Filters








323,834 Hits in 4.9 sec

Adaptive Methods for Aggregated Domain Generalization [article]

Xavier Thomas, Dhruv Mahajan, Alex Pentland, Abhimanyu Dubey
2021 arXiv   pre-print
In many settings, privacy concerns prohibit obtaining domain labels for the training data samples, and instead only have an aggregated collection of training points.  ...  using these pseudo-domains we learn a domain-adaptive classifier that makes predictions using information about both the input and the pseudo-domain it belongs to.  ...  in the domain generalization task by domain-adaptive classification.  ... 
arXiv:2112.04766v2 fatcat:xsi3whmwevfitjevpivjxkr2a4

Multi-source Domain Adaptation for Semantic Segmentation [article]

Sicheng Zhao, Bo Li, Xiangyu Yue, Yang Gu, Pengfei Xu, Runbo Hu, Hua Chai, Kurt Keutzer
2019 arXiv   pre-print
First, we generate an adapted domain for each source with dynamic semantic consistency while aligning at the pixel-level cycle-consistently towards the target.  ...  Second, we propose sub-domain aggregation discriminator and cross-domain cycle discriminator to make different adapted domains more closely aggregated.  ...  For each source domain, we generated adapted images with a novel dynamic semantic consistency loss.  ... 
arXiv:1910.12181v1 fatcat:djilqua37ndalbwm4ndj3ls3ja

MADAN: Multi-source Adversarial Domain Aggregation Network for Domain Adaptation [article]

Sicheng Zhao, Bo Li, Xiangyu Yue, Pengfei Xu, Kurt Keutzer
2020 arXiv   pre-print
First, an adapted domain is generated for each source with dynamic semantic consistency while aligning towards the target at the pixel-level cycle-consistently.  ...  For the segmentation adaptation, we further enforce category-level alignment and incorporate context-aware generation, which constitutes MADAN+.  ...  Further, we proposed a sub-domain aggregation discriminator and cross-domain cycle discriminator to better aggregate different adapted domains.  ... 
arXiv:2003.00820v1 fatcat:p3p2jyvurrb2zp4fefguunezp4

Seeking Similarities over Differences: Similarity-based Domain Alignment for Adaptive Object Detection [article]

Farzaneh Rezaeianaran, Rakshith Shetty, Rahaf Aljundi, Daniel Olmeda Reino, Shanshan Zhang, Bernt Schiele
2021 arXiv   pre-print
This has motivated research in Unsupervised Domain Adaptation (UDA) algorithms for detection.  ...  UDA methods learn to adapt from labeled source domains to unlabeled target domains, by inducing alignment between detector features from source and target domains.  ...  Our General Framework for UDA In this section we discuss our general framework for analyzing several aspects of unsupervised domain adaptation methods for object detection.  ... 
arXiv:2110.01428v1 fatcat:wa4f4ett6neorchltjbnnhqbni

Generalizable Person Re-identification with Relevance-aware Mixture of Experts [article]

Yongxing Dai, Xiaotong Li, Jun Liu, Zekun Tong, Ling-Yu Duan
2021 arXiv   pre-print
Besides, we design a voting network to adaptively integrate all the experts' features into the more generalizable aggregated features with domain relevance.  ...  domains for testing.  ...  As a result, our RaMoE method can generate very discriminative and generalizable aggregated features for the unseen target domains by adaptively integrating diverse domain experts with the domain relevance  ... 
arXiv:2105.09156v1 fatcat:cf3w2h22s5bcpnpgaaxah3eyqi

Strategies for adaptive optimization with aggregation constraints using interior-point methods

Graeme J. Kennedy
2015 Computers & structures  
We demonstrate that the proposed adaptive technique achieves significant computational savings compared to fixed-aggregation methods for a series of stress-constrained mass-minimization problems.  ...  In this paper, we present strategies to adaptively solve aggregation-constrained problems.  ...  For the adaptive KS aggregation method, the mass decreases as the aggregation parameter increases, while for the adaptive induced exponential aggregation method, the mass generally increases as the aggregation  ... 
doi:10.1016/j.compstruc.2015.02.024 fatcat:y72ent74ibhpvd5uzgmqzyfene

Reversing Gradients in Adversarial Domain Adaptation for Question Deduplication and Textual Entailment Tasks

Anush Kamath, Sparsh Gupta, Vitor Carvalho
2019 Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics  
We evaluate reversing gradients for adversarial adaptation on multiple domains, and demonstrate that it significantly outperforms other methods on question deduplication as well as on recognizing textual  ...  Adversarial domain adaptation has been recently introduced as an effective technique for textual matching tasks, such as question deduplication (Shah et al., 2018) .  ...  for Domain Adaptation Table 1: Comparison of Accuracy for different domain adaptation methods; Source domain for question duplicate detection: Quora (240k/ 80k/ 80k), Source domain for RTE: SNLI (550k  ... 
doi:10.18653/v1/p19-1556 dblp:conf/acl/KamathGC19 fatcat:foem3qumpndmfkyvxpiqqavutm

Exploiting Local Feature Patterns for Unsupervised Domain Adaptation [article]

Jun Wen, Risheng Liu, Nenggan Zheng, Qian Zheng, Zhefeng Gong, Junsong Yuan
2018 arXiv   pre-print
Unsupervised domain adaptation methods aim to alleviate performance degradation caused by domain-shift by learning domain-invariant representations.  ...  In this paper, we present a method for learning domain-invariant local feature patterns and jointly aligning holistic and local feature statistics.  ...  We are also partially supported by the Hunan Provincial Science and Technology Project Foundation (2018TP1018, 2018RS3065) and the Fundamental Research Funds for the Central Universities.  ... 
arXiv:1811.05042v2 fatcat:x4adfq4ydjhwxefrhqvtqlu7ge

Exploiting Local Feature Patterns for Unsupervised Domain Adaptation

Jun Wen, Risheng Liu, Nenggan Zheng, Qian Zheng, Zhefeng Gong, Junsong Yuan
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Unsupervised domain adaptation methods aim to alleviate performance degradation caused by domain-shift by learning domain-invariant representations.  ...  In this paper, we present a method for learning domain-invariant local feature patterns and jointly aligning holistic and local feature statistics.  ...  We are also partially supported by the Hunan Provincial Science and Technology Project Foundation (2018TP1018, 2018RS3065) and the Fundamental Research Funds for the Central Universities.  ... 
doi:10.1609/aaai.v33i01.33015401 fatcat:4vfeboxgtjhldilik3lyuvmwfa

Adaptive Boosting for Domain Adaptation: Towards Robust Predictions in Scene Segmentation [article]

Zhedong Zheng, Yi Yang
2021 arXiv   pre-print
We adopt one adaptive data sampler to progressively facilitate learning on hard samples and aggregate "weak" models to prevent over-fitting.  ...  adaptation methods, which can be combined with existing approaches to further improve the state-of-the-art performance.  ...  However, for domain adaptation, we can not foreknow the error e m on the unlabeled target-domain data.  ... 
arXiv:2103.15685v2 fatcat:ju5bi4w36jawzpeqpr2ksxuf4q

Adaptive Domain-Specific Normalization for Generalizable Person Re-Identification [article]

Jiawei Liu, Zhipeng Huang, Kecheng Zheng, Dong Liu, Xiaoyan Sun, Zheng-Jun Zha
2021 arXiv   pre-print
for learning more generalizable aggregated representation on unseen target domain.  ...  In this work, we propose a novel adaptive domain-specific normalization approach (AdsNorm) for generalizable person Re-ID.  ...  The compared methods are divided into three groups: supervised learning (S), unsupervised domain adaptation (UDA) and domain generalization (DG).  ... 
arXiv:2105.03042v2 fatcat:e4ukjdtdnrh7faizuljzjx3fri

Unsupervised Domain Adaptation for Video Semantic Segmentation [article]

Inkyu Shin, Kwanyong Park, Sanghyun Woo, In So Kweon
2021 arXiv   pre-print
In this work, we present a new video extension of this task, namely Unsupervised Domain Adaptation for Video Semantic Segmentation.  ...  Unsupervised Domain Adaptation for semantic segmentation has gained immense popularity since it can transfer knowledge from simulation to real (Sim2Real) by largely cutting out the laborious per pixel  ...  In second phase, VST, we design a aggregation based clip adaptive pseudo label generation strategy.  ... 
arXiv:2107.11052v2 fatcat:xoif3ki2ffaldapplslcoscpdu

A framework for self-supervised federated domain adaptation

Bin Wang, Gang Li, Chao Wu, WeiShan Zhang, Jiehan Zhou, Ye Wei
2022 EURASIP Journal on Wireless Communications and Networking  
Specifically, a multi-domain model generalization balance is proposed to aggregate the models from multiple source domains in each round of communication.  ...  A weighted strategy based on centroid similarity is also designed for SFDA. SFDA conducts self-supervised training on the target domain to tackle domain shift.  ...  Acknowledgements The authors thank the reviewers and editors for their efforts in valuable comments and suggestions. Authors' contributions All authors have contributed equally.  ... 
doi:10.1186/s13638-022-02104-8 fatcat:fflyzjadgffe3lnbrxnkob4wka

Attention-Aware Age-Agnostic Visual Place Recognition [article]

Ziqi Wang, Jiahui Li, Seyran Khademi, Jan van Gemert
2019 arXiv   pre-print
We propose an attention aggregation module that is robust to domain discrepancy between the train and the test data.  ...  Further, a multi-kernel maximum mean discrepancy (MK-MMD) domain adaptation loss is adopted to improve the cross-domain ranking performance.  ...  This work is also partly funded by the Netherlands Organization for Scientific Research (NWO) under the research program C2D-Horizontal Data Science for Evolving Content with project name DACCOMPLI and  ... 
arXiv:1909.05163v1 fatcat:ed75vdwsobbevk5kbqsge2pdkm

Domain and Content Adaptive Convolution for Domain Generalization in Medical Image Segmentation [article]

Shishuai Hu, Zehui Liao, Jianpeng Zhang, Yong Xia
2021 arXiv   pre-print
In this paper, we propose a multi-source domain generalization model, namely domain and content adaptive convolution (DCAC), for medical image segmentation.  ...  To address this issue, domain generalization methods have been proposed, which however usually use static convolutions and are less flexible.  ...  Based on the code, the trained domain-aware controller can generate the parameters for the domain-adaptive head.  ... 
arXiv:2109.05676v1 fatcat:tqufa3orfbdxtjeo7762ij6kyy
« Previous Showing results 1 — 15 out of 323,834 results