Filters








20,111 Hits in 3.7 sec

Online Unsupervised Learning of Visual Representations and Categories [article]

Mengye Ren, Tyler R. Scott, Michael L. Iuzzolino, Michael C. Mozer, Richard Zemel
2022 arXiv   pre-print
In this work, we propose an unsupervised model that simultaneously performs online visual representation learning and few-shot learning of new categories without relying on any class labels.  ...  Experiments show that our method can learn from an online stream of visual input data and its learned representations are significantly better at category recognition compared to state-of-the-art self-supervised  ...  Acknowledgments Resources used in preparing this research were provided, in part, by the Province of Ontario, the Government of Canada through CIFAR, and companies sponsoring the Vector Institute (www.  ... 
arXiv:2109.05675v4 fatcat:ryfqdeqgz5fzdbs4qxe46fytky

Collaborative Unsupervised Visual Representation Learning from Decentralized Data [article]

Weiming Zhuang, Xin Gan, Yonggang Wen, Shuai Zhang, Shuai Yi
2021 arXiv   pre-print
Unsupervised representation learning has achieved outstanding performances using centralized data available on the Internet.  ...  As such, a natural problem is how to leverage these data to learn visual representations for downstream tasks while preserving data privacy.  ...  Acknowledgements This study is supported by 1) the RIE2020 Industry Alignment Fund -Industry Collaboration Projects (IAF-ICP) Funding Initiative, as well as cash and in-kind contribution from the industry  ... 
arXiv:2108.06492v1 fatcat:hkunizhgongzzgg26xu55kgesi

Probabilistic Structural Latent Representation for Unsupervised Embedding

Mang Ye, Jianbing Shen
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
Extensive experiments demonstrate the superiority of PSLR over state-of-the-art unsupervised methods on both seen and unseen categories with cosine similarity.  ...  Unsupervised embedding learning aims at extracting low-dimensional visually meaningful representations from large-scale unlabeled images, which can then be directly used for similarity-based search.  ...  Why Latent Representation Learning? We visualize the similarity distribution of the training and testing set on the Car196 dataset.  ... 
doi:10.1109/cvpr42600.2020.00550 dblp:conf/cvpr/YeS20 fatcat:xkc65i5w7vd23llfknv25aft5i

Towards Using Unsupervised Learning for Comparing Traditional and Synchronous Online Learning in Assessing Students' Academic Performance

Mariana-Ioana Maier, Gabriela Czibula, Zsuzsanna-Edit Oneţ-Marian
2021 Mathematics  
Due to an accelerated expansion of online learning and digitalisation in education, there is a growing interest in understanding the impact of online learning on the academic performance of students.  ...  Experiments performed on real data sets collected in both online and traditional learning environments showed that autoencoders are able to detect hidden patterns in academic data sets unsupervised; these  ...  This work was supported by a grant of the Ministry of Research, Innovation and Digitization, CNCS/CCCDI -UEFISCDI, project number PN-III-P4-ID-PCE-2020-0800, within PNCDI III.  ... 
doi:10.3390/math9222870 fatcat:yvz7or5pbbhe3kdbofrf5uvrya

Discriminative stacked autoencoder for feature representation and classification

Yiping Gao, Xinyu Li, Liang Gao
2020 Science China Information Sciences  
Discriminative stacked autoencoder for feature representation and classification. Sci China Inf Sci, 2020, 63(2): 120111, https://doi.  ...  Figure 1 ( 1 Color online) (a) The diagram of the proposed DSA.  ...  (b) The data visualization of the learned feature; the top shows the feature learned from the original SAE, and the bottom is the feature learned from the proposed DSA. Y P, et al.  ... 
doi:10.1007/s11432-019-2722-3 fatcat:amdfgwce7zdwrii6cedh5eisty

Towards Unsupervised Sketch-based Image Retrieval [article]

Conghui Hu, Yongxin Yang, Yunpeng Li, Timothy M. Hospedales, Yi-Zhe Song
2022 arXiv   pre-print
Existing single-domain unsupervised representation learning methods perform poorly in this application, due to the unique cross-domain (sketch and photo) nature of the problem.  ...  We therefore introduce a novel framework that simultaneously performs sketch-photo domain alignment and semantic-aware representation learning.  ...  In more detail, our unsupervised learning strategy starts with state-of-the-art unsupervised representation learner SwAV [4] , which learns a semantically meaningful feature space by clustering and ensuring  ... 
arXiv:2105.08237v3 fatcat:hdtaq3zbivegnlzamohgpp5lv4

Novelty detection for unsupervised continual learning in image sequences

Ruiqi Dai, Mathieu Lefort, Frederic Armetta, Mathieu Guillermin, Stefan Duffner
2021 2021 IEEE 33rd International Conference on Tools with Artificial Intelligence (ICTAI)  
A further limitation that we could identify, in the case of online and unsupervised learning, is the tendency to oversegment categories into many additional clusters [4] , which makes the grouping of  ...  Comparatively, approaches based on deep learning are more suitable due to the powerful representation capacity for visual data and images.  ... 
doi:10.1109/ictai52525.2021.00080 fatcat:sy6atoamfvdl7par342ncm2qo4

Unsupervised Embedding Learning via Invariant and Spreading Instance Feature [article]

Mang Ye, Xu Zhang, Pong C. Yuen, Shih-Fu Chang
2019 arXiv   pre-print
It achieves significantly faster learning speed and higher accuracy than all existing methods. The proposed method performs well for both seen and unseen testing categories with cosine similarity.  ...  This paper studies the unsupervised embedding learning problem, which requires an effective similarity measurement between samples in low-dimensional embedding space.  ...  The main challenge of unsupervised embedding learning is to discover visual similarity or weak category information from unlabelled samples. Iscen et al.  ... 
arXiv:1904.03436v1 fatcat:re6dnvrm65ezrnly2mskdndcdu

Unsupervised Embedding Learning via Invariant and Spreading Instance Feature

Mang Ye, Xu Zhang, Pong C. Yuen, Shih-Fu Chang
2019 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
It achieves significantly faster learning speed and higher accuracy than all existing methods. The proposed method performs well for both seen and unseen testing categories with cosine similarity.  ...  This paper studies the unsupervised embedding learning problem, which requires an effective similarity measurement between samples in low-dimensional embedding space.  ...  The main challenge of unsupervised embedding learning is to discover visual similarity or weak category information from unlabelled samples. Iscen et al.  ... 
doi:10.1109/cvpr.2019.00637 dblp:conf/cvpr/YeZYC19 fatcat:nvd44iqqg5amhgwdn3fhjbaoly

Improving memory banks for unsupervised learning with large mini-batch, consistency and hard negative mining [article]

Adrian Bulat and Enrique Sánchez-Lozano and Georgios Tzimiropoulos
2021 arXiv   pre-print
An important component of unsupervised learning by instance-based discrimination is a memory bank for storing a feature representation for each training sample in the dataset.  ...  Overall, our approach greatly improves the vanilla memory-bank based instance discrimination and outperforms all existing methods for both seen and unseen testing categories with cosine similarity.  ...  Recently, there is a surge of interest in learning features in an unsupervised manner.  ... 
arXiv:2102.04442v1 fatcat:yoobifaxf5cjrlcesia3vbalza

Controlled-rearing studies of newborn chicks and deep neural networks [article]

Donsuk Lee, Pranav Gujarathi, Justin N. Wood
2021 arXiv   pre-print
However, there is a widely accepted critique of CNN models: unlike newborn animals, which learn rapidly and efficiently, CNNs are thought to be "data hungry," requiring massive amounts of training data  ...  CNNs are also the leading quantitative models in terms of predicting neural and behavioral responses in visual recognition tasks.  ...  As a starting point, we focused on three unsupervised learning algorithms: convolutional autoencoders, simple contrastive learning of representations (SimCLR) (Chen et al., 2020) , and "Bring Your Own  ... 
arXiv:2112.06106v1 fatcat:tpqzdtf4rfhrph3365inaboe2u

Online unsupervised cumulative learning for life-long robot operation

Y. Gatsoulis, C. Burbridge, T. M. Mcginnity
2011 2011 IEEE International Conference on Robotics and Biomimetics  
This paper describes a novel method for cumulative unsupervised learning of objects by visual inspection, using an online and expanding when required Bag-of-Words.  ...  Bag-of-Words is a generic and compact representation of visual perceptions which has commonly and successfully been used in object recognition problems.  ...  There is an extensive body of existing research into object recognition, and one popular approach is using the compact and generic Bag-of-Words representation of visual perceptions (BoW) [8] , [11] .  ... 
doi:10.1109/robio.2011.6181678 dblp:conf/robio/GatsoulisBM11 fatcat:uysgllvvezeabgscpfc5g5vfve

Modeling unsupervised perceptual category learning

Brenden M. Lake, Gautam K. Vallabha, James L. McClelland
2008 2008 7th IEEE International Conference on Development and Learning  
unsupervised learning of category structure in simple visual stimuli.  ...  Index Terms-human learning, mixture of Gaussians, online learning, unsupervised learning.  ...  Rosenthal for her comments and for providing data used in Fig. 5 . They also thank three anonymous reviewers for their helpful suggestions.  ... 
doi:10.1109/devlrn.2008.4640800 fatcat:667lnqnc4vaf3axdgjkpg42quq

Modeling Unsupervised Perceptual Category Learning

B.M. Lake, G.K. Vallabha, J.L. McClelland
2009 IEEE Transactions on Autonomous Mental Development  
unsupervised learning of category structure in simple visual stimuli.  ...  Index Terms-human learning, mixture of Gaussians, online learning, unsupervised learning.  ...  Rosenthal for her comments and for providing data used in Fig. 5 . They also thank three anonymous reviewers for their helpful suggestions.  ... 
doi:10.1109/tamd.2009.2021703 fatcat:c2p7nupuz5euxbd7bdeb6qywqy

Semi-Supervised Learning for Mars Imagery Classification and Segmentation [article]

Wenjing Wang, Lilang Lin, Zejia Fan, Jiaying Liu
2022 arXiv   pre-print
Contrastive learning is a powerful representation learning technique.  ...  For segmentation, we extend supervised inter-class contrastive learning into an element-wise mode and use online pseudo labels for supervision on unlabeled areas.  ...  Unlike existing methods, we solve the problem through representation learning. With a robust visual representation, the train-test gap and low data quality can be resolved simultaneously.  ... 
arXiv:2206.02180v1 fatcat:e6supf3fwrafpk5f75ckjarl4u
« Previous Showing results 1 — 15 out of 20,111 results