Filters








23,037 Hits in 1.9 sec

Unsupervised Learning of Dense Visual Representations [article]

Pedro O. Pinheiro, Amjad Almahairi, Ryan Y. Benmalek, Florian Golemo, Aaron Courville
2020 arXiv   pre-print
In this paper, we propose View-Agnostic Dense Representation (VADeR) for unsupervised learning of dense representations.  ...  Contrastive self-supervised learning has emerged as a promising approach to unsupervised visual representation learning.  ...  In this paper, we propose a method for unsupervised learning of dense representations.  ... 
arXiv:2011.05499v2 fatcat:eaxgc3f3trdwxjvttumon2jfei

Unsupervised representation learning using convolutional and stacked auto-encoders: a domain and cross-domain feature space analysis [article]

Gabriel B. Cavallari, Leonardo Sampaio Ferraz Ribeiro, Moacir Antonelli Ponti
2018 arXiv   pre-print
Our findings can be used as guidelines for the design of unsupervised representation learning methods within and across domains.  ...  A feature learning task involves training models that are capable of inferring good representations (transformations of the original space) from input data alone.  ...  Considering we learn the embedding using a different dataset, a 0.94% accuracy indicates the potential of unsupervised representation learning.  ... 
arXiv:1811.00473v1 fatcat:m7z6qivmzrd67ngfwztb2wkqyi

Dense Siamese Network [article]

Wenwei Zhang, Jiangmiao Pang, Kai Chen, Chen Change Loy
2022 arXiv   pre-print
It learns visual representations by maximizing the similarity between two views of one image with two types of consistency, i.e., pixel consistency and region consistency.  ...  This paper presents Dense Siamese Network (DenseSiam), a simple unsupervised learning framework for dense prediction tasks.  ...  The predictions of one view are learned to match those in another view. Table 1 . 1 Comparison of unsupervised dense representation learning methods.  ... 
arXiv:2203.11075v1 fatcat:2kde5tkpjbcezeiflsdpjau3te

Propagate Yourself: Exploring Pixel-Level Consistency for Unsupervised Visual Representation Learning [article]

Zhenda Xie and Yutong Lin and Zheng Zhang and Yue Cao and Stephen Lin and Han Hu
2021 arXiv   pre-print
Contrastive learning methods for unsupervised visual representation learning have reached remarkable levels of transfer performance.  ...  These results demonstrate the strong potential of defining pretext tasks at the pixel level, and suggest a new path forward in unsupervised visual representation learning. Code is available at .  ...  These results demonstrate the strong potential of defining pretext tasks at the pixel level, and suggest a new path forward in unsupervised visual representation learning.  ... 
arXiv:2011.10043v2 fatcat:o5bt6vxhtzhsvaqvaxtmopgkda

Mapping in a cycle: Sinkhorn regularized unsupervised learning for point cloud shapes [article]

Lei Yang, Wenxi Liu, Zhiming Cui, Nenglun Chen, Wenping Wang
2020 arXiv   pre-print
We propose an unsupervised learning framework with the pretext task of finding dense correspondences between point cloud shapes from the same category based on the cycle-consistency formulation.  ...  We also show that the learned pointwise features can be leveraged by supervised methods to improve the part segmentation performance with either the full training dataset or just a small portion of it.  ...  Table category Fig. 2 . 2 Visualization of the learned pointwise representations on rotated shapes from different categories, i.e., Airplane, Table, Guitar, Skateboard, and Chair.  ... 
arXiv:2007.09594v1 fatcat:5apqecvvy5fjdch3w3qitynori

Decomposition, discovery and detection of visual categories using topic models

Mario Fritz, Bernt Schiele
2008 2008 IEEE Conference on Computer Vision and Pattern Recognition  
The approach is capable of learning a compact and low dimensional representation for multiple visual categories from multiple view points without labeling of the training instances.  ...  Experiments on three databases show that the approach improves the state-of-the-art in unsupervised learning as well as supervised detection.  ...  Decomposition of Visual Categories In this section we describe our approach to decomposition of multiple visual categories by combining dense gradient representations and topic models.  ... 
doi:10.1109/cvpr.2008.4587803 dblp:conf/cvpr/FritzS08 fatcat:z2w22jyxezhwxd4fwmozi2u2xy

Supervised Understanding of Word Embeddings [article]

Halid Ziya Yerebakan, Parmeet Bhatia, Yoshihisa Shinagawa
2020 arXiv   pre-print
The embeddings are continuous and distributed representations of the words that preserve their similarities in compact Euclidean spaces.  ...  We further demonstrate the usefulness of supervised dimensions in revealing the polysemous nature of a keyword of interest by projecting it's embedding using learned classifiers in different sub-spaces  ...  These dense representations are more compact than prior bag of word methods (Joachims 1998), (McCallum and Nigam 1998) while providing transfer learning capability, thanks to learned similarities across  ... 
arXiv:2006.13299v1 fatcat:2qvoug75wvelpccifqslv6l7o4

Early Visual Concept Learning with Unsupervised Deep Learning [article]

Irina Higgins, Loic Matthey, Xavier Glorot, Arka Pal, Benigno Uria, Charles Blundell, Shakir Mohamed, Alexander Lerchner
2016 arXiv   pre-print
Addressing this problem, we propose an unsupervised approach for learning disentangled representations of the underlying factors of variation.  ...  We draw inspiration from neuroscience, and show how this can be achieved in an unsupervised generative model by applying the same learning pressures as have been suggested to act in the ventral visual  ...  Conclusion In this paper we have shown that deep unsupervised generative models are capable of learning disentangled representations of the visual data generative factors if put under similar learning  ... 
arXiv:1606.05579v3 fatcat:org54dkpgje2xdtss4bt5qtekm

Deforming Autoencoders: Unsupervised Disentangling of Shape and Appearance [article]

Zhixin Shu, Mihir Sahasrabudhe, Alp Guler, Dimitris Samaras, Nikos Paragios, Iasonas Kokkinos
2018 arXiv   pre-print
We introduce novel techniques that allow this approach to be deployed in the setting of autoencoders and show that this method can be used for unsupervised group-wise image alignment.  ...  A more powerful form of unsupervised disentangling becomes possible in template coordinates, allowing us to successfully decompose face images into shading and albedo, and further manipulate face images  ...  With latent representations learned in an unsupervised manner for shading, albedo, and defomation, the DAE allows us to simulate smooth transitions of the lighting direction.  ... 
arXiv:1806.06503v1 fatcat:2y3w7ofn6fhzrkac27gsabrg74

ViCE: Visual Concept Embedding Discovery and Superpixelization [article]

Robin Karlsson, Tomoki Hayashi, Keisuke Fujii, Alexander Carballo, Kento Ohtani, Kazuya Takeda
2022 arXiv   pre-print
We introduce superpixelization as a means to decompose images into a small set of visually coherent regions, allowing efficient learning of dense semantics by swapped prediction.  ...  However, these methods are classification-based and thus ineffective for learning dense feature maps required for unsupervised semantic segmentation.  ...  Our comparative baseline for dense representation learning is the recent SOTA unsupervised semantic segmentation model PiCIE [19] based on DeepCluster [10] .  ... 
arXiv:2111.12460v2 fatcat:vmr6f7pizjdvrluyesxqpklio4

Unsupervised learning of object frames by dense equivariant image labelling [article]

James Thewlis and Hakan Bilen and Andrea Vedaldi
2017 arXiv   pre-print
One of the key challenges of visual perception is to extract abstract models of 3D objects and object categories from visual measurements, which are affected by complex nuisance factors such as viewpoint  ...  Starting from the recent idea of viewpoint factorization, we propose a new approach that, given a large number of images of an object and no other supervision, can extract a dense object-centric coordinate  ...  Acknowledgments: This work acknowledges the support of the AIMS CDT (EPSRC EP/L015897/1) and ERC 677195-IDIU. Clipart: FreePik.  ... 
arXiv:1706.02932v2 fatcat:ae3jowrtibbj5o7a4brrd3ssfy

Towards bio-inspired unsupervised representation learning for indoor aerial navigation [article]

Ni Wang, Ozan Catal, Tim Verbelen, Matthias Hartmann, Bart Dhoedt
2021 arXiv   pre-print
We propose an unsupervised representation learning method that yields low-dimensional latent state descriptors, that mitigates the sensitivity to perceptual aliasing, and works on power-efficient, embedded  ...  Drones can perceive the environment from a richer set of viewpoints, while having more stringent compute and energy constraints than other autonomous platforms.  ...  ACKNOWLEDGMENT O.Ç . is funded by a Ph.D. grant of the Flanders Research Foundation (FWO).  ... 
arXiv:2106.09326v1 fatcat:m2fxiwavivgxfonexgy5xa3cee

An Ameliorated method for Fraud Detection using Complex Generative Model: Variational Autoencoder

2019 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
The current Supervised and Unsupervised Machine Learning Algorithm approaches to the discovery of fraud are their inability to learn and explore all possible information representation.  ...  The VAE-based fraud detection model is capable of learning latent variable probabilistic models by optimizing the average value of the information observed.  ...  EXISTING APPROACH The two main types of learning are supervised approach and unsupervised approach.  ... 
doi:10.35940/ijitee.b1005.1292s19 fatcat:seg34ssutbbqzktx6o52hcmfni

Unsupervised Representation Learning of Structured Radio Communication Signals [article]

Timothy J. O'Shea, Johnathan Corgan, T. Charles Clancy
2016 arXiv   pre-print
We explore unsupervised representation learning of radio communication signals in raw sampled time series representation.  ...  We demonstrate that we can learn modulation basis functions using convolutional autoencoders and visually recognize their relationship to the analytic bases used in digital communications.  ...  The views, opinions, and/or findings expressed are those of the author and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government.  ... 
arXiv:1604.07078v1 fatcat:dyybkn4ptzdmrcgnqgv2nt22my

Adversarial Framework for Unsupervised Learning of Motion Dynamics in Videos [article]

C. Spampinato, S. Palazzo, P. D'Oro, D. Giordano, M. Shah
2019 arXiv   pre-print
Our approach synthesizes videos by 1) factorizing the process into the generation of static visual content and motion, 2) learning a suitable representation of a motion latent space in order to enforce  ...  Unsupervised learning can instead leverage the vast amount of videos available on the web and it is a promising solution for overcoming the existing limitations.  ...  The learned representations are then employed to train a method for unsupervised video object segmentation; thus, in this section, we will focus on this last class of approaches.  ... 
arXiv:1803.09092v2 fatcat:tconl7knq5af3nqlbxthvx7br4
« Previous Showing results 1 — 15 out of 23,037 results