful! Could you please add Our NeurIPS2020 paper on One-shot UDA ( also can be regarded as an Unsupervised Domain-adaptive Semantic Segmentation ) in your repository?

Title: Adversarial Style Mining for One-shot Unsupervised Domain Adaptation
Authors: Yawei Luo, Ping Liu, Tao Guan, Junqing Yu, Yi Yang
Publisher: NeurIPS
Year: 2020
Code: https://github.com/RoyalVane/ASM

Best Regards,
Yawei

Thank you for maintaining this repository!
I would like to suggest a new conference paper belonging to Source Free DA (IJCAI2021).
Title: Source-free Domain Adaptation via Avatar Prototype Generation and Adaptation
Link: https://arxiv.org/abs/2106.15326
Code: https://github.com/SCUT-AILab/CPGA

I read a lot of literatures, most of which are based on the same feature space (the same dimension). How to use the DA method when the dimensions are different, and what keywords should be searched.

Thank you for your huge efforts in maintaining such a useful repository.

Can you pease add our CVPR 2021 Oral paper on Domain Generalisation:
Domain Generalization via Inference-time Label-Preserving Target Projections

https://arxiv.org/abs/2103.01134

Hi, I just found this repository，that's awesome ! Could you please add one early work Multi-Weight Partial Domain Adaptation(BMVC2019)(https://bmvc2019.org/wp-content/uploads/papers/0406-paper.pdf) to the partial DA part? Thank you~

hi @zhaoxin94
Please add the following paper related to Domain Adaptation
Paper: https://openaccess.thecvf.com/content/WACV2021/papers/Kurmi_Domain_Impression_A_Source_Data_Free_Domain_Adaptation_Method_WACV_2021_paper.pdf
Project page: https://delta-lab-iitk.github.io/SFDA/

Thanks

Hello, could you please add our new paper "Domain Conditioned Adaptation Network" that is accepted in AAAI2020

Domain Conditioned Adaptation Network AAAI2020 Pytorch

Many thanks,
Xie

Hi, our group published the new ADA paper, "Reducing Annotation Effort by Identifying and Labeling Contextually Diverse.
Classes for Semantic Segmentation Under Domain Shift" WACV23.
https://openaccess.thecvf.com/content/WACV2023/papers/Agarwal_Reducing_Annotation_Effort_by_Identifying_and_Labeling_Contextually_Diverse_Classes_WACV_2023_paper.pdf

Hi Xin, thank you very much for your effort in collating the numerous domain adaptation papers, which have helped me in my research. We have recently surveyed video domain adaptation and would like to request to add this new survey paper:
Video Unsupervised Domain Adaptation with Deep Learning: A Comprehensive Survey, the link to this paper is: https://arxiv.org/abs/2211.10412.

with the relevant repository:
https://github.com/xuyu0010/awesome-video-domain-adaptation

Thank you very much!

Thank you for meticulously maintaining this repository!

Could you add our recent work on unsupervised domain adaptation (UDA) for LiDAR segmentation? We built the first benchmark for UDA in LiDAR segmentation (range-view) and tested other related DA cases with minimum supervisions, such as SSDA and WSDA. We also proposed a new algorithm that has achieved promising results.

• Title: ConDA: Unsupervised Domain Adaptation for LiDAR Segmentation via Regularized Domain Concatenation
• Link: https://arxiv.org/abs/2111.15242
• Authors: Lingdong Kong, Niamul Quader, Venice Erin Liong, Hanwang Zhang
• Affiliations: Nanyang Technological University, Motional (nuTonomy)
• Tags: seg, uda, av, lidar

Hi,

Could you please add our new work into this list? The paper is about Unsupervised Domain Adaptation via Disentangled Representations.

You could find the paper here: https://arxiv.org/abs/1907.13590

Thank you!

Best,
Junlin

hi @zhaoxin94
Please add the following paper related to Domain Adaptation
Paper: https://arxiv.org/abs/1904.01341
Project page: https://vinodkkurmi.github.io/DiscriminatorDomainAdaptation/

Thanks

Please add following two papers on Unsupervised Domain Adaptation in your awesome repo:

1. https://link.springer.com/chapter/10.1007/978-3-030-58539-6_25 (ECCV 2020 Spotlight)
   Title : Unsupervised Domain Adaptation for Semantic Segmentation of NIR Images through Generative Latent Search
   Code: https://github.com/ambekarsameer96/GLSS

2. https://ieeexplore.ieee.org/document/9139471 (IEEE TMI 2020)
   Title: Target-Independent Domain Adaptation for WBC Classification using Generative Latent Search
   Code: https://github.com/prinshul/WBC-Classification-UDA

The above methods are unique in their design - they use test-time optimisation.

Thank you.

Hi Xin, could you please add our new ECCV-21 oral paper on video domain adaptation under the source-free setting:
Source-free Video Domain Adaptation by Learning Temporal Consistency for Action Recognition, the link is: https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136940144.pdf

with source code:
https://github.com/xuyu0010/ATCoN

and project page:
https://xuyu0010.github.io/sfvda.html

Thank you very much!

Hi, could you please add our cvpr 2021 work https://github.com/microsoft/ProDA into the repo? Thanks for this amazing collection!

Thank you for meticulously maintaining this repository!

We would like to recommend our recent work, TranSVAE, a disentanglement framework for unsupervised video domain adaptation, which has achieved SoTA performance among the UDA leaderboards of UCF-HMDB, Jester, and Epic-Kitchens.

• Title: Unsupervised Video Domain Adaptation: A Disentanglement Perspective
• Paper: https://arxiv.org/abs/2208.07365
• Code: https://github.com/ldkong1205/TranSVAE
• Project Page: https://ldkong.com/TranSVAE
• Gradio Demo: https://huggingface.co/spaces/ldkong/TranSVAE

Thanks again and best regards 🚀

Hi,

would you please add our new paper HoMM: Higher-order Moment Matching for Unsupervised Domain Adaptation which is accepted in AAAI-2020.

btw, our method is a Distance-based Methods

official codes are avalible at https://github.com/chenchao666/HoMM-Master

Thanks,
chao