Search NASA⌕ Search

SEARCH · Search NASA

Results for “Bayesian Framework”

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 829 records · Page 46

Past and present cosmic structure in the SDSS DR7 main sample

We present a chrono-cosmography project, aiming at the inference of the four dimensional formation history of the observed large scale structure from its origin to the present epoch. To do so, we perform a full-scale Bayesian analysis of the northern galactic cap of the Sloan Digital Sky Survey (SDSS) Data Release 7 main galaxy sample, relying on a fully probabilistic, physical model of the non-linearly evolved density field. Besides inferring initial conditions from observations, our methodology naturally and accurately reconstructs non-linear features at the present epoch, such as walls and filaments, corresponding to high-order correlation functions generated by late-time structure formation. Our inference framework self-consistently accounts for typical observational systematic and statistical uncertainties such as noise, survey geometry and selection effects. We further account for luminosity dependent galaxy biases and automatic noise calibration within a fully Bayesian approach. As a result, this analysis provides highly-detailed and accurate reconstructions of the present density field on scales larger than ∼ 3 Mpc/h, constrained by SDSS observations. This approach also leads to the first quantitative inference of plausible formation histories of the dynamic large scale structure underlying the observed galaxy distribution. The results described in this work constitute the first full Bayesian non-linear analysis of the cosmic large scale structure with the demonstrated capability of uncertainty quantification. Some of these results will be made publicly available along with this work. The level of detail of inferred results and the high degree of control on observational uncertainties pave the path towards high precision chrono-cosmography, the subject of simultaneously studying the dynamics and the morphology of the inhomogeneous Universe.

79 ASTRONOMY AND ASTROPHYSICS↗

A score-based diffusion model approach for adaptive learning of stochastic partial differential equation solutions

In this paper, we propose a novel framework for adaptively learning the time-evolving solutions of stochastic partial differential equations (SPDEs) using score-based diffusion models within a recursive Bayesian inference setting. SPDEs play a central role in modeling complex physical systems under uncertainty, but their numerical solutions often suffer from model errors and reduced accuracy due to incomplete physical knowledge and environmental variability. To address these challenges, we encode the governing physics into the score function of a diffusion model using simulation data and incorporate observational information via a likelihood-based correction in a reverse-time stochastic differential equation. This enables adaptive learning through iterative refinement of the solution as new data becomes available. To improve computational efficiency in high-dimensional settings, we introduce the ensemble score filter, a training-free approximation of the score function designed for real-time inference. Numerical experiments on benchmark SPDEs demonstrate the accuracy and robustness of the proposed method under sparse and noisy observations.

97 MATHEMATICS AND COMPUTING↗

Tipping point analysis of atmospheric oxygen concentration

We apply tipping point analysis to nine observational oxygen concentration records around the globe, analyse their dynamics and perform projections under possible future scenarios, leading to oxygen deficiency in the atmosphere. The analysis is based on statistical physics framework with stochastic modelling, where we represent the observed data as a composition of deterministic and stochastic components estimated from the observed data using Bayesian and wavelet techniques.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

A novel CMB component separation method: hierarchical generalized morphological component analysis

ABSTRACT We present a novel technique for cosmic microwave background (CMB) foreground subtraction based on the framework of blind source separation. Inspired by previous work incorporating local variation to generalized morphological component analysis (GMCA), we introduce hierarchical GMCA (HGMCA), a Bayesian hierarchical graphical model for source separation. We test our method on Nside = 256 simulated sky maps that include dust, synchrotron, free–free, and anomalous microwave emission, and show that HGMCA reduces foreground contamination by $25{{\ \rm per\ cent}}$ over GMCA in both the regions included and excluded by the Planck UT78 mask, decreases the error in the measurement of the CMB temperature power spectrum to the 0.02–0.03 per cent level at ℓ > 200 (and $\lt 0.26{{\ \rm per\ cent}}$ for all ℓ), and reduces correlation to all the foregrounds. We find equivalent or improved performance when compared to state-of-the-art internal linear combination type algorithms on these simulations, suggesting that HGMCA may be a competitive alternative to foreground separation techniques previously applied to observed CMB data. Additionally, we show that our performance does not suffer when we perturb model parameters or alter the CMB realization, which suggests that our algorithm generalizes well beyond our simplified simulations. Our results open a new avenue for constructing CMB maps through Bayesian hierarchical analysis.

79 ASTRONOMY AND ASTROPHYSICS↗

LISA extreme-mass-ratio inspiral events as probes of the black hole mass function

One of the sources of gravitational waves for the proposed space-based gravitational wave detector, the Laser Interferometer Space Antenna (LISA), are the inspirals of compact objects into supermassive black holes in the centers of galaxies--extreme-mass-ratio inspirals (EMRIs). Using LISA observations, we will be able to measure the parameters of each EMRI system detected to very high precision. However, the statistics of the set of EMRI events observed by LISA will be more important in constraining astrophysical models than extremely precise measurements for individual systems. The black holes to which LISA is most sensitive are in a mass range that is difficult to probe using other techniques, so LISA provides an almost unique window onto these objects. In this paper we explore, using Bayesian techniques, the constraints that LISA EMRI observations can place on the mass function of black holes at low redshift. We describe a general framework for approaching inference of this type--using multiple observations in combination to constrain a parametrized source population. Assuming that the scaling of the EMRI rate with the black-hole mass is known and taking a black-hole distribution given by a simple power law, dn/dlnM=A{sub 0}(M/M{sub *}){sup {alpha}}{sub 0}, we find that LISA could measure the parameters to a precision of {Delta}(lnA{sub 0}){approx}0.08, and {Delta}({alpha}{sub 0}){approx}0.03 for a reference model that predicts {approx}1000 events. Even with as few as 10 events, LISA should constrain the slope to a precision {approx}0.3, which is the current level of observational uncertainty in the low-mass slope of the black-hole mass function. We also consider a model in which A{sub 0} and {alpha}{sub 0} evolve with redshift, but find that EMRI observations alone do not have much power to probe such an evolution.

79 ASTRONOMY AND ASTROPHYSICS↗

A Singular Value Decomposition Framework for Retrievals with Vertical Distribution Information from Greenhouse Gas Column Absorption Spectroscopy Measurements

We review the singular value decomposition (SVD) framework and use it for quantifying and discerning vertical information in greenhouse gas retrievals from column integrated absorption measurements. While the commonly used traditional Bayesian optimal estimation (OE) assumes a prior distribution in order to regularize the inversion problem, the SVD approach identifies principal components that can be retrieved from the measurement without explicitly specifying a prior mean and prior covariance matrix. We review the SVD method, explicitly recognize the use of an uninformative prior and show it to incur no bias from the choice of the prior. We also make the connection between the SVD method and the pseudo-inverse, which makes it more intuitive and easy to understand. We illustrate the use of the SVD method on an integrated path differential absorption CO2 lidar measurement model and verify our derivations and bias-free properties versus optimal estimation using numerical simulations. In contrast, traditional OE retrievals exhibit bias when the prior mean used in the retrieval differs from the true mean. Hence, the SVD method is particularly useful for situations in which knowledge of the prior mean and prior covariance of the true state (e.g., greenhouse gas profiles) is inadequate.

Ramanathan, Anand K.↗

Setting the Bar for the Replacement of the Probability of Collision Metric

To date, satellite conjunction assessment (CA) risk analysis has largely embraced the probability of collision (Pc) as the omnibus metric to evaluate collision likelihood, and its use in such assessments has mostly been straightforward: at the point at which a conjunction mitigation decision is required, the calculated Pc is compared to a threshold; and if the calculated Pc exceeds that threshold, then a mitigation action is warranted. With only minor variation, this approach is employed by major CA risk assessment centers (e.g., NASA, EUSST, CNES, JAXA) and is advanced as the preferred method in the published CA best practices handbooks. Despite this near unanimity of operational practice, there is a major strain of secondary literature critical of the Pc and willing to propose alternatives. Alfano (2005) pointed out the ability of the Pc to underrepresent the risk in certain situations and counselled a maximum Pc construct. Carpenter (2017, 2019) reiterated this criticism and proposed using instead a confidence interval on the miss distance. Balch et al. (2019) identified what they argued was a defect in the entire Bayesian Pc construct and believed that the use of a more conservative methodology based on covariance ellipsoid overlap was necessary. Delande (2022) introduced the framework of collision “plausibility” to the risk assessment process and sketched out how this might be used operationally. Elkantassi (2022) published a full development of the miss distance confidence interval approach and applied it to several worked examples. While these different approaches to collision risk assessment do differ in their details, they all converge on two central points: first, the Pc’s failure to give an adequate expression of the risk in dilution region situations is a fatal flaw; and second, a conjunction should be presumed risky and in need of mitigation until the evidence of the situation can establish otherwise. These criticisms, if correct, would counsel a number of modifications to current CA operational practice; as such, they force a re-examination of fundamental aspects of the CA problem, including the following: 1. Is the CA risk assessment a probability problem, a statistics problem, or something else? 2. If it is a statistics problem, does it lend itself naturally to a hypothesis test construction? 3. If it can be construed as a hypothesis test, what form should the null hypothesis take, to wit: what constraints exist on the choice of the null hypothesis, what selections are in best alignment with all of the attendant parameters of the problem, and what is implied philosophically by different choices? 4. What are the implications of using the different proposed risk assessment parameters for CA? This question should be answered both in determining how frequently the dilution region situation cited by the critics of the Pc actually appears in an operationally significant manner and the missed detection and false alarm rates of all of the proposed risk assessment metrics, compared both to the Pc and to each other. This paper explores and offers preliminary answers to the above questions, presenting a researched treatment of the philosophical nature of the CA problem and the null hypothesis choice that achieves the greatest consistency with all of the different aspects of operational CA conduct. It then profiles all of the different proposed risk assessment metrics enumerated in the earlier paragraph against an extremely large database of conjunction events at both the 550km and 700km altitudes. The combination of the philosophical exploration of the CA problem and the results of the profiling activity articulates what a risk assessment metric will need to demonstrate, in terms of both innate construction and performance, in order to be a true competitor to the Pc.

conjunction assessment↗

Hierarchical Bayesian estimation for adsorption isotherm parameter determination

Estimation of isotherm model parameters from experimental data is necessary in adsorption process modeling. To facilitate isotherm modeling, databases that store experimental data exist. Nevertheless, data inconsistency among different researchers must be resolved to find a single set of isotherm parameters from multiple data sets. Herein, we propose a hierarchical Bayesian estimation method to quantify the discrepancy by a multiplicative factor while simultaneously obtaining the probability distributions of a single set of isotherm parameters. This approach also allows us to identify outliers, which are data sets that are not in agreement with other sets. The proposed approach is demonstrated for the case of CO 2 adsorption in the metal-organic framework UiO-66.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Dynamic Inversion for Hydrological Process Monitoring with Electrical Resistance Tomography Under Model Uncertainty

We propose an approach for imaging the dynamics of complex hydrological processes. The evolution of electrically conductive fluids in porous media is imaged using time-lapse electrical resistance tomography. The related dynamic inversion problem is solved using Bayesian filtering techniques, that is, it is formulated as a sequential state estimation problem in which the target is an evolving posterior probability density of the system state. The dynamical inversion framework is based on the state space representation of the system, which involves the construction of a stochastic evolution model and an observation model. The observation model used in this paper consists of the complete electrode model for ERT, with Archie's law relating saturations to electrical conductivity. The evolution model is an approximate model for simulating flow through partially saturated porous media. Unavoidable modeling and approximation errors in both the observation and evolution models are considered by computing approximate statistics for these errors. These models are then included in the construction of the posterior probability density of the estimated system state. This approximation error method allows the use of approximate - and therefore computationally efficient - observation and evolution models in the Bayesian filtering. We consider a synthetic example and show that the incorporation of an explicit model for the model uncertainties in the state space representation can yield better estimates than a frame-by-frame imaging approach.

54 ENVIRONMENTAL SCIENCES↗

Integrating System Dynamics and Bayesian Networks with Application to Counter-IED Scenarios

The practice of choosing a single modeling paradigm for predictive analysis can limit the scope and relevance of predictions and their utility to decision-making processes. Considering multiple modeling methods simultaneously may improve this situation, but a better solution provides a framework for directly integrating different, potentially complementary modeling paradigms to enable more comprehensive modeling and predictions, and thus better-informed decisions. The primary challenges of this kind of model integration are to bridge language and conceptual gaps between modeling paradigms, and to determine whether natural and useful linkages can be made in a formal mathematical manner. To address these challenges in the context of two specific modeling paradigms, we explore mathematical and computational options for linking System Dynamics (SD) and Bayesian network (BN) models and incorporating data into the integrated models. We demonstrate that integrated SD/BN models can naturally be described as either state space equations or Dynamic Bayes Nets, which enables the use of many existing computational methods for simulation and data integration. To demonstrate, we apply our model integration approach to techno-social models of insurgent-led attacks and security force counter-measures centered on improvised explosive devices.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

A general Bayesian algorithm for the autonomous alignment of beamlines

Autonomous methods to align beamlines can decrease the amount of time spent on diagnostics, and also uncover better global optima leading to better beam quality. The alignment of these beamlines is a high-dimensional expensive-to-sample optimization problem involving the simultaneous treatment of many optical elements with correlated and nonlinear dynamics. Bayesian optimization is a strategy of efficient global optimization that has proved successful in similar regimes in a wide variety of beamline alignment applications, though it has typically been implemented for particular beamlines and optimization tasks. In this paper, we present a basic formulation of Bayesian inference and Gaussian process models as they relate to multi-objective Bayesian optimization, as well as the practical challenges presented by beamline alignment. We show that the same general implementation of Bayesian optimization with special consideration for beamline alignment can quickly learn the dynamics of particular beamlines in an online fashion through hyperparameter fitting with no prior information. We present the implementation of a concise software framework for beamline alignment and test it on four different optimization problems for experiments on X-ray beamlines at the National Synchrotron Light Source II and the Advanced Light Source, and an electron beam at the Accelerator Test Facility, along with benchmarking on a simulated digital twin. We discuss new applications of the framework, and the potential for a unified approach to beamline alignment at synchrotron facilities.

47 OTHER INSTRUMENTATION↗

Reliability Prediction using FMEA, FTA, and Related Techniques [Slides]

Summary: We presented a reliability analysis framework. We made point estimates of reliability using reliability block diagrams, fault trees, success trees and estimates with uncertainty using expert elicitation, Monte Carlo simulation, Bayesian analysis. Expert elicitation of failure modes and probabilities is labor-intensive, but critical. Bayesian analysis updates information from expert elicitation with data from reliability and aging tests (aging/compatibility data are needed to estimate lower-bound reliabilities at end of life). Estimation by more than one method helps insure consistency and accuracy.

97 MATHEMATICS AND COMPUTING↗

Learning time series for intelligent monitoring

We address the problem of classifying time series according to their morphological features in the time domain. In a supervised machine-learning framework, we induce a classification procedure from a set of preclassified examples. For each class, we infer a model that captures its morphological features using Bayesian model induction and the minimum message length approach to assign priors. In the performance task, we classify a time series in one of the learned classes when there is enough evidence to support that decision. Time series with sufficiently novel features, belonging to classes not present in the training set, are recognized as such. We report results from experiments in a monitoring domain of interest to NASA.

Manganaris, Stefanos↗

A Model-Based Probabilistic Inversion Framework for Wire Fault Detection Using TDR

Time-domain reflectometry (TDR) is one of the standard methods for diagnosing faults in electrical wiring and interconnect systems, with a long-standing history focused mainly on hardware development of both high-fidelity systems for laboratory use and portable hand-held devices for field deployment. While these devices can easily assess distance to hard faults such as sustained opens or shorts, their ability to assess subtle but important degradation such as chafing remains an open question. This paper presents a unified framework for TDR-based chafing fault detection in lossy coaxial cables by combining an S-parameter based forward modeling approach with a probabilistic (Bayesian) inference algorithm. Results are presented for the estimation of nominal and faulty cable parameters from laboratory data.

Schuet, Stefan R.↗

Improving Site-Specific Radiological Performance Assessments - 13431

An improved approach is presented for conducting complete and defensible radiological site-specific performance assessments (PAs) to support radioactive waste disposal decisions. The basic tenets of PA were initiated some thirty years ago, focusing on geologic disposals and evaluating compliance with regulations. Some of these regulations were inherently probabilistic (i.e., addressing uncertainty in a quantitative fashion), such as the containment requirements of the U.S. Environmental Protection Agency's (EPA's) 40 CFR 191, Environmental Radiation Protection Standards for Management and Disposal of Spent Nuclear Fuel, High-Level and Transuranic Radioactive Wastes, Chap. 191.13 [1]. Methods of analysis were developed to meet those requirements, but at their core early PAs used 'conservative' parameter values and modeling approaches. This limited the utility of such PAs to compliance evaluation, and did little to inform decisions about optimizing disposal, closure and long-term monitoring and maintenance, or, in general, maintaining doses 'as low as reasonably achievable' (ALARA). This basic approach to PA development in the United States was employed essentially unchanged through the end of the 20. century, principally by the U.S. Department of Energy (DOE). Performance assessments developed in support of private radioactive waste disposal operations, regulated by the U.S. Nuclear Regulatory Commission (NRC) and its agreement states, were typically not as sophisticated. Discussion of new approaches to PA is timely, since at the time of this writing, the DOE is in the midst of revising its Order 435.1, Radioactive Waste Management [2], and the NRC is revising 10 CFR 61, Licensing Requirements for Land Disposal of Radioactive Waste [3]. Over the previous decade, theoretical developments and improved computational technology have provided the foundation for integrating decision analysis (DA) concepts and objective-focused thinking, plus a Bayesian approach to probabilistic modeling and risk analysis, to guide improvements in PA. This decision-making approach, [4, 5, 6] provides a transparent formal framework for using a value- or objective-focused approach to decision-making. DA, as an analytical means to implement structured decision making, provides a context for both understanding how uncertainty affects decisions and for targeting uncertainty reduction. The proposed DA approach improves defensibility and transparency of decision-making. The DA approach is fully consistent with the need to perform realistic modeling (rather than conservative modeling), including evaluation of site-specific factors. Instead of using generic stylized scenarios for radionuclide fate and transport and for human exposures to radionuclides, site-specific scenarios better represent the advantages and disadvantages of alternative disposal sites or engineered designs, thus clarifying their differences as well as providing a sound basis for evaluation of site performance. The full DA approach to PA is described, from explicitly incorporating societal values through stakeholder involvement to model building. Model building involves scoping by considering features, events, processes, and exposure scenarios (FEPSs), development of a conceptual site model (CSM), translation into numerical models and subsequent computation, and model evaluation. These are implemented in a cycle of uncertainty analysis, sensitivity analysis and value of information analysis so that uncertainty can be reduced until sufficient confidence is gained in the decisions to be made. This includes the traditional focus on hydrogeological processes, but also places emphasis on other FEPSs such as biotically-induced transport and human exposure phenomena. The significance of human exposure scenarios is emphasized by modifying the traditional acronym 'FEPs' to include them, hence 'FEPSs'. The radioactive waste community is also recognizing that disposal sites are to be considered a national (or even global) resource. As such, there is a pressing need to optimize their utility within the constraints of protecting human health and the environment. Failing to do so will result in the need for additional sites or options for storing radioactive waste temporarily, assuming a continued need for radioactive waste disposal. Optimization should be performed using DA, including economic analysis, invoked if necessary through the ALARA process. The economic analysis must recognize the cost of implementation (disposal design, closure, maintenance, etc.), and intra- and inter-generational equity in order to ensure that the best possible radioactive waste management decisions are made for the protection of both current and future generations. In most cases this requires consideration of population or collective risk. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Resolving Mixtures of Soot Characterized by SP-AMS Spectra Using a Latent Dirichlet Allocation Model

Soot produced by detonation or combustion events exhibits different chemical properties depending on the fuel, device construction, and environmental conditions in which the event occurs. These properties can be useful for defining relevant signatures for probabilistically identifying the different types of events that occurred, based on the soot that is produced from these events. However, it is rare to observe samples of soot from a detonation or combustion that are not contaminated by outside particles. In this paper, we present a method for resolving mixtures of soot to determine the contributions of sources that may be present in samples of recovered soot. We use Latent Dirichlet Allocation to describe the generative process for a sample of recovered soot, and use Variational Bayesian Inference to learn about the parameters associated with the generative model. We demonstrate the utility of this method by considering real samples of mixtures of soot under various frameworks to show that the model is able to identify the different components present in a sample of soot as well as their mixing proportions.

54 ENVIRONMENTAL SCIENCES↗

Bayesian Neural Networks for Uncertainty Analysis of Hydrologic Modeling: A Comparison of Two Schemes

Bayesian Neural Networks (BNNs) have been shown as useful tools to analyze modeling uncertainty of Neural Networks (NNs). This research focuses on the comparison of two BNNs. The first BNNs (BNN-I) use statistical methods to describe the characteristics of different uncertainty sources (input, parameter, and model structure) and integrate these uncertainties into a Markov Chain Monte Carlo (MCMC) framework to estimate total uncertainty. The second BNNs (BNN-II) lump all uncertainties into a single error term (i.e. the residual between model prediction and measurement). In this study, we propose a simple BNN-II, which use Genetic Algorithms (GA) and Bayesian Model Averaging (BMA) to calibrate Neural Networks with different structures (number of hidden units) and combine the predictions from different NNs to derive predictions and uncertainty analysis. We tested these two BNNs in two watersheds for daily and monthly hydrologic simulation. The BMA based BNNs developed in this study outperforms BNN-I in the two watersheds in terms of both accurate prediction and uncertainty estimation. These results show that, given incomplete understanding of the characteristics associated with each uncertainty source, the simple lumped error approach may yield better prediction and uncertainty estimation.

99 GENERAL AND MISCELLANEOUS↗

Parallelizing autotuning for HPC applications: Unveiling the potential of the speculation strategy in Bayesian optimization

In the exascale computing era, tuning High-Performance Computing (HPC) applications has become a significant computational challenge. Although Bayesian optimization (BO) has emerged as a promising tool for HPC performance tuning, the BO workflow is inherently sequential (i.e., one function evaluation at a time) and cannot leverage the huge amount of parallel resources present in modern supercomputers, resulting in a considerable underutilization of their computational capabilities. This paper explores the trade-off between search quality and parallelism in BO, investigating a diverse set of methods. Building upon both previous approaches from the literature and novel methodologies introduced in this work, our study provides a deep analysis to accelerate BO performance tuning. By examining a set of synthetic functions and practical HPC applications, our exploration analyzes the interaction among various BO methods for parallelization, the quantity of parallel resources, the runtime distribution of target HPC applications, and the costs associated with different search orchestration mechanisms that have been overlooked in previous studies. Compared to sequential BO, our novel methodology achieves comparable quality while demonstrating robust scalability in search time as the amount of parallel resources increases; it also outperforms a state-of-the-art tuner, which supports parallelization, achieving up to 3.67x faster search time. We provide high-value insights for practitioners seeking to leverage the power of parallel computing for efficient HPC application tuning. Additionally, to further assist researchers in accelerating the performance tuning of their HPC applications, we provide an extension of an existing open-source tuning framework that incorporates our methods.

Bayesian optimization↗