1,407 Hits in 5.0 sec

Robust Estimation and Generative Adversarial Nets [article]

Chao Gao, Jiyi Liu, Yuan Yao, Weizhi Zhu
2019 arXiv   pre-print
This connection opens the door of computing robust estimators using tools developed for training GANs.  ...  In particular, we show in both theory and experiments that some appropriate structures of discriminator networks with hidden layers in GANs lead to statistically optimal robust location estimators for  ...  Robust Mean Estimation via GAN In this section, we focus on the problem of robust mean estimation under Huber's -contamination model.  ... 
arXiv:1810.02030v3 fatcat:p34ofwcqyje63o3p7gmaewma5i

Robust W-GAN-Based Estimation Under Wasserstein Contamination [article]

Zheng Liu, Po-Ling Loh
2021 arXiv   pre-print
Specifically, we analyze properties of Wasserstein GAN-based estimators for location estimation, covariance matrix estimation, and linear regression and show that our proposed estimators are minimax optimal  ...  Although minimax rates of estimation have been established in recent years, many existing robust estimators with provably optimal convergence rates are also computationally intractable.  ...  Optimal Transport: Old and New, Volume 338. Springer Science & Business Media. Wu, K., G. W. Ding, R. Huang, and Y. Yu (2020). On minimax optimality of GANs for robust mean estimation.  ... 
arXiv:2101.07969v1 fatcat:ow2hi65tkjaodph5vpreklhbqa

Robust Density Estimation under Besov IPM Losses [article]

Ananya Uppal, Shashank Singh, Barnabas Poczos
2021 arXiv   pre-print
Specifically, under a range of smoothness assumptions on the population and outlier distributions, we show that a re-scaled thresholding wavelet series estimator achieves minimax optimal convergence rates  ...  Finally, based on connections that have recently been shown between nonparametric density estimation under IPM losses and generative adversarial networks (GANs), we show that certain GAN architectures  ...  is a GAN estimate P that converges at the same rate and so is robust minimax optimal.  ... 
arXiv:2004.08597v2 fatcat:t7xd44talzhdhoaxaqzuccaqsy

Bregman learning for generative adversarial networks

Jian Gao, Hamidou Tembine
2018 2018 Chinese Control And Decision Conference (CCDC)  
This connection opens the door of computing robust estimators using tools developed for training GANs.  ...  In particular, we show in both theory and experiments that some appropriate structures of discriminator networks with hidden layers in GANs lead to statistically optimal robust location estimators for  ...  ROBUST MEAN ESTIMATION VIA GAN In this section, we focus on the problem of robust mean estimation under Huber's -contamination model.  ... 
doi:10.1109/ccdc.2018.8407110 fatcat:f3vt7tzw2re7nf4g6qdnpq2suq

Adversarial Learning and Augmentation for Speaker Recognition

Jen-Tzung Chien, Kang-Ting Peng
2018 Odyssey 2018 The Speaker and Language Recognition Workshop  
Our idea is to incorporate the class label into GAN which involves a minimax optimization problem for adversarial learning.  ...  In addition to the minimax optimization of adversarial loss, the posterior probabilities of class labels given real and fake samples are maximized.  ...  The layer-wise generator and discriminator are jointly estimated by solving a minimax optimization problem based on stochastic gradient descent (SGD) algorithm.  ... 
doi:10.21437/odyssey.2018-48 dblp:conf/odyssey/ChienP18 fatcat:tac6nolat5e4fim44qrw7jktv4

Minimax Defense against Gradient-based Adversarial Attacks [article]

Blerta Lindqvist, Rauf Izmailov
2020 arXiv   pre-print
Our minimax classifier is the discriminator of a generative adversarial network (GAN) that plays a minimax game with the GAN generator.  ...  Our Minimax adversarial approach presents a significant shift in defense strategy for neural network classifiers.  ...  Based on minimax optimization in GANs, the first, novel approach counters the assumption of gradient-based attacks that neural network classifiers perform gradient descent optimization.  ... 
arXiv:2002.01256v1 fatcat:ivus7rlt4na3nb6rvrxth5mvle

Train simultaneously, generalize better: Stability of gradient-based minimax learners [article]

Farzan Farnia, Asuman Ozdaglar
2020 arXiv   pre-print
The success of minimax learning problems of generative adversarial networks (GANs) has been observed to depend on the minimax optimization algorithm used for their training.  ...  This dependence is commonly attributed to the convergence speed and robustness properties of the underlying optimization algorithm.  ...  The primary focus of optimization-related studies of minimax learning problems has been on the convergence and robustness properties of minimax optimization algorithms.  ... 
arXiv:2010.12561v1 fatcat:tju5hfqcnnbmth3rumlw6tregi

GAN Slimming: All-in-One GAN Compression by A Unified Optimization Framework [article]

Haotao Wang, Shupeng Gui, Haichuan Yang, Ji Liu, Zhangyang Wang
2020 arXiv   pre-print
To this end, we propose the first unified optimization framework combining multiple compression means for GAN compression, dubbed GAN Slimming (GS).  ...  GS seamlessly integrates three mainstream compression techniques: model distillation, channel pruning and quantization, together with the GAN minimax objective, into one unified optimization form, that  ...  compression means for GANs?  ... 
arXiv:2008.11062v1 fatcat:567bndbj5re2himu2kwwyiz7y4

Generative Adversarial Network: Some Analytical Perspectives [article]

Haoyang Cao, Xin Guo
2021 arXiv   pre-print
This subchapter will start from an introduction of GANs from an analytical perspective, then move on to the training of GANs via SDE approximations and finally discuss some applications of GANs in computing  ...  Ever since its debut, generative adversarial networks (GANs) have attracted tremendous amount of attention.  ...  The connection between mean-field games and GANs via the minimax structure presents GANs' potential computing power for high dimensional control and optimization problems with variational structures.  ... 
arXiv:2104.12210v2 fatcat:kzzab5vr4bfilont4fdgu24bla

Learning Implicit Generative Models by Teaching Explicit Ones [article]

Chao Du, Kun Xu, Chongxuan Li, Jun Zhu, Bo Zhang
2020 arXiv   pre-print
Generative adversarial nets (GANs) present a minimax framework to train such models, which however can suffer from mode collapse due to the nature of the JS-divergence.  ...  GANs.  ...  Also, we find both GAN and VAE-GAN are sensitive to the architectures of G and D, whereas LBT-GAN is much more robust.  ... 
arXiv:1807.03870v3 fatcat:4ervyi3oezfpdesvqeb2c46dge

Generative Multi-Adversarial Networks [article]

Ishan Durugkar, Ian Gemp, Sridhar Mahadevan
2017 arXiv   pre-print
Generative adversarial networks (GANs) are a framework for producing a generative model by way of a two-player minimax game.  ...  In previous work, the successful training of GANs requires modifying the minimax objective to accelerate training early on.  ...  Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF.  ... 
arXiv:1611.01673v3 fatcat:ghxxid475betjiantrlln4h2zm

Page 3831 of Mathematical Reviews Vol. , Issue 97F [page]

1997 Mathematical Reviews  
on the compound parameter is asymptotically optimal.  ...  The entropy loss function has been used extensively to dis- cuss the minimax estimator. For example, M. C. Yang [Statist.  ... 

Generative Minimization Networks: Training GANs Without Competition [article]

Paulina Grnarova, Yannic Kilcher, Kfir Y. Levy, Aurelien Lucchi, Thomas Hofmann
2021 arXiv   pre-print
However, recent applications of generative models, particularly GANs, have triggered interest in solving min-max games for which standard optimization techniques are often not suitable.  ...  At the heart of these problems is the min-max structure of the GAN objective which creates non-trivial dependencies between the players.  ...  Note that our optimization setting can be applied to any minimax formulation of GANs by optimizing the DG for the specific game objective respectively (GAN, WGAN etc.).  ... 
arXiv:2103.12685v1 fatcat:pn6ddwxv7jgv5oypsq3zfodyse

Generative Adversarial Nets for Robust Scatter Estimation: A Proper Scoring Rule Perspective [article]

Chao Gao, Yuan Yao, Weizhi Zhu
2019 arXiv   pre-print
techniques that optimize GANs.  ...  The recent discovery on the connection between robust estimation and generative adversarial nets (GANs) by Gao et al. (2018) suggests that it is possible to compute depth-like robust estimators using similar  ...  The authors also thank Jiantao Jiao for pointing out references on scoring rules and for helpful discussion during this project.  ... 
arXiv:1903.01944v1 fatcat:y7lcp4k7cjceda32fegosv466u

A Robust Adversarial Network-Based End-to-End Communications System With Strong Generalization Ability Against Adversarial Attacks [article]

Yudi Dong and Huaxia Wang and Yu-Dong Yao
2021 arXiv   pre-print
We propose a novel defensive mechanism based on a generative adversarial network (GAN) framework to defend against adversarial attacks in end-to-end communications systems.  ...  Specifically, we utilize a generative network to model a powerful adversary and enable the end-to-end communications system to combat the generative attack network via a minimax game.  ...  Consensus Optimization For GAN Training The stability and convergence of GAN training is a very challenge task, which suffers from the problems of nonconvergence, mode collapse, and diminished gradient  ... 
arXiv:2103.02654v1 fatcat:wszuivbvgrdbnhs5icubs6y6yy
« Previous Showing results 1 — 15 out of 1,407 results