548 Hits in 4.3 sec

Semi-supervised Learning with GANs: Manifold Invariance with Improved Inference [article]

Abhishek Kumar, Prasanna Sattigeri, P. Thomas Fletcher
2017 arXiv   pre-print
Semi-supervised learning methods using Generative Adversarial Networks (GANs) have shown promising empirical success recently.  ...  Motivated by the ability of the GANs generator to capture the data manifold well, we propose to estimate the tangent space to the data manifold using GANs and employ it to inject invariances into the classifier  ...  In this work we make following contributions: • We propose to use the tangents from the generator's mapping to automatically infer the desired invariances and further improve on semi-supervised learning  ... 
arXiv:1705.08850v2 fatcat:bhvyhsbujfenvprmi6ttvpy4vu

Semi-Supervised Learning with GANs: Revisiting Manifold Regularization [article]

Bruno Lecouat, Chuan-Sheng Foo, Houssam Zenati, Vijay R. Chandrasekhar
2018 arXiv   pre-print
When incorporated into the feature-matching GAN of Improved GAN, we achieve state-of-the-art results for GAN-based semi-supervised learning on the CIFAR-10 dataset, with a method that is significantly  ...  We leverage this property to perform manifold regularization by approximating the Laplacian norm using a Monte Carlo approximation that is easily computed with the GAN.  ...  When applied to the feature-matching GAN, we achieve state-of-the-art performance amongst GAN-based methods for semi-supervised learning.  ... 
arXiv:1805.08957v1 fatcat:47yci44ahnawxiszwwbhpz2wtu

Consistency Regularization with Generative Adversarial Networks for Semi-Supervised Learning [article]

Zexi Chen, Bharathkumar Ramachandra, Ranga Raju Vatsavai
2020 arXiv   pre-print
Generative Adversarial Networks (GANs) based semi-supervised learning (SSL) approaches are shown to improve classification performance by utilizing a large number of unlabeled samples in conjunction with  ...  Our experiments show that this new composite consistency regularization based semi-GAN significantly improves its performance and achieves new state-of-the-art performance among GAN-based SSL approaches  ...  discriminator of semi-GAN, encouraging it to make consistent predictions for data under perturbations, thus leading to improved semi-supervised classification.  ... 
arXiv:2007.03844v2 fatcat:2iu34ftqtbgzvlgflzjx6w4lqy

Local Clustering with Mean Teacher for Semi-supervised Learning [article]

Zexi Chen, Benjamin Dutton, Bharathkumar Ramachandra, Tianfu Wu, Ranga Raju Vatsavai
2020 arXiv   pre-print
semi-supervised learning.  ...  We demonstrate on semi-supervised benchmark datasets SVHN and CIFAR-10 that adding our LC loss to MT yields significant improvements compared to MT and performance comparable to the state of the art in  ...  on a data manifold.  ... 
arXiv:2004.09665v2 fatcat:uk766o5ynzhmdhudtk3gziyiie

Semi-Supervised Learning with the Deep Rendering Mixture Model [article]

Tan Nguyen, Wanjia Liu, Ethan Perez, Richard G. Baraniuk, Ankit B. Patel
2016 arXiv   pre-print
Taken together, our work provides a unified framework for supervised, unsupervised, and semi-supervised learning.  ...  Semi-supervised learning algorithms reduce the high cost of acquiring labeled training data by using both labeled and unlabeled data during learning.  ...  For semi-supervised learning tasks on CIFAR10, the Improved GAN has the best classification error (18.63% and 17.72% test errors when N L ∈ {4K, 8K}).  ... 
arXiv:1612.01942v1 fatcat:q2onfll7qvcynm3n5azwrlb7em

Semi-supervised Learning using Adversarial Training with Good and Bad Samples [article]

Wenyuan Li, Zichen Wang, Yuguang Yue, Jiayun Li, William Speier, Mingyuan Zhou, Corey W. Arnold
2019 arXiv   pre-print
In this work, we investigate semi-supervised learning (SSL) for image classification using adversarial training.  ...  Triple-GAN, which aims to jointly optimize model components by incorporating three players, generates suitable image-label pairs to compensate for the lack of labeled data in SSL with improved benchmark  ...  Conclusions We have presented unified-GAN (UGAN), a new GAN framework for semi-supervised learning.  ... 
arXiv:1910.08540v1 fatcat:522qihzzunesldxoqvsjfs6mwi

Attribute Guided Unpaired Image-to-Image Translation with Semi-supervised Learning [article]

Xinyang Li, Jie Hu, Shengchuan Zhang, Xiaopeng Hong, Qixiang Ye, Chenglin Wu, Rongrong Ji
2019 arXiv   pre-print
Especially, AGUIT benefits from two-fold: (1) It adopts a novel semi-supervised learning process by translating attributes of labeled data to unlabeled data, and then reconstructing the unlabeled data  ...  AGUIT considers multi-modal and multi-domain tasks of UIT jointly with a novel semi-supervised setting, which also merits in representation disentanglement and fine control of outputs.  ...  Several recent works leveraged GANs for semi-supervised learning of classification Step I. Representation Decomposition Step II. Reconstruction and Translation Step III.  ... 
arXiv:1904.12428v1 fatcat:vifa3uqt5ralrb7pdnd2ojrzii

Partition-Guided GANs [article]

Mohammadreza Armandpour, Ali Sadeghian, Chunyuan Li, Mingyuan Zhou
2021 arXiv   pre-print
Despite the success of Generative Adversarial Networks (GANs), their training suffers from several well-known problems, including mode collapse and difficulties learning a disconnected set of manifolds  ...  In this paper, we break down the challenging task of learning complex high dimensional distributions, supporting diverse data samples, to simpler sub-tasks.  ...  [49] , motivated by the better performance of supervised-GANs, propose using a small set of labels and a semi-supervised method to infer the labels for the entire data.  ... 
arXiv:2104.00816v2 fatcat:n2q5zfbx7jfybfawg26xk6orgm

RoCGAN: Robust Conditional GAN

Grigorios G. Chrysos, Jean Kossaifi, Stefanos Zafeiriou
2020 International Journal of Computer Vision  
The focus so far has largely been on performance improvement, with little effort in making cGANs more robust to noise.  ...  We prove that RoCGAN share similar theoretical properties as GAN and establish with both synthetic and real data the merits of our model.  ...  We learn this structure with an unsupervised module which is included along with our supervised pathway.  ... 
doi:10.1007/s11263-020-01348-5 fatcat:dep7zvp4ene23dlg42qw42vdza

Improving GAN Training via Binarized Representation Entropy (BRE) Regularization [article]

Yanshuai Cao, Gavin Weiguang Ding, Kry Yik-Chau Lui, Ruitong Huang
2018 arXiv   pre-print
The approach also leads to higher classification accuracies in semi-supervised learning.  ...  We propose a novel regularizer to improve the training of Generative Adversarial Networks (GANs).  ...  Table. 2 shows results on CIFAR10 with feature matching semi-supervised learning GAN.  ... 
arXiv:1805.03644v1 fatcat:pq3levwpcbhcjaornzetrnqhu4

Yes, we GAN: Applying Adversarial Techniques for Autonomous Driving [article]

Michal Uricar, Pavel Krizek, David Hurych, Ibrahim Sobh, Senthil Yogamani, Patrick Denny
2019 arXiv   pre-print
This paper focuses on the application of GAN in autonomous driving including topics such as advanced data augmentation, loss function learning, semi-supervised learning, etc.  ...  Research on GAN is rapidly growing and there are many variants of the original GAN focusing on various aspects of deep learning.  ...  We would also like to thank the company Borealis AI for the GAN T-shirts which inspired the title of this paper 4 .  ... 
arXiv:1902.03442v1 fatcat:ev4dkg6wq5cjvbuy6htwlojzda

Semi-supervised and weakly-supervised learning with spatio-temporal priors in medical image segmentation

Gabriele Valvano
We make this possible by constraining the data representation learned by our method to be semantic or by regularising the model predictions to satisfy data-driven spatio-temporal priors.  ...  With the advent of faster and higher-quality imaging technologies, the amount of data that is possible to collect for each patient is paving the way toward personalised medicine.  ...  model and improve test-time performance in semi-supervised learning.  ... 
doi:10.13118/imtlucca/e-theses/344/ fatcat:qru63k6hibed3pwtxemhd523ua

Pairwise Augmented GANs with Adversarial Reconstruction Loss [article]

Aibek Alanov, Max Kochurov, Daniil Yashkov, Dmitry Vetrov
2018 arXiv   pre-print
We propose a novel autoencoding model called Pairwise Augmented GANs. We train a generator and an encoder jointly and in an adversarial manner.  ...  Here we train a discriminator to distinguish two types of pairs: an object with its augmentation and the one with its reconstruction.  ...  For example, it is used in semi-supervised learning , in a manipulation of object properties using low dimensional manifold (Creswell et al., 2017) and in an optimization utilizing the known structure  ... 
arXiv:1810.04920v1 fatcat:cgmhncnl4vbuhbb4yzulcgw5ty

DNA-GAN: Learning Disentangled Representations from Multi-Attribute Images [article]

Taihong Xiao, Jiapeng Hong, Jinwen Ma
2018 arXiv   pre-print
In this paper, we propose a supervised learning model called DNA-GAN which tries to disentangle different factors or attributes of images.  ...  images with the existence of the corresponding attribute being changed.  ...  In the semi-supervised setting, Siddharth et al. (2016) tried to learn a disentangled representations by using an auxiliary variable.  ... 
arXiv:1711.05415v2 fatcat:upgqokc7afb5tppmqz2hkuu6hm

Survey on Implementations of Generative Adversarial Networks for Semi-Supervised Learning

Ali Reza Sajun, Imran Zualkernan
2022 Applied Sciences  
Given recent advances in deep learning, semi-supervised techniques have seen a rise in interest.  ...  Generative adversarial networks (GANs) represent one recent approach to semi-supervised learning (SSL). This paper presents a survey method using GANs for SSL.  ...  GANs for semi-supervised learning.  ... 
doi:10.3390/app12031718 fatcat:x4skf2zvvvfornkmhelaijiwwu
« Previous Showing results 1 — 15 out of 548 results