Search NASA⌕ Search

SEARCH · Search NASA

Results for “Bayesian method”

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 901 records · Page 50

Thermodynamics and Kinetics of Phase Transformations in Energy Materials (Final Report)

The primary goal of this program is to develop computational tools and methodology for enabling a shift from empirical trial-and-error driven materials discovery towards data and information technology driven design of materials with targeted functional properties. Both thermodynamic and kinetic factors need to be taken into account when designing new materials. Thermodynamics sets the necessary-but-not-sufficient conditions for the suitability of a particular material in a given application, while the kinetics controls the rates of microscopic processes and hence the performance in energy applications (e.g., power in energy storage, conversion efficiencies in solar cells and thermoelectrics, chemical activity in catalysis). The complexity of these phenomena makes computational design difficult, not least due to the lack of suitable theoretical methods and computational tools. To address these needs, we develop systematic, quantitatively accurate first-principles methods to address two important problems: (i) predict and optimize the thermodynamics of materials over a large composition and structure space, and (ii) model kinetic processes such as mass and heat transport in crystalline solids under non-equilibrium conditions. These tasks are addressed by combining high-throughput computation with automatic construction of accurate lattice models using mathematically rigorous methods of information theory, such as compressive sensing and Bayesian inference.

36 MATERIALS SCIENCE↗

We Know Where We Don't Know: 3D Bayesian CNNs for Credible Geometric Uncertainty

Deep learning has been successfully applied to the segmentation of 3D Computed Tomography (CT) scans. Establishing the credibility of these segmentations requires uncertainty quantification (UQ) to identify untrustworthy predictions. Recent UQ architectures include Monte Carlo dropout networks (MCDNs), which approximate deep Gaussian processes, and Bayesian neural networks (BNNs), which learn the distribution of the weight space. BNNs are advantageous over MCDNs for UQ but are thought to be computationally infeasible in high dimension, and neither architecture has produced interpretable geometric uncertainty maps. We propose a novel 3D Bayesian convolutional neural network (BCNN), the first deep learning method which generates statistically credible geometric uncertainty maps and scales for application to 3D data. We present experimental results on CT scans of graphite electrodes and laser-welded metals and show that our BCNN outperforms an MCDN in recent uncertainty metrics. The geometric uncertainty maps generated by our BCNN capture distributions of sigmoid values that are interpretable as confidence intervals, critical for applications that rely on deep learning for high-consequence decisions.

97 MATHEMATICS AND COMPUTING↗

Mapping Stochastic Devices to Probabilistic Algorithms

Probabilistic and Bayesian neural networks have long been proposed as a method to incorporate uncertainty about the world (both in training data and operation) into artificial intelligence applications. One approach to making a neural network probabilistic is to leverage a Monte Carlo sampling approach that samples a trained network while incorporating noise. Such sampling approaches for neural networks have not been extensively studied due to the prohibitive requirement of many computationally expensive samples. While the development of future microelectronics platforms that make this sampling more efficient is an attractive option, it has not been immediately clear how to sample a neural network and what the quality of random number generation should be. This research aimed to start addressing these two fundamental questions by examining basic “off the shelf” neural networks can be sampled through a few different mechanisms (including synapse “dropout” and neuron “dropout”) and examine how these sampling approaches can be evaluated both in terms of evaluating algorithm effectiveness and the required quality of random numbers.

97 MATHEMATICS AND COMPUTING↗

Top Quark Produced Through the Electroweak Force: Discovery Using the Matrix Element Analysis and Search for Heavy Gauge Bosons Using Boosted Decision Trees

The top quark produced through the electroweak channel provides a direct measurement of the V tb element in the CKM matrix which can be viewed as a transition rate of a top quark to a bottom quark. This production channel of top quark is also sensitive to different theories beyond the Standard Model such as heavy charged gauged bosons termed W'. This thesis measures the cross section of the electroweak produced top quark using a technique based on using the matrix elements of the processes under consideration. The technique is applied to 2.3 fb -1 of data from the D0 detector. From a comparison of the matrix element discriminants between data and the signal and background model using Bayesian statistics, we measure the cross section of the top quark produced through the electroweak mechanism σ(p$\bar{p}$ → tb + X, tqb + X) = 4.30 -1.20 +0.98 pb. The measured result corresponds to a 4.9σ Gaussian-equivalent significance. By combining this analysis with other analyses based on the Bayesian Neural Network (BNN) and Boosted Decision Tree (BDT) method, the measured cross section is 3.94 ± 0.88 pb with a significance of 5.0σ, resulting in the discovery of electroweak produced top quarks. Using this measured cross section and constraining |V tb | < 1, the 95% confidence level (C.L.) lower limit is |V tb | > 0.78. Additionally, a search is made for the production of W' using the same samples from the electroweak produced top quark. An analysis based on the BDT method is used to separate the signal from expected backgrounds. No significant excess is found and 95% C.L. upper limits on the production cross section are set for W' with masses within 600-950 GeV. For four general models of W{prime} boson production using decay channel W' → t$\bar{p}$, the lower mass limits are the following: M(W' L with SM couplings) > 840 GeV; M(W' R ) > 880 GeV or 890 GeV if the right-handed neutrino is lighter or heavier than W{prime}{sub R}; and M(W' L+R ) > 915 GeV.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Bayesian optimization algorithms for accelerator physics

Accelerator physics relies on numerical algorithms to solve optimization problems in online accelerator control and tasks such as experimental design and model calibration in simulations. The effectiveness of optimization algorithms in discovering ideal solutions for complex challenges with limited resources often determines the problem complexity these methods can address. The accelerator physics community has recognized the advantages of Bayesian optimization algorithms, which leverage statistical surrogate models of objective functions to effectively address complex optimization challenges, especially in the presence of noise during accelerator operation and in resource-intensive physics simulations. In this review article, we offer a conceptual overview of applying Bayesian optimization techniques toward solving optimization problems in accelerator physics. We begin by providing a straightforward explanation of the essential components that make up Bayesian optimization techniques. We then give an overview of current and previous work applying and modifying these techniques to solve accelerator physics challenges. Finally, we explore practical implementation strategies for Bayesian optimization algorithms to maximize their performance, enabling users to effectively address complex optimization challenges in real-time beam control and accelerator design. Published by the American Physical Society 2024

43 PARTICLE ACCELERATORS↗

Developing a Continuous Nuclear Data Evaluation Method for the Resolved and Unresolved Resonance Regions With FITAPI and SAMMY [Poster]

SAMMY is an R-matrix Bayesian fitting program used in the analysis of nuclear cross section data. SAMMY produces separate cross section models in the resolved and unresolved resonance regions. This project develops a new method of continuous evaluation joining both energy regions with FITAPI. Full energy spectrum model introduces cross covariance between physics models allowing more accurate description of uncertainty. Consistency developed between physics models by the use of a unique set of prior parameters. An unresolved region cross section developed with sensitivity to resolved cross section. Incorporation into ORNL FITAPI provides access to greater number of parameter estimation algorithms.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Variational inference of effective range parameters for 3 He− 4 He scattering

We use two different methods, Monte Carlo sampling and variational inference (VI), to perform a Bayesian calibration of the effective-range parameters in 3 He– 4 He elastic scattering. The parameters are calibrated to data from a recent set of 3 He– 4 He elastic scattering differential cross section measurements. Analysis of these data for E lab ≤ 4.3 MeV yields a unimodal posterior for which both methods obtain the same structure. However, the effective-range expansion amplitude does not account for the 7/2 − state of 7 Be so, even after calibration, the description of data at the upper end of this energy range is poor. The data up to E lab = 2.6 MeV can be well described, but calibration to this lower-energy subset of the data yields a bimodal posterior. After adapting VI to treat such a multi-modal posterior we find good agreement between the VI results and those obtained with parallel-tempered Monte Carlo sampling.

effective field theory↗

Risk assessment of turbine rotor failure using probabilistic ultrasonic non-destructive evaluations

The study presents a method and application of risk assessment methodology for turbine rotor fatigue failure using probabilistic ultrasonic nondestructive evaluations. A rigorous probabilistic modeling for ultrasonic flaw sizing is developed by incorporating the model-assisted probability of detection, and the probability density function (PDF) of the actual flaw size is derived. Two general scenarios, namely the ultrasonic inspection with an identified flaw indication and the ultrasonic inspection without flaw indication, are considered in the derivation. To perform estimations for fatigue reliability and remaining useful life, uncertainties from ultrasonic flaw sizing and fatigue model parameters are systematically included and quantified. The model parameter PDF is estimated using Bayesian parameter estimation and actual fatigue testing data. The overall method is demonstrated using a realistic application of steam turbine rotor, and the risk analysis under given safety criteria is provided to support maintenance planning.

42 ENGINEERING↗

Bayesian sequential optimal experimental design for nonlinear models using policy gradient reinforcement learning

We present a mathematical framework and computational methods for optimally designing a finite sequence of experiments. This sequential optimal experimental design (sOED) problem is formulated as a finite-horizon partially observable Markov decision process (POMDP) under a Bayesian setting and with information-theoretic utilities. The formulation is general and may accommodate continuous random variables, non-Gaussian posteriors, and nonlinear forward models. The sOED design policy incorporates elements of feedback and lookahead simultaneously, and we show it to generalize the commonly-used batch and greedy design strategies. We solve for the sOED policy using the policy gradient (PG) method from reinforcement learning, and provide a derivation for the PG expression in the sOED context. Adopting an actor-critic approach, the policy and value functions are parameterized using deep neural networks and improved via PG estimates produced from simulated episodes of designs and observations. The new PG-sOED algorithm is first validated on a linear-Gaussian benchmark, and then compared against other design baselines on a sensor movement problem for contaminant source inversion in a convection-diffusion field. As a result, we provide explanation for the policy behaviors using knowledge of the underlying physical process.

97 MATHEMATICS AND COMPUTING↗

Quantifying and reducing model-form uncertainties in Reynolds-averaged Navier–Stokes simulations: A data-driven, physics-informed Bayesian approach

Despite their well-known limitations, Reynolds-Averaged Navier–Stokes (RANS) models are still the workhorse tools for turbulent flow simulations in today's engineering analysis, design and optimization. While the predictive capability of RANS models depends on many factors, for many practical flows the turbulence models are by far the largest source of uncertainty. As RANS models are used in the design and safety evaluation of many mission-critical systems such as airplanes and nuclear power plants, quantifying their model-form uncertainties has significant implications in enabling risk-informed decision-making. In this work we develop a data-driven, physics-informed Bayesian framework for quantifying model-form uncertainties in RANS simulations. Uncertainties are introduced directly to the Reynolds stresses and are represented with compact parameterization accounting for empirical prior knowledge and physical constraints (e.g., realizability, smoothness, and symmetry). An iterative ensemble Kalman method is used to assimilate the prior knowledge and observation data in a Bayesian framework, and to propagate them to posterior distributions of velocities and other Quantities of Interest (QoIs). We use two representative cases, the flow over periodic hills and the flow in a square duct, to evaluate the performance of the proposed framework. Both cases are challenging for standard RANS turbulence models. Simulation results suggest that, even with very sparse observations, the obtained posterior mean velocities and other QoIs have significantly better agreement with the benchmark data compared to the baseline results. At most locations the posterior distribution adequately captures the true model error within the developed model form uncertainty bounds. The framework is a major improvement over existing black-box, physics-neutral methods for model-form uncertainty quantification, where prior knowledge and details of the models are not exploited. This approach has potential implications in many fields in which the governing equations are well understood but the model uncertainty comes from unresolved physical processes. - Highlights: • Proposed a physics–informed framework to quantify uncertainty in RANS simulations. • Framework incorporates physical prior knowledge and observation data. • Based on a rigorous Bayesian framework yet fully utilizes physical model. • Applicable for many complex physical systems beyond turbulent flows.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Automated integration gate selection for Gaussian mixture model pulse shape discrimination

Pulse shapes differ between neutron and gamma particles when measured with detector devices employing pulse shape discriminating (PSD) scintillators. Digitized waveforms can be used in detection systems to perform pulse shape discrimination for this application. Prior Gaussian Mixture Model (GMM) methods require access to the pulse full-waveform. Reducing the waveform to a smaller set of combined samples reduces computational cost while affecting PSD performance. In this work, we develop a method for selecting the best performing combination of integration gates, or contiguous summed segments of the digitized pulse for PSD. The method uses a discrimination score based on the GMM PSD approach. Furthermore, the final selection is performed using Bayesian Optimization. PSD detection results are compared with varying numbers of selected gates on time-of-flight (TOF) data. This method can be used to fully automate the selection of gates in an unsupervised (without ground truth labels) setting.

42 ENGINEERING↗

RECALIBRATION OF THE VIRIAL FACTOR AND M{sub BH}-{sigma}{sub *} RELATION FOR LOCAL ACTIVE GALAXIES

Determining the virial factor of the broad-line region gas is crucial for calibrating active galactic nucleus black hole mass estimators, since the measured line-of-sight velocity needs to be converted into the intrinsic virial velocity. The average virial factor has been empirically calibrated based on the M{sub BH}-{sigma}{sub *} relation of quiescent galaxies, but the claimed values differ by a factor of two in recent studies. We investigate the origin of the difference by measuring the M{sub BH}-{sigma}{sub *} relation using an updated galaxy sample from the literature and explore the dependence of the virial factor on various fitting methods. We find that the discrepancy is primarily caused by the sample selection, while the difference stemming from the various regression methods is marginal. However, we generally prefer the FITEXY and Bayesian estimators based on Monte Carlo simulations for the M{sub BH}-{sigma}{sub *} relation. In addition, the choice of independent variable in the regression leads to {approx}0.2 dex variation in the virial factor inferred from the calibration process. Based on the determined virial factor, we present the updated M{sub BH}-{sigma}{sub *} relation of local active galaxies.

79 ASTRONOMY AND ASTROPHYSICS↗

Bayesian and variational Bayesian approaches for flows in heterogeneous random media

In this paper, we study porous media flows in heterogeneous stochastic media. We propose an efficient forward simulation technique that is tailored for variational Bayesian inversion. As a starting point, the proposed forward simulation technique decomposes the solution into the sum of separable functions (with respect to randomness and the space), where each term is calculated based on a variational approach. This is similar to Proper Generalized Decomposition (PGD). Next, we apply a multiscale technique to solve for each term (as in ) and, further, decompose the random function into 1D fields. As a result, our proposed method provides an approximation hierarchy for the solution as we increase the number of terms in the expansion and, also, increase the spatial resolution of each term. We use the hierarchical solution distributions in a variational Bayesian approximation to perform uncertainty quantification in the inverse problem. We conduct a detailed numerical study to explore the performance of the proposed uncertainty quantification technique and show the theoretical posterior concentration.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A statistical method for estimating wood thermal diffusivity and probe geometry using in situ heat response curves from sap flow measurements

The heat pulse method is widely used to measure water flux through plants; it works by inferring the velocity of water through a porous medium from the speed at which a heat pulse is propagated through the system. No systematic, non-destructive calibration procedure exists to determine the site-specific parameters necessary for calculating sap velocity, e.g., wood thermal diffusivity and probe spacing. Such parameter calibration is crucial to obtain the correct transpiration flux density from the sap flow measurements at the plant scale; and consequently, to up-scale tree-level water fluxes to canopy and landscape scales. The purpose of this study is to present a statistical framework for estimating the wood thermal diffusivity and probe spacing simutaneously from in-situ heat response curves collected by the implanted probes of a heat ratio apparatus. Conditioned on the time traces of wood temperature following a heat pulse, the parameters are inferred using a Bayesian inversion technique, based on the Markov chain Monte Carlo sampling method. The primary advantage of the proposed methodology is that it does not require known probe spacing or any further intrusive sampling of sapwood. The Bayesian framework also enables direct quantification of uncertainty in estimated sap flow velocity. Experiments using synthetic data show that repeated tests using the same apparatus are essential to obtain reliable and accurate solutions. When applied to field conditions, these tests are conducted during different seasons and automated using the existing data logging system. The seasonality of wood thermal diffusivity is obtained as a by-product of the parameter estimation process, and it is shown to be affected by both moisture content and temperature. Empirical factors are often introduced to account for the influence of non-ideal probe geometry on the estimation of heat pulse velocity, and they are estimated in this study as well. The proposed methodology can be applied for the calibration of existing heat ratio sap flow systems at other sites. It is especially useful when an alternative transpiration calibration device, such as a lysimeter, is not available.

Bayesian inversion↗

Uncertainty quantification in multivariable regression for material property prediction with Bayesian neural networks

With the increased use of data-driven approaches and machine learning-based methods in material science, the importance of reliable uncertainty quantification (UQ) of the predicted variables for informed decision-making cannot be overstated. UQ in material property prediction poses unique challenges, including multi-scale and multi-physics nature of materials, intricate interactions between numerous factors, limited availability of large curated datasets, etc. In this work, we introduce a physics-informed Bayesian Neural Networks (BNNs) approach for UQ, which integrates knowledge from governing laws in materials to guide the models toward physically consistent predictions. To evaluate the approach, we present case studies for predicting the creep rupture life of steel alloys. Experimental validation with three datasets of creep tests demonstrates that this method produces point predictions and uncertainty estimations that are competitive or exceed the performance of conventional UQ methods such as Gaussian Process Regression. Additionally, we evaluate the suitability of employing UQ in an active learning scenario and report competitive performance. The most promising framework for creep life prediction is BNNs based on Markov Chain Monte Carlo approximation of the posterior distribution of network parameters, as it provided more reliable results in comparison to BNNs based on variational inference approximation or related NNs with probabilistic outputs.

36 MATERIALS SCIENCE↗

Bayesian Radiation Source Localization

Locating illicit radiological sources using gamma ray or neutron detection is a key challenge for both homeland security and nuclear nonproliferation. Localization methods using an array of detectors or a sequence of observations in time and space must provide rapid results while accounting for a dynamic attenuating environment. In the presence of significant attenuation and scatter, more extensive numerical transport calculations in place of the standard analytical approximations may be required to achieve accurate results. Numerical adjoints based on deterministic transport codes provide relatively efficient detector response calculations needed to determine the most likely location of a true source. Probabilistic representations account for uncertainty in the source location resulting from uncertainties in detector responses, the approximations that are used, and the potential for nonunique solutions. A Bayesian approach improves on previous likelihood methods for source localization by allowing the incorporation of all available information to help constrain solutions.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

Transverse mode analysis for free-space laser beams using Bayesian analysis

Reliable and rapid assessment of the transverse mode quality of a free-space laser beam has a wide range of applications in laser development, research, and utilization. It has become even more important with recent advances in developing orbital angular momentum photon beams across a broad spectral region. In this work, a general modal analysis method for a free-space multimode laser beam has been developed based on Bayesian analysis. After transforming mode decomposition into a linear system problem, a Gaussian probabilistic model is used to find a closed-form solution. The method is found to be robust with the presence of Gaussian noise. Prior knowledge about the mode content can be incorporated into the method to improve the solution for situations when coherent disturbances or contamination are present in the laser beam. This method can be used to analyze the mode content for laser beams in different bases, such as Hermite–Gaussian (HG) modes and Laguerre–Gaussian (LG) modes. Three applications of this method are presented: a detailed modal analysis of the beam image from the incoherent intensity addition of HG modes and two examples of mode decomposition using the complex wavefront from the coherent superposition of HG and LG modes. The feasibility of this method is demonstrated using various simulation results. Based on digital images of a laser beam recorded without complex wavelength-limiting optics, in principle, this method can be used in a wide wavelength range from infrared to ultraviolet, and possibly x ray.

Liu, Peifan (ORCID:0000000273816153)↗