Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Bayesian Inference”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 937 records · Page 52

An analysis pipeline for the inference of protein-protein interaction networks

An analysis pipeline has been created for deployment of a novel algorithm, the Bayesian Estimator of Protein-Protein Association Probabilities (BEPro), for use in the reconstruction of protein-protein interaction networks. We have combined the Software Environment for BIological Network Inference (SEBINI), an interactive environment for the deployment and testing of network inference algorithms that use high-throughput data, and the Collective Analysis of Biological Interaction Networks (CABIN), software that allows integration and analysis of protein-protein interaction and gene-to-gene regulatory evidence obtained from multiple sources, to allow interactions computed by BEPro to be stored, visualized, and further analyzed. Incorporating BEPro into SEBINI and automatically feeding the resulting inferred network into CABIN, we have created a structured workflow for protein-protein network inference and supplemental analysis from sets of mass spectrometry bait-prey experiment data. SEBINI demo site: https://www.emsl.pnl.gov /SEBINI/ Contact: ronald.taylor@pnl.gov. BEPro is available at http://www.pnl.gov/statistics/BEPro3/index.htm. Contact: ds.daly@pnl.gov. CABIN is available at http://www.sysbio.org/dataresources/cabin.stm. Contact: mudita.singhal@pnl.gov.

59 BASIC BIOLOGICAL SCIENCES↗

Quantification of neural networks uncertainties with applications to SAFARI-1 axial neutron flux profiles

Deep Neural Networks (DNNs) have been widely used as a data-driven modelling tool in nuclear engineering. However, as a Machine Learning model, Artificial Neural Network (ANN) predictions are subjected to uncertainties originating from the noise in training data, incomplete coverage of the domain, and imperfect neural network architectures. In this work, we target at quantifying the prediction/approximation uncertainties of ANNs using Monte Carlo Dropout (MCD), as well as Bayesian Neural Networks (BNNs) which are solved by variational inference. With a demonstration problem in which neural networks are used to predict the assembly axial neutron flux profiles, the results have shown that the three different neural network models (regular DNNs, DNNs solved with MCD and BNNs) can produce results that agree very well among each other and with the measurement data, on cycles that are not used in the training process. Besides the excellent generalization capability, the uncertainty bands produced by MCD and BNN agree very well, and in general, they can fully envelop the noisy measurement data points. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Learning from crowds with variational Gaussian processes

Solving a supervised learning problem requires to label a training set. This task is traditionally performed by an expert, who provides a label for each sample. The proliferation of social web services (e.g., Amazon Mechanical Turk) has introduced an alternative crowdsourcing approach. Anybody with a computer can register in one of these services and label, either partially or completely, a dataset. The effort of labeling is then shared between a great number of annotators. However, this approach introduces scientifically challenging problems such as combining the unknown expertise of the annotators, handling disagreements on the annotated samples, or detecting the existence of spammer and adversarial annotators. All these problems require probabilistic sound solutions which go beyond the naive use of majority voting plus classical classification methods. In this work we introduce a new crowdsourcing model and inference procedure which trains a Gaussian Process classifier using the noisy labels provided by the annotators. Variational Bayes inference is used to estimate all unknowns. The proposed model can predict the class of new samples and assess the expertise of the involved annotators. Moreover, the Bayesian treatment allows for a solid uncertainty quantification. Since when predicting the class of a new sample we might have access to some annotations for it, we also show how our method can naturally incorporate this additional information. Furthermore, a comprehensive experimental section evaluates the proposed method with synthetic and real experiments, showing that it consistently outperforms other state-of-the-art crowdsourcing approaches.

97 MATHEMATICS AND COMPUTING↗

Adaptive sampling for accelerating neutron diffraction-based strain mapping *

Abstract Neutron diffraction is a useful technique for mapping residual strains in dense metal objects. The technique works by placing an object in the path of a neutron beam, measuring the diffracted signals and inferring the local lattice strain values from the measurement. In order to map the strains across the entire object, the object is stepped one position at a time in the path of the neutron beam, typically in raster order, and at each position a strain value is estimated. Typical dwell times at neutron diffraction instruments result in an overall measurement that can take several hours to map an object that is several tens of centimeters in each dimension at a resolution of a few millimeters, during which the end users do not have an estimate of the global strain features and are at risk of incomplete information in case of instruments outages. In this paper, we propose an object adaptive sampling strategy to measure the significant points first. We start with a small initial uniform set of measurement points across the object to be mapped, compute the strain in those positions and use a machine learning technique to predict the next position to measure in the object. Specifically, we use a Bayesian optimization based on a Gaussian process regression method to infer the underlying strain field from a sparse set of measurements and predict the next most informative positions to measure based on estimates of the mean and variance in the strain fields estimated from the previously measured points. We demonstrate our real-time measure-infer-predict workflow on additively manufactured steel parts—demonstrating that we can get an accurate strain estimate even with 30%–40% of the typical number of measurements—leading the path to faster strain mapping with useful real-time feedback. We emphasize that the proposed method is general and can be used for fast mapping of other material properties such as phase fractions from time-consuming point-wise neutron measurements.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Uncertainty quantification of material parameters in modeling coupled metal and high explosive experiments

Experiments involving the coupling of metal and high explosives (HE) are of notable defense-related interest, and we seek to refine the uncertainty quantification associated with models of such experiments. In particular, our focus is on how uncertainty related to the metal constitutive model challenges our ability to infer high explosive model parameters when analyzing focused science experiments. We consider three focused experiments involving an HE accelerating metal: small plate tests with tantalum/LX-14 and tantalum/LX-17 pairings as well as a tantalum/LX-17 cylinder test. For all three models, we perform sensitivity analysis to ascertain the influence of metal strength on the coupled experimental response. Moreover, we calibrate each model in a Bayesian setting and study the quantification of metal strength on the inference of the HE parameters. Based on our results, we offer guidance for future metal/HE experiments.

36 MATERIALS SCIENCE↗

Dealing with Measurement Uncertainties as Nuisance Parameters in Bayesian Model Calibration

In the presence of model discrepancy, the calibration of physics-based models for physical parameter inference is a challenging problem. Lack of identifiability between calibration parameters and model discrepancy requires additional identifiability constraints to be placed on the model discrepancy to obtain unique physical parameter estimates. If these assumptions are violated, the inference for the calibration parameters can be systematically biased. In many applications, such as in dynamic material property experiments, many of the calibration inputs refer to measurement uncertainties. Here, we develop a metric for identifying overfitting of these measurement uncertainties, propose a prior capable of reducing this overfitting, and show how this leads to a diagnostic tool for validation of physical parameter inference. The approach is demonstrated for a benchmark example and applied for a material property application to perform inference on the equation of state parameters of tantalum.

97 MATHEMATICS AND COMPUTING↗

Bayesian approach for three-dimensional aquifer characterization at the Hanford 300 Area

This study presents a stochastic, three-dimensional characterization of a heterogeneous hydraulic conductivity field within DOE's Hanford 300 Area site, Washington, by assimilating large-scale, constant-rate injection test data with small-scale, three-dimensional electromagnetic borehole flowmeter (EBF) measurement data. We first inverted the injection test data to estimate the transmissivity field, using zeroth-order temporal moments of pressure buildup curves. We applied a newly developed Bayesian geostatistical inversion framework, the method of anchored distributions (MAD), to obtain a joint posterior distribution of geostatistical parameters and local log-transmissivities at multiple locations. The unique aspects of MAD that make it suitable for this purpose are its ability to integrate multi-scale, multi-type data within a Bayesian framework and to compute a nonparametric posterior distribution. After we combined the distribution of transmissivities with depth-discrete relative-conductivity profile from EBF data, we inferred the three-dimensional geostatistical parameters of the log-conductivity field, using the Bayesian model-based geostatistics. Such consistent use of the Bayesian approach throughout the procedure enabled us to systematically incorporate data uncertainty into the final posterior distribution. The method was tested in a synthetic study and validated using the actual data that was not part of the estimation. Results showed broader and skewed posterior distributions of geostatistical parameters except for the mean, which suggests the importance of inferring the entire distribution to quantify the parameter uncertainty.

Murakami, Haruko↗

Fast Bayesian blind deconvolution with Huber Super Gaussian priors

We report expectation Maximization (EM) based inference has already proven to be a very powerful tool to solve blind image deconvolution (BID) problems. Unfortunately, three important problems still impede the application of EM in BID: the undesirable saddle points and local minima caused by highly nonconvex priors, the instability around zero of some of the most interesting sparsity promoting priors, and the intrinsic high computational cost of the corresponding BID algorithm. In this paper we first show how Super Gaussian priors can be made numerically tractable around zero by introducing the family of Huber Super Gaussian priors and then present a fast EM based blind deconvolution method formulated in the image space. In the proposed computational approach, image and kernel estimation are performed by using the Alternating Direction Method of Multipliers (ADMM), which allows to exploit the advantages of FFT computation. For highly nonconvex priors, we propose a Smooth ADMM (SADMM) approach to avoid poor BID estimates. In conclusion, extensive experiments demonstrate that the proposed method significantly outperforms state-of-the-art BID methods in terms of quality of the reconstructions and speed.

97 MATHEMATICS AND COMPUTING↗

Quantitative Uncertainty Analysis for Diffraction

Current software uses a least squares statistical method. For each model parameter, current software yields a point estimate and estimated standard error for each point estimate. In contrast, we have employed a Bayesian approach with sampling algorithm to provide a full posterior probability density function on each model parameter. This yields a different type of solution (probabilistic) with better fidelity. The Quantitative Uncertainty Analysis for Diffraction (QUAD) package returns a matrix of affect model parameters as a function of iteration through the MCMC analysis. QUAD is a Python code that allows for analysis of X-ray and neutron diffraction data to infer the structure of materials with quantifiable certainty. The code uses Bayesian statistics and Markov chain sampling algorithms to create posterior probability distributions on all material structure parameters that are modeled by a researcher. The statistical packages use the open source GSASII project as a library to calculate model diffraction patterns.

Fancher, ChristopherM↗

Towards Compact Neural Networks via End-to-End Training: A Bayesian Tensor Approach with Automatic Rank Determination

Post-training model compression can reduce the inference costs of deep neural networks, but uncompressed training still consumes enormous hardware resources and energy. To enable low-energy training on edge devices, it is highly desirable to directly train a compact neural network from scratch with a low memory cost. Low-rank tensor decomposition is an effective approach to reduce the memory and computing costs of large neural networks. However, directly training low-rank tensorized neural networks is a very challenging task because it is hard to determine a proper tensor rank a priori, and the tensor rank controls both model complexity and accuracy. Here, this paper presents a novel end-to-end framework for low-rank tensorized training. We first develop a Bayesian model that supports various low-rank tensor formats (e.g., CANDECOMP/PARAFAC, Tucker, tensor-train, and tensor-train matrix) and reduces neural network parameters with automatic rank determination during training. Then we develop a customized Bayesian solver to train large-scale tensorized neural networks. Our training methods shows orders-of-magnitude parameter reduction and little accuracy loss (or even better accuracy) in the experiments. On a very large deep learning recommendation system with over 4.2 ×10 9 model parameters, our method can reduce the parameter number to 1.6 ×10 5 automatically in the training process (i.e., by 2.6 ×10 4 times) while achieving almost the same accuracy. Code is available at https://github.com/colehawkins/bayesian-tensor-rank-determination.

compact neural networks↗

Bayesian model-data comparison incorporating theoretical uncertainties

Accurate comparisons between theoretical models and experimental data are critical for scientific progress. However, inferred physical model parameters can vary significantly with the chosen physics model, highlighting the importance of properly accounting for theoretical uncertainties. In this Letter, we present a Bayesian framework that explicitly quantifies these uncertainties by statistically modeling theory errors, guided by qualitative knowledge of a theory’s varying reliability across the input domain. We demonstrate the effectiveness of this approach using two systems: a simple ball drop experiment and multi-stage heavy-ion simulations. In both cases incorporating model discrepancy leads to improved parameter estimates, with systematic improvements observed as additional experimental observables are integrated.

Bayesian methods↗

Atmospheric inversion of surface carbon flux with consideration of the spatial distribution of US crop production and consumption

In order to improve quantification of the spatial distribution of carbon sinks and sources in the conterminous US, we conduct a nested global atmospheric inversion with detailed spatial information on crop production and consumption. County-level cropland net primary productivity, harvested biomass, soil carbon change, and human and livestock consumption data over the conterminous US are used for this purpose. Time-dependent Bayesian synthesis inversions are conducted based on CO₂ observations at 210 stations to infer CO₂ fluxes globally at monthly time steps with a nested focus on 30 regions in North America. Prior land surface carbon fluxes are first generated using a biospheric model, and the inversions are constrained using prior fluxes with and without adjustments for crop production and consumption over the 2002–2007 period. After these adjustments, the inverted regional carbon sink in the US Midwest increases from 0.25 ± 0.03 to 0.42 ± 0.13 Pg C yr⁻¹, whereas the large sink in the US southeast forest region is weakened from 0.41 ± 0.12 to 0.29 ± 0.12 Pg C yr⁻¹. These adjustments also reduce the inverted sink in the west region from 0.066 ± 0.04 to 0.040 ± 0.02 Pg C yr⁻¹ because of high crop consumption and respiration by humans and livestock. The general pattern of sink increases in crop production areas and sink decreases (or source increases) in crop consumption areas highlights the importance of considering the lateral carbon transfer in crop products in atmospheric inverse modeling, which provides a reliable atmospheric perspective of the overall carbon balance at the continental scale but is unreliable for separating fluxes from different ecosystems.

54 ENVIRONMENTAL SCIENCES↗

Visible core spectroscopy at Wendelstein 7-X

This paper presents an overview of recent hardware extensions and data analysis developments to the Wendelstein 7-X visible core spectroscopy systems. These include upgrades to prepare the in-vessel components for long-pulse operation, nine additional spectrometers, a new line of sight array for passive spectroscopy, and a coherence imaging charge exchange spectroscopy diagnostic. Progress in data analysis includes ion temperatures and densities from multiple impurity species, a statistical comparison with x-ray crystal spectrometer measurements, neutral density measurements from thermal passive Balmer-alpha emission, and a Bayesian analysis of active hydrogen emission, which is able to infer electron density and main ion temperature profiles.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

PHOTOMETRIC SUPERNOVA COSMOLOGY WITH BEAMS AND SDSS-II

Supernova (SN) cosmology without spectroscopic confirmation is an exciting new frontier, which we address here with the Bayesian Estimation Applied to Multiple Species (BEAMS) algorithm and the full three years of data from the Sloan Digital Sky Survey II Supernova Survey (SDSS-II SN). BEAMS is a Bayesian framework for using data from multiple species in statistical inference when one has the probability that each data point belongs to a given species, corresponding in this context to different types of SNe with their probabilities derived from their multi-band light curves. We run the BEAMS algorithm on both Gaussian and more realistic SNANA simulations with of order 10{sup 4} SNe, testing the algorithm against various pitfalls one might expect in the new and somewhat uncharted territory of photometric SN cosmology. We compare the performance of BEAMS to that of both mock spectroscopic surveys and photometric samples that have been cut using typical selection criteria. The latter typically either are biased due to contamination or have significantly larger contours in the cosmological parameters due to small data sets. We then apply BEAMS to the 792 SDSS-II photometric SNe with host spectroscopic redshifts. In this case, BEAMS reduces the area of the {Omega}{sub m}, {Omega}{sub {Lambda}} contours by a factor of three relative to the case where only spectroscopically confirmed data are used (297 SNe). In the case of flatness, the constraints obtained on the matter density applying BEAMS to the photometric SDSS-II data are {Omega}{sup BEAMS}{sub m} = 0.194 {+-} 0.07. This illustrates the potential power of BEAMS for future large photometric SN surveys such as Large Synoptic Survey Telescope.

79 ASTRONOMY AND ASTROPHYSICS↗

Phylogenetic diversity, trichothecene potential, and pathogenicity within Fusarium sambucinum species complex

The Fusarium sambucinum species complex (FSAMSC) is one of the most taxonomically challenging groups of fusaria, comprising prominent mycotoxigenic plant pathogens and other species with various lifestyles. Among toxins produced by members of the FSAMSC, trichothecenes pose the most significant threat to public health. Herein a global collection of 171 strains, originating from diverse hosts or substrates, were selected to represent FSAMSC diversity. This strain collection was used to assess their species diversity, evaluate their potential to produce trichothecenes, and cause disease on wheat. Maximum likelihood and Bayesian analyses of a combined 3-gene dataset used to infer evolutionary relationships revealed that the 171 strains originally received as 48 species represent 74 genealogically exclusive phylogenetically distinct species distributed among six strongly supported clades: Brachygibbosum, Graminearum, Longipes, Novel, Sambucinum, and Sporotrichioides. Most of the strains produced trichothecenes in vitro but varied in type, indicating that the six clades correspond to type A, type B, or both types of trichothecene-producing lineages. Furthermore, five strains representing two putative novel species within the Sambucinum Clade produced two newly discovered type A trichothecenes, 15-keto NX-2 and 15-keto NX-3. Strains of the two putatively novel species together with members of the Graminearum Clade were aggressive toward wheat when tested for pathogenicity on heads of the susceptible cultivar Apogee. In planta, the Graminearum Clade strains produced nivalenol or deoxynivalenol and the aggressive Sambucinum Clade strains synthesized NX-3 and 15-keto NX-3. Other strains within the Brachygibbosum, Longipes, Novel, Sambucinum, and Sporotrichioides Clades were nonpathogenic or could infect the inoculated floret without spreading within the head. Moreover, most of these strains did not produce any toxin in the inoculated spikelets. These data highlight aggressiveness toward wheat appears to be influenced by the type of toxin produced and that it is not limited to members of the Graminearum Clade.

59 BASIC BIOLOGICAL SCIENCES↗

CONCURRENT, CONDENSED STEIN VARIATIONAL GRADIENT DESCENT FOR UNCERTAINTY QUANTIFICATION OF NEURAL NETWORKS

In this work, we propose a Stein variational gradient descent (SVGD) method to concurrently sparsify, train, and provide uncertainty quantification (UQ) of a complexly parameterized model, such as a neural network (NN). It employs a graph reconciliation and condensation process to reduce complexity and increase similarity in the Stein ensemble of parameterizations. Therefore, the proposed concurrent, condensed SVGD (ccSVGD) method can provide UQ on parameters, not just outputs. Furthermore, the parameter reduction speeds up the convergence of the Stein gradient descent as it reduces the combinatorial complexity by aligning and differentiating the sensitivity to parameters. These properties are demonstrated with an illustrative example and an application to a mechanical response representation problem in solid mechanics.

42 ENGINEERING↗

AGNfitter: a Bayesian MCMC approach to fitting spectral energy distributions of AGNs

We present AGNfitter, a publicly available open-source algorithm implementing a fully Bayesian Markov Chain Monte Carlo method to fit the spectral energy distributions (SEDs) of active galactic nuclei (AGNs) from the sub-millimeter to the UV, allowing one to robustly disentangle the physical processes responsible for their emission. AGNfitter makes use of a large library of theoretical, empirical, and semi-empirical models to characterize both the nuclear and host galaxy emission simultaneously. The model consists of four physical emission components: an accretion disk, a torus of AGN heated dust, stellar populations, and cold dust in star-forming regions. AGNfitter determines the posterior distributions of numerous parameters that govern the physics of AGNs with a fully Bayesian treatment of errors and parameter degeneracies, allowing one to infer integrated luminosities, dust attenuation parameters, stellar masses, and star-formation rates. We tested AGNfitter’s performance on real data by fitting the SEDs of a sample of 714 X-ray selected AGNs from the XMM-COSMOS survey, spectroscopically classified as Type1 (unobscured) and Type2 (obscured) AGNs by their optical–UV emission lines. We find that two independent model parameters, namely the reddening of the accretion disk and the column density of the dusty torus, are good proxies for AGN obscuration, allowing us to develop a strategy for classifying AGNs as Type1 or Type2, based solely on an SED-fitting analysis. Our classification scheme is in excellent agreement with the spectroscopic classification, giving a completeness fraction of ∼86% and ∼70%, and an efficiency of ∼80% and ∼77%, for Type1 and Type2 AGNs, respectively.

79 ASTRONOMY AND ASTROPHYSICS↗

Bayesian Model Calibration for Extrapolative Prediction via Gibbs Posteriors

The current standard Bayesian approach to model calibration, which assigns a Gaussian process prior to the discrepancy term, often suffers from issues of unidentifiability and computational complexity and instability. When the goal is to quantify uncertainty in physical parameters for extrapolative prediction, then there is no need to perform inference on the discrepancy term. With this in mind, we introduce Gibbs posteriors as an alternative Bayesian method for model calibration, which updates the prior with a loss function connecting the data to the parameter. The target of inference is the physical parameter value which minimizes the expected loss. We propose to tune the loss scale of the Gibbs posterior to maintain nominal frequentist coverage under assumptions of the form of model discrepancy, and present a bootstrap implementation for approximating coverage rates. Our approach is highly modular, allowing an analyst to easily encode a wide variety of such assumptions. Furthermore, we provide a principled method of combining posteriors calculated from data subsets. We apply our methods to data from an experiment measuring the material properties of tantalum.

36 MATERIALS SCIENCE↗