Filters








1,020 Hits in 7.2 sec

CrossRectify: Leveraging Disagreement for Semi-supervised Object Detection [article]

Chengcheng Ma, Xingjia Pan, Qixiang Ye, Fan Tang, Weiming Dong, Changsheng Xu
2022 arXiv   pre-print
Semi-supervised object detection has recently achieved substantial progress.  ...  These self-errors can hurt performance even worse than random-errors, and can be neither discerned nor rectified during the self-labeling process.  ...  Acknowledgement This work was supported by the National Natural Science Foundation of China 61832016, U20B2070, 6210070958, 62102162.  ... 
arXiv:2201.10734v2 fatcat:nai3wods5fhc5jcpryd4gmbmo4

Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identification [article]

Yixiao Ge, Dapeng Chen, Hongsheng Li
2020 arXiv   pre-print
In order to mitigate the effects of noisy pseudo labels, we propose to softly refine the pseudo labels in the target domain by proposing an unsupervised framework, Mutual Mean-Teaching (MMT), to learn  ...  In addition, the common practice is to adopt both the classification loss and the triplet loss jointly for achieving optimal performances in person re-ID models.  ...  In order to avoid error amplification, we propose to use the temporally average model of each network to generate reliable soft pseudo labels for supervising the other network.  ... 
arXiv:2001.01526v2 fatcat:7nqggtsmwvhp3kkz5ofm76xuli

Modeling Techniques for Machine Learning Fairness: A Survey [article]

Mingyang Wan, Daochen Zha, Ninghao Liu, Na Zou
2022 arXiv   pre-print
In recent years, various techniques have been developed to mitigate the unfairness for machine learning models.  ...  Based on where the fairness is achieved in the model, we categorize them into explicit and implicit methods, where the former directly incorporates fairness metrics in training objectives, and the latter  ...  For instance, semi-supervised and unsupervised settings demand tailored designs to leverage unlabeled data to mitigate unfairness [72, 85] .  ... 
arXiv:2111.03015v2 fatcat:didcuo2yabbcrb2fuhveqgng3y

Dual-Correction Adaptation Network for Noisy Knowledge Transfer [article]

Yunyun Wang and Weiwen Zheng and Songcan Chen
2022 arXiv   pre-print
Secondly, source domain usually contains innate noises, which will inevitably aggravate the target noises, leading to noise amplification across domains.  ...  To our best knowledge, this is the first naive attempt of dual-directional adaptation for noisy UDA, and naturally applicable to noise-free UDA.  ...  ACKNOWLEDGMENTS This work is supported in part by the NSFC under Grant No.s 62076124 and 62176118.  ... 
arXiv:2207.04423v1 fatcat:n7dbfpgkt5cihpxsnjzecauure

An Overview on Application of Machine Learning Techniques in Optical Networks

Francesco Musumeci, Cristina Rottondi, Avishek Nag, Irene Macaluso, Darko Zibar, Marco Ruffini, Massimo Tornatore
2018 IEEE Communications Surveys and Tutorials  
We classify and survey relevant literature dealing with the topic, and we also provide an introductory tutorial on ML for researchers and practitioners interested in this field.  ...  The adoption of ML techniques in the field of optical communication networks is motivated by the unprecedented growth of network complexity faced by optical networks in the last few years.  ...  Supervised learning Supervised learning is used in a variety of applications, such as speech recognition, spam detection and object recognition.  ... 
doi:10.1109/comst.2018.2880039 fatcat:ql662vgph5hjdejxtl5yvdysom

An Overview on Application of Machine Learning Techniques in Optical Networks [article]

Francesco Musumeci, Cristina Rottondi, Avishek Nag, Irene Macaluso, Darko Zibar, Marco Ruffini, Massimo Tornatore
2018 arXiv   pre-print
We classify and survey relevant literature dealing with the topic, and we also provide an introductory tutorial on ML for researchers and practitioners interested in this field.  ...  The adoption of ML techniques in the field of optical communication networks is motivated by the unprecedented growth of network complexity faced by optical networks in the last few years.  ...  Supervised learning Supervised learning is used in a variety of applications, such as speech recognition, spam detection and object recognition.  ... 
arXiv:1803.07976v4 fatcat:rhzumocnrzfxpismbpkjejytfm

A Review of Machine Learning-based Failure Management in Optical Networks [article]

Danshi Wang, Chunyu Zhang, Wenbin Chen, Hui Yang, Min Zhang, Alan Pak Tao Lau
2022 arXiv   pre-print
Finally, the future directions on ML for failure management are discussed from the perspective of data, model, task, and emerging techniques.  ...  An overview of the applications of ML in failure management is provided in terms of alarm analysis, failure prediction, failure detection, failure localization, and failure identification.  ...  The RA is always combined with EDFA for hybrid amplification.  ... 
arXiv:2208.10677v1 fatcat:qrsqkb7kpza6jfakb6t5kvci7q

Deep Learning-Based Object Detection Techniques for Remote Sensing Images: A Survey

Zheng Li, Yongcheng Wang, Ning Zhang, Yuxi Zhang, Zhikang Zhao, Dongdong Xu, Guangli Ben, Yunxiao Gao
2022 Remote Sensing  
Object detection in remote sensing images (RSIs) requires the locating and classifying of objects of interest, which is a hot topic in RSI analysis research.  ...  However, although some scholars have authored reviews on DL-based object detection systems, the leading DL-based object detection improvement strategies have never been summarized in detail.  ...  Semi-supervised learning can mitigate the model's requirement for labels and thus reduce the number of labeled samples, which can effectively solve the problem of labeling complex samples of RSIs.  ... 
doi:10.3390/rs14102385 fatcat:sgoqy33cdbe2xopqqfsa2hyowq

A Survey on Bias in Visual Datasets [article]

Simone Fabbrizzi, Symeon Papadopoulos, Eirini Ntoutsi, Ioannis Kompatsiaris
2022 arXiv   pre-print
Hence, this work aims to: i) describe the biases that might manifest in visual datasets; ii) review the literature on methods for bias discovery and quantification in visual datasets; iii) discuss existing  ...  A key conclusion of our study is that the problem of bias discovery and quantification in visual datasets is still open, and there is room for improvement in terms of both methods and the range of biases  ...  This work has received funding from the European Union's Horizon 2020 research and  ... 
arXiv:2107.07919v2 fatcat:sppbqoftqjhlnpc35zo4xwnwdy

Deep Probabilistic Logic: A Unifying Framework for Indirect Supervision [article]

Hai Wang, Hoifung Poon
2018 arXiv   pre-print
end task and refining uncertain formula weights for indirect supervision, using variational EM.  ...  In this paper, we propose deep probabilistic logic (DPL) as a general framework for indirect supervision, by composing probabilistic logic with deep learning.  ...  Acknowledgements We thank David McAllester, Chris Quirk, and Scott Yih for useful discussions, and the three anonymous reviewers for helpful comments.  ... 
arXiv:1808.08485v1 fatcat:stmfezgu3zholekpfpgcomubie

Deep Probabilistic Logic: A Unifying Framework for Indirect Supervision

Hai Wang, Hoifung Poon
2018 Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing  
end task and refining uncertain formula weights for indirect supervision, using variational EM.  ...  In this paper, we propose deep probabilistic logic (DPL) as a general framework for indirect supervision, by composing probabilistic logic with deep learning.  ...  Acknowledgements We thank David McAllester, Chris Quirk, and Scott Yih for useful discussions, and the three anonymous reviewers for helpful comments.  ... 
doi:10.18653/v1/d18-1215 dblp:conf/emnlp/WangP18 fatcat:3xbhlmii55ac5p2u3vsnvpylhi

Deep Learning in Spatial Transcriptomics: A Survey of Deep Learning Methods for Spatially-Resolved Transcriptomics [article]

A. Ali Heydari, Suzanne S. Sindi
2022 bioRxiv   pre-print
for accurate and robust analysis.  ...  In this review, we provide an overview of existing state-of-the-art tools for analyzing spatially-resolved transcriptomics, while delving deeper into the DL-based approaches.  ...  In addition to supervised and unsupervised algorithms, there are also semi-supervised learning, where a model uses a mix of both supervised and unsupervised tasks, and self-supervised, where the computer  ... 
doi:10.1101/2022.02.28.482392 fatcat:mw5vz673urbv5egvwbpvflitoi

Euclid preparation. IV. Impact of undetected galaxies on weak-lensing shear measurements

N. Martinet, T. Schrabback, H. Hoekstra, M. Tewes, R. Herbonnet, P. Schneider, B. Hernandez-Martin, A.N. Taylor, J. Brinchmann, C.S. Carvalho, M. Castellano, G. Congedo (+4 others)
2019 Astronomy and Astrophysics  
In this paper we study the importance of faint galaxies below the observational detection limit of a survey.  ...  shear bias calibration of a few times 10 −4 , in line with the 2 × 10 −3 total accuracy budget required by the scientific objectives of the Euclid survey.  ...  We are grateful to the GalSim and HST-UDF team for the public release of their software and data, respectively.  ... 
doi:10.1051/0004-6361/201935187 fatcat:xqvy5rekjngxrkslbp3odunbpa

Precoding for High Throughput Satellite Communication Systems: A Survey [article]

Malek Khammassi, Abla Kammoun, Mohamed-Slim Alouini
2022 arXiv   pre-print
From a problem formulation point of view, precoding techniques are classified according to the precoding objective, group, and level.  ...  With the expanding demand for high data rates and extensive coverage, high throughput satellite (HTS) communication systems are emerging as a key technology for future communication generations.  ...  These conditions ensure mutuality of constructive interference.  ... 
arXiv:2208.08542v1 fatcat:6nzq52blobawpc2ouixsd37pte

Comprehensive Review of Artificial Intelligence and Statistical Approaches in Distributed Denial of Service Attack and Defense Methods

Bashar Ahmed Khalaf, Salama A. Mostafa, Aida Mustapha, Mazin Abed Mohammed, Wafaa Mustafa Abduallah
2019 IEEE Access  
Incidents of serious damage due to DDoS attacks have been increasing, thereby leading to an urgent need for new attack identification, mitigation, and prevention mechanisms.  ...  However, the most appropriate and effective defense features, mechanisms, techniques, and methods for handling such attacks remain to be an open question.  ...  Adjacent nodes and attack information supervise the node activity sent to the expert unit for integrated assessment.  ... 
doi:10.1109/access.2019.2908998 fatcat:wasnqcj4bnc23pk4wzsl7n7rze
« Previous Showing results 1 — 15 out of 1,020 results