Filters








26,992 Hits in 4.5 sec

Curriculum Manager for Source Selection in Multi-Source Domain Adaptation [article]

Luyu Yang, Yogesh Balaji, Ser-Nam Lim, Abhinav Shrivastava
2020 arXiv   pre-print
In this paper, we proposed an adversarial agent that learns a dynamic curriculum for source samples, called Curriculum Manager for Source Selection (CMSS).  ...  The performance of Multi-Source Unsupervised Domain Adaptation depends significantly on the effectiveness of transfer from labeled source domain samples.  ...  Conclusion In this paper, we proposed Curriculum Manager for Source Selection (CMSS) that learns a curriculum for Multi-Source Unsupervised Domain Adaptation.  ... 
arXiv:2007.01261v1 fatcat:nrgf2sbqa5a7pcrqfkgcr6qpaq

Curriculum CycleGAN for Textual Sentiment Domain Adaptation with Multiple Sources [article]

Sicheng Zhao, Yang Xiao, Jiang Guo, Xiangyu Yue, Jufeng Yang, Ravi Krishna, Pengfei Xu, Kurt Keutzer
2021 arXiv   pre-print
Existing multi-source domain adaptation (MDA) methods either fail to extract some discriminative features in the target domain that are related to sentiment, neglect the correlations of different sources  ...  with curriculum instance-level adaptation which bridges the gap across source and target domains, and (3) task classifier trained on the intermediate domain for final sentiment classification.  ...  , U1933114), Natural Science Foundation of Tianjin, China (Nos. 20JCJQJC00020, 18JCYBJC15400, 18ZXZNGX00110), and the Fundamental Research Funds for the Central Universities.  ... 
arXiv:2011.08678v2 fatcat:7gyajhdcynblvjgtcjcjrpgy4i

Class-Conditional Domain Adaptation on Semantic Segmentation [article]

Yue Wang, Yuke Li, James H. Elder, Runmin Wu, Huchuan Lu
2019 arXiv   pre-print
Unsupervised domain adaptation can potentially address these problems, allowing systems trained on labelled datasets from one or more source domains (including less expensive synthetic domains) to be adapted  ...  We address this problem by introducing a Class-Conditional Domain Adaptation method (CCDA). It includes a class-conditional multi-scale discriminator and the class-conditional loss.  ...  Acknowledgements: We would like to thank the York University Vision: Science to Applications (VISTA) program for its support.  ... 
arXiv:1911.11981v2 fatcat:nx5orxg7sffybk3r2nex2vztay

Unsupervised Domain Adaptation for Semantic Image Segmentation: a Comprehensive Survey [article]

Gabriela Csurka, Riccardo Volpi, Boris Chidlovskii
2021 arXiv   pre-print
, domain generalization, test-time adaptation or source-free domain adaptation; we conclude this survey by describing datasets and benchmarks most widely used in semantic segmentation research.  ...  We present the most important semantic segmentation methods; we provide a comprehensive survey on domain adaptation techniques for semantic segmentation; we unveil newer trends such as multi-domain learning  ...  Curriculum Manager for Source Selection in domain adaptation method for semantic segmentation. In Multi-Source Domain Adaptation.  ... 
arXiv:2112.03241v1 fatcat:uzlehddvuvfwzf4dfbjimja45e

Lifelong Personalization via Gaussian Process Modeling for Long-Term HRI

Samuel Spaulding, Jocelyn Shen, Hae Won Park, Cynthia Breazeal
2021 Frontiers in Robotics and AI  
Inspired by the way in which agents use active learning to select new training data based on domain context, we augment a Gaussian Process-based multitask personalization model with a mechanism to actively  ...  Across a wide variety of domains, artificial agents that can adapt and personalize to users have potential to improve and transform how social services are provided.  ...  "short, well-defined lessons delivered with limited adaptation to individual learners or flexibility in curriculum."  ... 
doi:10.3389/frobt.2021.683066 pmid:34164437 pmcid:PMC8215502 fatcat:wetgbxbkbnconkcfkwnhy3jcxi

Curriculum-style Local-to-global Adaptation for Cross-domain Remote Sensing Image Segmentation [article]

Bo Zhang, Tao Chen, Bin Wang
2022 arXiv   pre-print
The proposed curriculum-style adaptation performs the adaptation process in an easy-to-hard way according to the adaptation difficulties that can be obtained using an entropy-based score for each patch  ...  To address these challenges, we propose a curriculum-style local-to-global cross-domain adaptation framework for the segmentation of VHR RSIs.  ...  First, the given source-domain model is adapted from the source domain to the selected easy-to-adapt patches of the target domain.  ... 
arXiv:2203.01539v1 fatcat:6ostmgkl5ndhzkazasair2pvg4

Page 34 of Communication Abstracts Vol. 22, Issue 1 [page]

1999 Communication Abstracts  
Emergency management of chemical spills was selected to exemplify the rule-based decision task. An expert system in this domain was developed to serve as the training tool.  ...  To con- tribute to the human success in playing such a role, this study examines the effec- tiveness of using expert systems to train for the time-constrained decision domain.  ... 

Medical Image Classification Based on Curriculam Learning

2019 International journal of recent technology and engineering  
This paper is an attempt to apply SSL through Multi-Modal Curriculum Learning (MMCL) strategy over medical images. Through this, medical images can be categorized into normal and abnormal images.  ...  Experimental results demonstrate good accuracy for classification.  ...  For this purpose binary selection matrix is used and it ensures each image is selected only once. To improve the performance, single modal curriculum is expanded to multi-modal curriculum.  ... 
doi:10.35940/ijrte.b1001.0782s219 fatcat:l725rxtlqjezfg6lfp6bg2646e

Towards a Capabilities Taxonomy for Prognostics and Health Management

Jeff Bird, Nancy Madge, Karl Reichard
2020 International Journal of Prognostics and Health Management  
This communication proposes the development by the PHM Society of a classification or taxonomy for the skills needed for the prognostics and health management (PHM) field.  ...  Preliminary results of the development of Analytics, Test and Experiment Design and Cost Benefit Studies sub-domains within the PHM field are reported based on workshops at the PHM 2012 and 2013 Annual  ...  She specializes in curriculum development for adult learning and training in industrial and professional settings.  ... 
doi:10.36001/ijphm.2014.v5i1.2201 fatcat:3r4md4pi5je5la3qqovlzn76va

Unsupervised Domain Adaptation in Semantic Segmentation: a Review [article]

Marco Toldo, Andrea Maracani, Umberto Michieli, Pietro Zanuttigh
2020 arXiv   pre-print
The aim of this paper is to give an overview of the recent advancements in the Unsupervised Domain Adaptation (UDA) of deep networks for semantic segmentation.  ...  analysis of the classifier discrepancies, self-teaching, entropy minimization, curriculum learning and multi-task learning.  ...  Multi-Tasking Some works exploit additional types of information available in the source domain dataset, e.g., depth maps, to improve the performance in the target domain.  ... 
arXiv:2005.10876v1 fatcat:7t5v6qibxnfcxhwtohqqunhd2u

Learning and Knowledge Transfer with Memory Networks for Machine Comprehension

Mohit Yadav, Lovekesh Vig, Gautam Shroff
2017 Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers  
Motivated by these practical issues, we propose a novel curriculum inspired training procedure for Memory Networks to improve the performance for machine comprehension with relatively small volumes of  ...  Additionally, we explore various training regimes for Memory Networks to allow knowledge transfer from a closely related domain having larger volumes of labelled data.  ...  in the source domain N SD .  ... 
doi:10.18653/v1/e17-1080 dblp:conf/eacl/YadavVS17 fatcat:q7r7uklx6zgopmjudrtgu3wfn4

Edison Data Science Framework: Part 3. Data Science Model Curriculum (Mc-Ds) Release 1

Yuri Demchenko, Adam Belloum, Tomasz Wiktorski
2016 Zenodo  
When coupled with individual or group competence benchmarking, MC-DS can also be used for building individual training curricula and professional (self/up) skilling for effective career management.  ...  Further MC-DS refinement will be done based on consultation with the universities community and experts both in Data Science and scientific or industry domains.  ...  or evaluating the existing curriculum for compliance to the selected Data Science professional profiles.  ... 
doi:10.5281/zenodo.167592 fatcat:axjcv7zj6jffxobrj2o52624zi

Unsupervised Domain Adaptation in Semantic Segmentation: A Review

Marco Toldo, Andrea Maracani, Umberto Michieli, Pietro Zanuttigh
2020 Technologies  
The aim of this paper is to give an overview of the recent advancements in the Unsupervised Domain Adaptation (UDA) of deep networks for semantic segmentation.  ...  analysis of the classifier discrepancies, self-teaching, entropy minimization, curriculum learning and multi-task learning.  ...  Multi-Tasking Some works exploit additional types of information available in the source domain dataset, for example, depth maps, to improve the performance in the target domain.  ... 
doi:10.3390/technologies8020035 fatcat:qzgjjiw5p5bldk76mh3s3pwlfq

Zero-Shot Deep Domain Adaptation [article]

Kuan-Chuan Peng, Ziyan Wu, Jan Ernst
2018 arXiv   pre-print
Domain adaptation is an important tool to transfer knowledge about a task (e.g. classification) learned in a source domain to a second, or target domain.  ...  Therefore, the source-domain task of interest solution (e.g. a classifier for classification tasks) which is jointly trained with the source-domain representation can be applicable to both the source and  ...  [17] in a typical domain adaptation (DA) task, where source-domain training data, target-domain training data, and a task of interest (TOI) are given.  ... 
arXiv:1707.01922v5 fatcat:4zl3hiiegjbtxjh523de4xn6dq

KD3A: Unsupervised Multi-Source Decentralized Domain Adaptation via Knowledge Distillation [article]

Hao-Zhe Feng, Zhaoyang You, Minghao Chen, Tianye Zhang, Minfeng Zhu, Fei Wu, Chao Wu, Wei Chen
2021 arXiv   pre-print
Conventional unsupervised multi-source domain adaptation (UMDA) methods assume all source domains can be accessed directly.  ...  In this study, we propose a privacy-preserving UMDA paradigm named Knowledge Distillation based Decentralized Domain Adaptation (KD3A), which performs domain adaptation through the knowledge distillation  ...  selection based methods, i.e. the curriculum manager (CMSS) (Yang et al., 2020). (4) Decentralized UMDA, i.e.  ... 
arXiv:2011.09757v7 fatcat:h3elgrahsvdf7mt37n3miu2vw4
« Previous Showing results 1 — 15 out of 26,992 results