Filters








1,798 Hits in 4.3 sec

Variable Skipping for Autoregressive Range Density Estimation [article]

Eric Liang, Zongheng Yang, Ion Stoica, Pieter Abbeel, Yan Duan, Xi Chen
2020 arXiv   pre-print
In this paper, we explore a technique, variable skipping, for accelerating range density estimation over deep autoregressive models.  ...  ., database cardinality estimation) require estimating range densities, a capability that is under-explored by current neural density estimation literature.  ...  Building on prior work, we distill and evaluate a more general optimization for accelerating range density estimation termed variable skipping.  ... 
arXiv:2007.05572v1 fatcat:6eksq3er4fhj3hr7i4vexn7hai

Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Images [article]

Rewon Child
2021 arXiv   pre-print
Despite this, autoregressive models have historically outperformed VAEs in log-likelihood.  ...  We present a hierarchical VAE that, for the first time, generates samples quickly while outperforming the PixelCNN in log-likelihood on all natural image benchmarks.  ...  We also thank the detailed anonymous reviews we received for helping improve our submission.  ... 
arXiv:2011.10650v2 fatcat:dihrzakdejgoji4vfxnge3gjpe

True versus Spurious Long Memory in Cryptocurrencies

Dooruj Rambaccussing, Murat Mazibas
2020 Journal of Risk and Financial Management  
The estimated memory parameters show that volatility is persistent, and when volatility is measured by log range, it is borderline nonstationary.  ...  Panel A exhibits the GPH estimates, bias test, and skip-sampling tests with different rates of h = 4, 8 months for returns.  ...  For instance, a researcher estimates a Generalized Autoregressive Conditional Heteroscedasticity (GARCH) process to model volatility when the observations are generated by a Fractionally Integrated Generalized  ... 
doi:10.3390/jrfm13090186 fatcat:7jh2ibrygbfypk55kdeln4l4bi

BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modeling [article]

Lars Maaløe, Marco Fraccaro, Valentin Liévin, Ole Winther
2019 arXiv   pre-print
We observe that BIVA, in contrast to recent results, can be used for anomaly detection. We attribute this to the hierarchy of latent variables which is able to extract high-level semantic features.  ...  However, their performance in terms of test likelihood and quality of generated samples has been surpassed by autoregressive models without stochastic units.  ...  Figure 7 : Distribution of the KL(q(z L )||p(z L ))) estimate for each model, each target density p(z L ) and for different initial random seeds.  ... 
arXiv:1902.02102v3 fatcat:trarar7u5fdepeqo7qb7vz6pjy

A Test of the Long Memory Hypothesis Based on Self-Similarity

James Davidson, Dooruj Rambaccussing
2015 Journal of Time Series Econometrics  
AbstractThis paper develops a new test of true versus spurious long memory, based on log-periodogram estimation of the long memory parameter using skip-sampled data.  ...  A correction factor is derived to overcome the bias in this estimator due to aliasing.  ...  Acknowledgement: We thank David Peel for helpful discussions on this problem, and an anonymous referee for perceptive comments which have materially improved the paper.  ... 
doi:10.1515/jtse-2013-0036 fatcat:ja5pesyocvhn5m5kmqu2su7eea

Asymptotic Theory of Outlier Detection Algorithms for Linear Time Series Regression Models

Søren Johansen, Bent Nielsen
2016 Scandinavian Journal of Statistics  
We de...ne a number of outlier detection algorithms related to the Huber-skip and the Least Trimmed Squares estimators, including the 1-step Huber skip estimator and the Forward Search.  ...  Finally, we analyze the gauge, the fraction of wrongly detected outliers, for a number of outlier detection algorithms and establish an asymptotic normal and a Poisson theory for the gauge.  ...  , Christophe Croux, Jurgen Doornik and Silvelyn Zwanzig for comments to the manuscript.  ... 
doi:10.1111/sjos.12174 fatcat:cjrtmqfc5ng73kkflyaf3waepu

Autoregressive Energy Machines [article]

Charlie Nash, Conor Durkan
2019 arXiv   pre-print
We propose the Autoregressive Energy Machine, an energy-based model which simultaneously learns an unnormalized density and computes an importance-sampling estimate of the normalizing constant for each  ...  The Autoregressive Energy Machine achieves state-of-the-art performance on a suite of density-estimation tasks.  ...  Acknowledgements The authors thank George Papamakarios, Iain Murray, and Chris Williams for helpful discussion.  ... 
arXiv:1904.05626v1 fatcat:ylygw6j2lnbbtjkdkb3jm4phr4

A Unified Deep Model of Learning from both Data and Queries for Cardinality Estimation [article]

Peizhi Wu, Gao Cong
2021 arXiv   pre-print
Cardinality estimation is a fundamental problem in database systems.  ...  First, to enable using the supervised query information in the deep autoregressive model, we develop differentiable progressive sampling using the Gumbel-Softmax trick.  ...  We would like to thank Zizhong Meng (NTU) for helping with some of the experiments, and the anonymous reviewers for providing constructive feedback and valuable suggestions.  ... 
arXiv:2107.12295v1 fatcat:hmvedppunnbihfc5m72estcweu

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
In particular, this compendium covers energy-based models, variational autoencoders, generative adversarial networks, autoregressive models, normalizing flows, in addition to numerous hybrid approaches  ...  The neural autoregressive density estimator (NADE) [131] , which can be viewed as a mean-field approximation of a restricted Boltzmann machine, achieves this for binary data by placing time-dependent  ...  For the column "Exact Density", represents tractable densities, () approximate densities, and intractable densities.  ... 
arXiv:2103.04922v2 fatcat:nivlg3whyjhadhwdl2tsh5yciy

Learning high-level structures in HEP data with novel Deep Auto-Regressive Networks for Fast Simulation

Anna Zaborowska, Ioana Ifrim, Witold Pokorski
2019 Zenodo  
The aim is for the network to be able to capture nonlinear, long-range correlations and input varying dependencies with tractable, explicit probability densities.  ...  The following research report analyses the benefits of employing autoregressive models in comparison with previously proposed models and their ability for generalisation in the attempt of fitting multiple  ...  networks -In contrast with VAEs, the autoregressive model provides tractable likelihoods -Autoregressive sequential models have worked for audio (WaveNet), images (PixelCNN++) and text (Transformer):  ... 
doi:10.5281/zenodo.3599392 fatcat:bn5fkfjrizfytjy3rypiag5vsa

Outlier Detection Algorithms for Least Squares Time Series Regression

Soren Johansen, Bent Nielsen
2014 Social Science Research Network  
The methods include the Huber-skip M-estimator, 1-step Huber-skip M-estimators, in particular the Impulse Indicator Saturation, iterated 1-step Huber-skip M-estimators and the Forward Search.  ...  From the asymptotic results we establish a new asymptotic theory for the gauge of these methods, which is the expected frequency of falsely detected outliers.  ...  Asymptotic results for Huber-skip M-estimators We consider recent results on the Huber-skip M-estimator as well as for 1-step Huber-skip M-estimators and iterations thereof.  ... 
doi:10.2139/ssrn.2510281 fatcat:fhumhkycq5ccvmd4fxofwnwevi

Locally Masked Convolution for Autoregressive Models [article]

Ajay Jain and Pieter Abbeel and Deepak Pathak
2020 arXiv   pre-print
State-of-the-art estimators for natural images are autoregressive, decomposing the joint distribution over pixels into a product of conditionals parameterized by a deep neural network, e.g. a convolutional  ...  Using LMConv, we learn an ensemble of distribution estimators that share parameters but differ in generation order, achieving improved performance on whole-image density estimation (2.89 bpd on unconditional  ...  Acknowledgements We thank Paras Jain, Nilesh Tripuraneni, Joseph Gonzalez and Jonathan Ho for helpful discussions, and reviewers for helpful suggestions.  ... 
arXiv:2006.12486v3 fatcat:wbz2rnvhtjcepja7vfifcp4xey

Degeneration in VAE: in the Light of Fisher Information Loss [article]

Huangjie Zheng, Jiangchao Yao, Ya Zhang, Ivor W. Tsang
2018 arXiv   pre-print
We call this class of VAE equipped with skip connections as SCVAE and perform a range of experiments to show its advantages in information preservation and degeneration mitigation.  ...  Specifically, we propose a Fisher Information measure for the layer-wise analysis.  ...  X, but a function w.r.t. the probabilistic density p θ and useful for parametric estimation of distributions.  ... 
arXiv:1802.06677v3 fatcat:dxnlzropqvhufgef7wmp6bsie4

How to apply dynamic panel bootstrap-corrected fixed-effects (xtbcfe) and heterogeneous dynamics (panelhetero)

Samuel Asumadu SARKODIE, Phebe Asantewaa OWUSU
2020 MethodsX  
We further demonstrate how to use empirical CDF, moments and kernel density estimation to investigate heterogeneous effects.  ...  •We demonstrate how the dynamic panel bootstrap-corrected fixed-effects estimator is useful in estimating higher-order panel data models and accounting for challenges such as omitted-variable bias, convergence  ...  y i, t represents dependent variables, x i, t denotes strongly exogenous independent variables, γ denotes the autoregressive coefficient of lagged dependent variable, β represents estimated vector coefficients  ... 
doi:10.1016/j.mex.2020.101045 pmid:32939352 pmcid:PMC7479353 fatcat:yuqrqhd6bzb6xjvnfck6uzfuaq

Articulatory-WaveNet: Autoregressive Model For Acoustic-to-Articulatory Inversion [article]

Narjes Bozorg, Michael T.Johnson
2020 arXiv   pre-print
This paper presents Articulatory-WaveNet, a new approach for acoustic-to-articulator inversion.  ...  Results show significant improvement in both correlation and RMSE between the generated and true articulatory trajectories for the new method, with an average correlation of 0.83, representing a 36% relative  ...  [33] utilized DNN and a deep trajectory-Mixture Density Network (MDN) for estimating articulatory trajectories from acoustic signals.  ... 
arXiv:2006.12594v1 fatcat:y3xq5czyhjbkvhr4ilbqfwhztu
« Previous Showing results 1 — 15 out of 1,798 results