Filters








221 Hits in 3.5 sec

The Convolution Exponential and Generalized Sylvester Flows [article]

Emiel Hoogeboom, Victor Garcia Satorras, Jakub M. Tomczak, Max Welling
2020 arXiv   pre-print
In addition, we generalize Sylvester Flows and propose Convolutional Sylvester Flows which are based on the generalization and the convolution exponential as basis change.  ...  Empirically, we show that the convolution exponential outperforms other linear transformations in generative flows on CIFAR10 and the graph convolution exponential improves the performance of graph normalizing  ...  Convolutional Sylvester Flows Generalized Sylvester flows and the convolution exponential can be naturally combined to obtain Convolutional Sylvester Flows (CSFs).  ... 
arXiv:2006.01910v2 fatcat:t3bdtfbkbbfqphe6ww2pf22fbm

A characterization of the Frobenius problem and its application to arithmetic progressions [article]

Hans J. H. Tuenter
2006 arXiv   pre-print
We give a functional relationship that completely characterizes the set NR, and apply it to the case when the numbers are in an arithmetic progression.  ...  ,a_k, and are interested in the set of positive numbers NR that have no representation by the linear form ∑_i a_ix_i in nonnegative integers x_1, x_2,...,x_k.  ...  By expanding the exponential in the left-hand side of (3.3), and changing the order of summation we see that we have also obtained the exponential generating function of the Sylvester sums: ∞ m=0 S m z  ... 
arXiv:math/0606610v1 fatcat:yfzhrlv27fet5pyamgnl22yomy

MintNet: Building Invertible Neural Networks with Masked Convolutions [article]

Yang Song and Chenlin Meng and Stefano Ermon
2019 arXiv   pre-print
Additionally, the determinant of the Jacobian can be computed analytically and efficiently, enabling their generative use as flow models.  ...  When trained as generative models, our invertible networks achieve competitive likelihoods on MNIST, CIFAR-10 and ImageNet 32x32, with bits per dimension of 0.98, 3.32 and 4.06 respectively.  ...  For example, planar flows [27] and Sylvester flow [2] constrain the number of hidden units to be smaller than the input dimension.  ... 
arXiv:1907.07945v2 fatcat:c36v2br2mvcahn3muzropbrd6e

Quasi-Autoregressive Residual (QuAR) Flows [article]

Achintya Gopal
2020 arXiv   pre-print
Compared to the standard residual flow approach, this simplification retains many of the benefits of residual flows while dramatically reducing the compute time and memory requirements, thus making flow-based  ...  The current state of the art results are built upon residual flows as these can model a larger hypothesis space than coupling layers.  ...  We thank Aaron Key (Bloomberg Quant Research) for discussions about normalizing flows and the plots in the paper.  ... 
arXiv:2009.07419v1 fatcat:rmcgj6n7czhuhi3uxztxzaiqe4

Turbidite bed thickness statistics of architectural elements in a deep-marine confined mini-basin setting: Examples from the Grès d'Annot Formation, SE France

G. Pantopoulos, B.C. Kneller, A.D. McArthur, S. Courivaud, A.E. Grings, J. Kuchle
2018 Marine and Petroleum Geology  
Several datasets exhibit power law as well as exponential thick-bedded tails. The data also exhibit non-random clustering of bed thickness.  ...  Discrimination of architectural elements in this confined turbidite succession seems to be feasible based on the characteristics of the observed composite lognormal distributions such as number and variability  ...  We also thank Shell and former BG Brasil for financial support and permission to publish results.  ... 
doi:10.1016/j.marpetgeo.2018.04.008 fatcat:qthopc7hebeajfmmwjje2alhyq

Low-rate turbo-Hadamard codes

Li Ping, W.K. Leung, K.Y. Wu
2003 IEEE Transactions on Information Theory  
Traditional low-rate codes, such as Hadamard codes [5], [6] and super-orthogonal convolutional codes [7], are not well suited for this purpose due to their relatively low coding gain.  ...  Both simulation and analytical results are provided to demonstrate the advantages of the proposed scheme.  ...  ACKNOWLEDGMENT The authors wish to thank the Associate Editor and the anonymous reviewers for their careful reading and constructive comments to improve the paper. They would also like to thank S.  ... 
doi:10.1109/tit.2003.820018 fatcat:pot4fylm5nguhg6sv4r4gvorly

Stable Blind Deconvolution over the Reals from Additional Autocorrelations [article]

Philipp Walk, Babak Hassibi
2017 arXiv   pre-print
Moreover, the stability constant depends on the signal dimension and on the signals magnitude of the first and last coefficients.  ...  We will show in this work that under a sufficient zero separation of the corresponding signal in the z-domain, a stable reconstruction against additive noise is possible.  ...  Since the convolution is a product of two input signals in the frequency domain, it can be generated by uncountably many input signals.  ... 
arXiv:1710.07879v1 fatcat:bnlyksbbfbdytafjpqxgl2rcfy

Distributed control of spatially invariant systems

B. Bamieh, F. Paganini, M.A. Dahleh
2002 IEEE Transactions on Automatic Control  
in the spring 1993. He has held consulting positions with several companies in the United States and abroad.  ...  His interests include robust control and identification, the development of computational methods for linear and nonlinear controller design, and applications of feedback control in several disciplines  ...  However, in general the Riccati solutions and will not be differential operators ( and are not rational in general, see the example in Section V-B), and their convolution kernels will have a spread, reflecting  ... 
doi:10.1109/tac.2002.800646 fatcat:ygyou56z3vagnmfqheswpnt3hy

Quantitative approach in environmental interpretations of deep-marine sediments (Dukla Unit, Western Carpathian Flysch Zone)

Marta Prekopová, Juraj Janočko
2009 Geologica Carpathica  
We analysed Paleogene deep-water sediments belonging to the Cisna, Sub-Menilite, and Menilite Formations of the Dukla Unit, Outer Carpathian Flysch Zone and, using two independent quantitative methods,  ...  The second one is the lognormal mixture model of Talling (2001). Based on a quantitative approach, we suggest deposition of the lowermost Cisna Formation in the channel-levee environment.  ...  Sylvester and W. Winkler, who reviewed, commented and considerably improved the manuscript.  ... 
doi:10.2478/v10096-009-0035-y fatcat:jd2rliqvabf6dgywxb6g5mgzgi

Model Selection for Bayesian Autoencoders [article]

Ba-Hien Tran and Simone Rossi and Dimitrios Milios and Pietro Michiardi and Edwin V. Bonilla and Maurizio Filippone
2021 arXiv   pre-print
We carry out posterior estimation of the BAE parameters via stochastic gradient Hamiltonian Monte Carlo and turn our BAE into a generative model by fitting a flexible Dirichlet mixture model in the latent  ...  distance (DSWD) between the output of the autoencoder and the empirical data distribution.  ...  Acknowledgments and Disclosure of Funding MF gratefully acknowledges support from the AXA Research Fund and the Agence Nationale de la Recherche (grant ANR-18-CE46-0002 and ANR-19-P3IA-0002).  ... 
arXiv:2106.06245v1 fatcat:tztxolyjpngs5pmjvtpiqxi6pm

Generative Time-series Modeling with Fourier Flows

Ahmed M. Alaa, Alex James Chan, Mihaela van der Schaar
2021 International Conference on Learning Representations  
Most of the recently proposed methods for generating synthetic time-series rely on implicit likelihood modeling using generative adversarial networks (GANs)-but such models can be difficult to train, and  ...  We show that, by virtue of the DFT analytic properties, the Jacobian determinants and inverse mapping for the Fourier flow can be computed efficiently in linearithmic time, without imposing explicit structural  ...  The research presented in this paper was supported by the US Office of Naval Research (ONR), and by the National Science Foundation (NSF) under grant numbers 1407712, 1462245, 1524417, 1533983, and 1722516  ... 
dblp:conf/iclr/AlaaCS21 fatcat:4y4fdywf5bes5owj5yrb5e74kq

Relative contribution of V-H+ATPase and NA+/H+exchanger to bicarbonate reabsorption in proximal convoluted tubules of old rats

Mariana Fiori, Martin Radrizzani, Paula Diaz-Sylvester, Angelica Muller, Tomas Corti, Alberto Monserrat, Carlos Amorena
2006 Aging Cell  
(10 −6 M) (B) and with EIPA (10 −4 M) (EIPA) in young (3-month) (A) and old (18-month) (B) rats compared with the value observed without treatment (C).  ...  ) and old(18-month) rats.  ...  The authors would like to thank Dr. A. A. Altamirano for critical reading of this manuscript and Dr. H. Schteingart and Lic. G.  ... 
doi:10.1111/j.1474-9726.2006.00229.x pmid:16968310 fatcat:zxbpvss2x5fadbx3cqltzwgerm

Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models [article]

Sam Bond-Taylor, Adam Leach, Yang Long, Chris G. Willcocks
2021 arXiv   pre-print
These techniques are drawn under a single cohesive framework, comparing and contrasting to explain the premises behind each, while reviewing current state-of-the-art advances and implementations.  ...  Deep generative modelling is a class of techniques that train deep neural networks to model the distribution of training samples.  ...  A higher rank generalisation of the matrix determinant lemma has been applied to planar flows, known as Sylvester flows, removing the severe bottleneck thus allowing greater representation ability [14  ... 
arXiv:2103.04922v2 fatcat:nivlg3whyjhadhwdl2tsh5yciy

List of contents and Author Index, Volume 19, 2006

2006 Applied Mathematics Letters  
Chan and H.Y. Tian 298 A criterion for a multi-dimensional explosion due to a concentrated nonlinear source NUMBER 4 J. Sokół 303 Convolution and subordination in the convex hull of convex mappings J.  ...  Bouagada and P. Van Dooren 451 Stability margins for generalized state space systems V.V. Gubernov, H.S. Sidhu and G.N. Mercer 458 Generalized compound matrix method J. Li and N.-j.  ...  Hu and D. Cheng 859 The polynomial solution to the Sylvester matrix equation M. Silimbani 865 Semistandard tableaux associated with generalized labellings of posets J.  ... 
doi:10.1016/s0893-9659(06)00219-9 fatcat:yaqob7cf6jenxoegamqj6icw5u

Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Models

Sam Bond-Taylor, Adam Leach, Yang Long, Chris George Willcocks
2021 IEEE Transactions on Pattern Analysis and Machine Intelligence  
These techniques are compared and contrasted, explaining the premises behind each and how they are interrelated, while reviewing current state-of-the-art advances and implementations.  ...  Deep generative models are a class of techniques that train deep neural networks to model the distribution of training samples.  ...  A higher rank generalisation of the matrix determinant lemma has been applied to planar flows, known as Sylvester flows, removing the severe bottleneck thus allowing greater representation ability [14  ... 
doi:10.1109/tpami.2021.3116668 pmid:34591756 fatcat:yjpayhmrfnaeziahmrgiyvtxkm
« Previous Showing results 1 — 15 out of 221 results