280 Hits in 4.0 sec

f-GANs in an Information Geometric Nutshell [article]

Richard Nock and Zac Cranko and Aditya Krishna Menon and Lizhen Qu and Robert C. Williamson
2017 arXiv   pre-print
their concinnity in the f-GAN game.  ...  Nowozin et al showed last year how to extend the GAN principle to all f-divergences.  ...  , the variational f -GAN formulation can be captured in an information-geometric framework by the following identity using Theorems 4, 7, 10, 11.  ... 
arXiv:1707.04385v1 fatcat:larmk3pz4ne6boy7vng7oakkfy

GraphAnoGAN: Detecting Anomalous Snapshots from Attributed Graphs [article]

Siddharth Bhatia, Yiwei Wang, Bryan Hooi, Tanmoy Chakraborty
2021 arXiv   pre-print
In this paper, we propose GraphAnoGAN, an anomalous snapshot ranking framework, which consists of two core components -- generative and discriminative models.  ...  These models do not take into account the multifaceted interactions between the structure and attributes in the network.  ...  For Anomalous and Dominant, an anomalous snapshot is determined using geometric mean of the per-node anomaly scores.  ... 
arXiv:2106.15504v1 fatcat:nd2i6qf5rbabzcpgunzdjef6se

Adversarial Networks and Autoencoders: The Primal-Dual Relationship and Generalization Bounds [article]

Hisham Husain, Richard Nock, Robert C. Williamson
2019 arXiv   pre-print
In this work, we study the f-GAN and WAE models and make two main discoveries.  ...  First, we find that the f-GAN and WAE objectives partake in a primal-dual relationship and are equivalent under some assumptions, which then allows us to explicate the success of WAE.  ...  (20) can be reparametrized by setting ϕ(x) ← x 2 /2−ϕ(x) and ψ(y) ← y 2 /2 − ψ(y), and so the constraint changes: The optimization problem in Equation  ... 
arXiv:1902.00985v2 fatcat:monu5zgn6bc5lmfavwmgwgroam

LOLNeRF: Learn from One Look [article]

Daniel Rebain, Mark Matthews, Kwang Moo Yi, Dmitry Lagun, Andrea Tagliasacchi
2022 arXiv   pre-print
Our experiments show that we achieve state-of-the-art results in novel view synthesis and high-quality results for monocular depth prediction.  ...  We demonstrate this by training models to reconstruct object categories using datasets that contain only one view of each subject without depth or geometry information.  ...  Introduction A long-standing challenge in computer vision is the extraction of 3D geometric information from images of the real world [37] .  ... 
arXiv:2111.09996v2 fatcat:flf4cqu3c5fhvhx7llxgn5zoj4

All One Needs to Know about Priors for Deep Image Restoration and Enhancement: A Survey [article]

Yunfan Lu, Yiqi Lin, Hao Wu, Yunhao Luo, Xu Zheng, Lin Wang
2022 arXiv   pre-print
research in the community; (5) An open-source repository that provides a taxonomy of all mentioned works and code links.  ...  However, the importance of priors has not been systematically studied and analyzed by far in the research community.  ...  Face restorations from ESRGAN [40] ,Super-FAN [41] ,HiFaceGAN [42] ,DFDNet [43] ,PSFR-GAN [44] . The PSFR-GAN employs semantic information. , (e)PULSE [46] , and (f)GFP-GAN [36] .  ... 
arXiv:2206.02070v1 fatcat:icu7hwua3jggbp7owl2l5mgyfu

A Convex Duality Framework for GANs [article]

Farzan Farnia, David Tse
2018 arXiv   pre-print
However, in practice the discriminator is constrained to be in a smaller class F such as neural nets.  ...  Unlike the f-divergence, we prove that the proposed hybrid divergence changes continuously with the generative model, which suggests regularizing the discriminator's Lipschitz constant in f-GAN and vanilla  ...  Acknowledgments: We are grateful for support under a Stanford Graduate Fellowship, the National Science Foundation grant under CCF-1563098, and the Center for Science of Information (CSoI), an NSF Science  ... 
arXiv:1810.11740v1 fatcat:uqurcy24lbfgthquds6fyjm24u

The Inductive Bias of Restricted f-GANs [article]

Shuang Liu, Kamalika Chaudhuri
2018 arXiv   pre-print
in a given function class, the distribution induced by the generator is restricted to lie in a pre-specified distribution class and the objective is similar to a variational form of the f-divergence.  ...  Specifically, we provide a theoretical characterization of the distribution inferred by a simple form of generative adversarial learning called restricted f-GANs -- where the discriminator is a function  ...  "f-GANs in an information geometric nutshell". In: Advances in Neural Information Processing Systems. 2017, pp. 456-464. [NWJ10] XuanLong Nguyen, Martin J Wainwright, and Michael I Jordan.  ... 
arXiv:1809.04542v1 fatcat:ihpvr6q555gdxfluq53kcygqee

Unsupervised Discovery of Interpretable Directions in the GAN Latent Space [article]

Andrey Voynov, Artem Babenko
2020 arXiv   pre-print
In this paper, we introduce an unsupervised method to identify interpretable directions in the latent space of a pretrained GAN model.  ...  The latent spaces of GAN models often have semantically meaningful directions.  ...  In essence, GANs consist of two networks -a generator and a discriminator, which are trained jointly in an adversarial manner.  ... 
arXiv:2002.03754v3 fatcat:iq6pttabwvavlc5qbw6ntochli

Optimizing the Latent Space of Generative Networks [article]

Piotr Bojanowski, Armand Joulin, David Lopez-Paz, Arthur Szlam
2019 arXiv   pre-print
In most successful applications, GAN models share two common aspects: solving a challenging saddle point optimization problem, interpreted as an adversarial game between a generator and a discriminator  ...  Generative Adversarial Networks (GANs) have achieved remarkable results in the task of generating realistic natural images.  ...  In a nutshell, GLO can be viewed as an "encoderless" autoencoder, or as a "discriminator-less" GAN. Choice of Z. A common choice of Z in the GAN literature is from a Normal distribution on R d .  ... 
arXiv:1707.05776v2 fatcat:phgf7vtzgzc53oocu3yvtr4yxa

Learning disconnected manifolds: a no GAN's land

Ugo Tanielian, Thibaut Issenhuth, Elvis Dohmatob, Jérémie Mary
2020 International Conference on Machine Learning  
We formalize this problem by establishing a "no free lunch" theorem for the disconnected manifold learning stating an upper-bound on the precision of the targeted distribution.  ...  In a nutshell, our contributions are the following: • We discuss evaluation of GANs and formally link the PR measure (Sajjadi et al., 2018) and its Improved PR version (Kynkäänniemi et al., 2019) .  ...  Learning disconnected manifolds leads to the apparition of an area with high gradients and data sampled in between modes. (a) WGAN 4 classes: visualisation of JG(z) F .  ... 
dblp:conf/icml/TanielianIDM20 fatcat:5tbudjczlne7fo4sfbb7no6tba

MarginGAN: Adversarial Training in Semi-Supervised Learning

Jinhao Dong, Tong Lin
2019 Neural Information Processing Systems  
Recently, generative adversarial networks (GANs) has been applied to SSL and obtained amazing results. The method of Feature Matching (FM) GANs proposed in [8] substitutes the original binary  ...  Pseudo labels are used for generated and unlabeled examples in training.  ...  Margins in semi-supervised learning [21] propose the margin of an unlabeled example denoted as |f (x)|, that can be also represented as yf (x) with pseudo label y = sign(f (x)).  ... 
dblp:conf/nips/DongL19 fatcat:bkdcpbzbefaefovsbc2eehkv6q

AlGaN-Based 1.55 µm Phototransistor as a Crucial Building Block for Optical Computers

Daniel Hofstetter, Cynthia Aku-Leh, Hans Beck, David P. Bour
2021 Crystals  
Due to partial screening of the strong internal polarization fields between GaN quantum wells and AlN barriers, this slightly diagonal transition generates an optical rectification voltage.  ...  An optically activated, enhancement mode heterostructure field effect transistor is proposed and analytically studied.  ...  A more comprehensive description can be found in reference [11] . In a nutshell, the epitaxial layers are grown on top of a C-face sapphire substrate.  ... 
doi:10.3390/cryst11111431 fatcat:lfpqh63egrcyrkfchhvpgvohki

High-throughput, high-resolution deep learning microscopy based on registration-free generative adversarial network

Hao Zhang, Chunyu Fang, Xinlin Xie, Yicong Yang, Wei Mei, Di Jin, Peng Fei
2019 Biomedical Optics Express  
By appropriately adopting prior microscopy data in an adversarial training, the neural network can recover a high-resolution, accurate image of new specimen from its single low-resolution measurement.  ...  The gigapixel, multi-color reconstruction of these samples verifies a successful GAN-based single image super-resolution procedure.  ...  Acknowledgements The authors acknowledge the selfless sharing of the GAN source codes from Hao Dong (, as well as the contributions of Tinting Zhu for assistance with fluorescent  ... 
doi:10.1364/boe.10.001044 pmid:30891329 pmcid:PMC6420277 fatcat:zryz3fw35fayhejhthogwkjxle

Deep CNN and Deep GAN in Computational Visual Perception-Driven Image Analysis

R. Nandhini Abirami, P. M. Durai Raj Vincent, Kathiravan Srinivasan, Usman Tariq, Chuan-Yu Chang, Dr Shahzad Sarfraz
2021 Complexity  
This survey incorporates an overview of existing applications of deep learning in computational visual perception.  ...  The computer vision's goal is to surpass the capabilities of biological vision in extracting useful information from visual data.  ...  In a nutshell, the dropout technique assumes that a randomly selected portion of the network is muted for each training case [75] .  ... 
doi:10.1155/2021/5541134 fatcat:xluxbl7kojbvxpjq5u726d3djm

Atomistic simulation of surface passivated wurtzite nanowires: electronic bandstructure and optical emission

Vinay U. Chimalgi, Md Rezaul Karim Nishat, Krishna K. Yalavarthi, Shaikh S. Ahmed
2014 Advances in nano research  
reported values in bulk structures, and b) the localization of wavefunctions and the optical anisotropy in GaN/AlN disk-in-wire nanowires.  ...  The three-dimensional Nano-Electronic Modeling toolkit is an open source software package that allows the atomistic calculation of single-particle electronic states and optical response of various semiconductor  ...  As an example, a two-atom (anion and cation) Hamiltonian construction is described via the following matrix ac i j i j u v V f l m n V  (2) where, f i,j (l,m,n) are the two center Slater-Koster integrals  ... 
doi:10.12989/anr.2014.2.3.157 fatcat:iufmf7mk5nhgtjxbko4l6wv5ji
« Previous Showing results 1 — 15 out of 280 results