Filters








48,622 Hits in 4.4 sec

Bayesian graph convolutional neural networks via tempered MCMC [article]

Rohitash Chandra, Ayush Bhagat, Manavendra Maharana, Pavel N. Krivitsky
2021 arXiv   pre-print
Recent advances in parallel computing and advanced proposal schemes in sampling, such as incorporating gradients has allowed Bayesian deep learning methods to be implemented.  ...  Bayesian inference provides a principled and robust approach to uncertainty quantification of model parameters for deep learning models.  ...  The comparison of results with the literature motivates the implementation of Bayesian framework for other deep learning model, which includes LSTM and CNN models [48] .  ... 
arXiv:2104.08438v1 fatcat:ot3wt2mobzggxn3zusrpkxaj6u

Bayesian graph convolutional neural networks via tempered MCMC

Rohitash Chandra, Ayush Bhagat, Manavendra Maharana, Pavel N. Krivitsky
2021 IEEE Access  
The comparison of results with the literature motivates the implementation of Bayesian framework for other deep learning model, which includes LSTM and CNN models [48] .  ...  incorporating transfer learning to take advantage of multiple sources of data in a Bayesian framework via Langevin MCMC sampling [75] .  ... 
doi:10.1109/access.2021.3111898 fatcat:kwwwa7vdcrgv3hm5ainkkmpiba

Bayesian analysis of Box–Cox transformed linear mixed models with ARMA() dependence

Jack C. Lee, Tsung I. Lin, Kuo J. Lee, Ying L. Hsu
2005 Journal of Statistical Planning and Inference  
Two priors are proposed and put into comparisons in parameter estimation and prediction of future values.  ...  In this paper, we present a Bayesian inference methodology for Box-Cox transformed linear mixed model with ARMA(p, q) errors using approximate Bayesian and Markov chain Monte Carlo methods.  ...  Bayesian inference via MCMC sampling The algorithm The following sampling scheme is used to obtain the posterior distributions of , 2 , , and .  ... 
doi:10.1016/j.jspi.2004.03.015 fatcat:iamin3fckvgbpb4pltv3ynwvny

Data Mining of Perishable Food Safety Sampling based on Voting

Anqi Hu, Tongjuan Liu
2018 DEStech Transactions on Computer Science and Engineering  
The prediction model can predict the condition of perishable food, which innovatively guiding the safety inspection work; by choosing the safety rapid detection for typical effective sample of perishable  ...  Different models had been made, which by selecting the neural network algorithm, classification and regression tree algorithm and Bayesian network algorithm in the data mining software.  ...  network modeling The optimal prediction model of Bayesian network End Bayesian network modeling Maximum assessment (Comparativ e)?  ... 
doi:10.12783/dtcse/csae2017/17467 fatcat:43gppej6ojdevlx6txfhny5z4y

Surrogate-assisted parallel tempering for Bayesian neural learning [article]

Rohitash Chandra, Konark Jain, Arpit Kapoor, Ashray Aman
2020 arXiv   pre-print
Hence, we present surrogate-assisted parallel tempering for Bayesian neural learning for simple to computationally expensive models.  ...  Parallel tempering MCMC addresses some of these limitations given that they can sample multimodal posterior distributions and utilize high-performance computing.  ...  term for each surrogate model and integrating it out via MCMC sampling.  ... 
arXiv:1811.08687v3 fatcat:yzsduvrojjaajihutzyrcnz5fy

Bayesian model selection in the M-open setting – Approximate posterior inference and probability-proportional-to-size subsampling for efficient large-scale leave-one-out cross-validation [article]

Riko Kelter
2020 arXiv   pre-print
First, we discuss several model views and the available Bayesian model comparison methods in each.  ...  Here, we provide a tutorial on approximate Pareto-smoothed importance sampling leave-one-out cross-validation (PSIS-LOO), a computationally efficient method for Bayesian model comparison.  ...  via the Monte-Carlo average elpd LOO The computational costs of Bayesian model comparison via LOO-CV problematically strongly depend on two quantities as shown in table 1.  ... 
arXiv:2005.13199v1 fatcat:iq3gbcczprdodmzdodnynu5k7e

Improved Bayesian Logistic Supervised Topic Models with Data Augmentation [article]

Jun Zhu, Xun Zheng, Bo Zhang
2013 arXiv   pre-print
Our augment-and-collapse sampling algorithm has analytical forms of each conditional distribution without making any restricting assumptions and can be easily parallelized.  ...  We address these issues by: 1) introducing a regularization constant to better balance the two parts based on an optimization formulation of Bayesian inference; and 2) developing a simple Gibbs sampling  ...  Comparison with MedLDA: The above formulation of logistic supervised topic models as an instance of regularized Bayesian inference provides a direct comparison with the max-margin supervised topic model  ... 
arXiv:1310.2408v1 fatcat:sp5d6fnfvfalbj2m4nnvqxpd7m

MCMC Techniques for Parameter Estimation of ODE Based Models in Systems Biology

Gloria I. Valderrama-Bahamóndez, Holger Fröhlich
2019 Frontiers in Applied Mathematics and Statistics  
The comparison included Metropolis-Hastings, parallel tempering MCMC, adaptive MCMC, and parallel adaptive MCMC.  ...  Here we performed a systematic comparison of different MCMC techniques for this purpose using five public domain models.  ...  Full Bayesian inference can in principle address both aspects, but usually requires computationally costly sampling via MCMC.  ... 
doi:10.3389/fams.2019.00055 fatcat:oni6ecfdfjbpvnhxlil5ksbw4y

Analysis of multiple sclerosis lesions via spatially varying coefficients

Tian Ge, Nicole Müller-Lenke, Kerstin Bendfeldt, Thomas E. Nichols, Timothy D. Johnson
2014 Annals of Applied Statistics  
Here we present a Bayesian spatial model to accurately model binary lesion maps and to determine if there is spatial dependence between lesion location and subject specific covariates such as MS subtype  ...  We apply our model to binary lesion maps derived from T_2-weighted MRI images from 250 multiple sclerosis patients classified into five clinical subtypes, and demonstrate unique modeling and predictive  ...  An efficient leave-one-out cross-validation (LOOCV) approach based on importance sampling [7] was also used to predict the clinical subtype of a held-out subject (with knowledge of their covariates).  ... 
doi:10.1214/14-aoas718 pmid:25431633 pmcid:PMC4243942 fatcat:5gnqdegldvf6bcpukj5vqskium

Semiparametric regression models for spatial prediction and uncertainty quantification of soil attributes

Hunter R. Merrill, Sabine Grunwald, Nikolay Bliznyuk
2016 Stochastic environmental research and risk assessment (Print)  
Bayesian semiparametric models yielded the best predictive results and provided empirical coverage probability nearly equal to nominal.  ...  , based an array of point and interval measures of out-of-sample forecast quality.  ...  In addition to the above models, we also obtained predictions via kriging (Model 7) and Bayesian kriging (Model 8) for comparison with Models 3 and 4, respectively.  ... 
doi:10.1007/s00477-016-1337-0 fatcat:4h3q4u6klzgjll7r66rci3amny

Model of Combined Transport of Perishable Foodstuffs and Safety Inspection Based on Data Mining

Tongjuan Liu, Anqi Hu
2017 Food and Nutrition Sciences  
The relative perfect prediction model can guide the actual sampling work about food quality and safety by prognosticating the occurrence of unqualified food to select the typical and effective samples  ...  The relative optimal prediction model of the perishable food transportation metamorphism monitoring system could be got.  ...  , via track monitoring, predict the metamorphism of perishable food.  ... 
doi:10.4236/fns.2017.87054 fatcat:elm7wsej25gahlzpezf25dnoqq

Bayesian Optimization with Robust Bayesian Neural Networks

Jost Tobias Springenberg, Aaron Klein, Stefan Falkner, Frank Hutter
2016 Neural Information Processing Systems  
We present a general approach for using flexible parametric models (neural networks) for Bayesian optimization, staying as close to a truly Bayesian treatment as possible.  ...  Experiments including multi-task Bayesian optimization with 21 tasks, parallel optimization of deep neural networks and deep reinforcement learning show the power and flexibility of this approach.  ...  To approximate EI we used 50 samples acquired via SGHMC sampling. Maximization of the acquision function was performed via gradient ascent.  ... 
dblp:conf/nips/SpringenbergKFH16 fatcat:vmmf3aodjrgc7nqxx7qua6jgfa

Accelerated Parallel Non-conjugate Sampling for Bayesian Non-parametric Models [article]

Michael Minyi Zhang, Sinead A. Williamson, Fernando Perez-Cruz
2022 arXiv   pre-print
In special cases, where the integration is tractable, we can sample new feature assignments according to a predictive likelihood.  ...  Inference of latent feature models in the Bayesian nonparametric setting is generally difficult, especially in high dimensional settings, because it usually requires proposing features from some prior  ...  Table 1 1 Comparison of final predictive log likelihood results for MNIST, Yale, and CIFAR data with Dirichlet-multinomial model.  ... 
arXiv:1705.07178v7 fatcat:7c33fd37ebcfnjzfklq2uqma5m

Frequentist Approximation of the Bayesian Posterior Inclusion Probability by Stochastic Subsampling

V Ly, E Fokoué
2016 British Journal of Mathematics & Computer Science  
Finally, the scheme proposed is very general and can therefore be easily adapted to all kinds of statistical prediction tasks.  ...  in comparatively comparative computational times thanks to the availability of parallel computing facilities through cloud and cluster computing.  ...  As expected, the out of sample prediction results for the frequentist and Bayesian PIP because we are applying the same variables to the training sets.  ... 
doi:10.9734/bjmcs/2016/27023 fatcat:4miz36mmrve2feer6qgvy27csi

Accurate Machine Learning Atmospheric Retrieval via a Neural Network Surrogate Model for Radiative Transfer [article]

Michael D. Himes, Joseph Harrington, Adam D. Cobb, Atilim Gunes Baydin, Frank Soboczenski, Molly D. O'Beirne, Simone Zorzan, David C. Wright, Zacchaeus Scheffer, Shawn D. Domagal-Goldman, Giada N. Arney
2021 arXiv   pre-print
This inference is computationally expensive, as it requires many executions of a costly radiative transfer (RT) simulation for each set of sampled model parameters.  ...  Here we present a novel approach to retrieval by training a forward ML surrogate model that predicts spectra given model parameters, providing a fast approximate RT simulation that can be used in a conventional  ...  Figure 2 shows example comparisons between the spectra predicted by MARGE and true spectra calculated with transit.  ... 
arXiv:2003.02430v3 fatcat:3higkqzztvflbd2qgwkxzo5axi
« Previous Showing results 1 — 15 out of 48,622 results