SEARCH · Search NASA
Results for “inference methods”
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.
Real-Time Pathogen Detection in the Era of Whole-Genome Sequencing and Big Data: Comparison of k-mer and Site-Based Methods for Inferring the Genetic Distances among Tens of Thousands of Salmonella Samples
The adoption of whole-genome sequencing within the public health realm for molecular characterization of bacterial pathogens has been followed by an increased emphasis on real-time detection of emerging outbreaks (e.g., food-borne Salmonellosis). In turn, large databases of whole-genome sequence data are being populated. These databases currently contain tens of thousands of samples and are expected to grow to hundreds of thousands within a few years. For these databases to be of optimal use one must be able to quickly interrogate them to accurately determine the genetic distances among a set of samples. Being able to do so is challenging due to both biological (evolutionary diverse samples) and computational (petabytes of sequence data) issues. We evaluated seven measures of genetic distance, which were estimated from either k-mer profiles (Jaccard, Euclidean, Manhattan, Mash Jaccard, and Mash distances) or nucleotide sites (NUCmer and an extended multi-locus sequence typing (MLST) scheme). Finally, when analyzing empirical data (wholegenome sequence data from 18,997 Salmonella isolates) there are features (e.g., genomic, assembly, and contamination) that cause distances inferred from k-mer profiles, which treat absent data as informative, to fail to accurately capture the distance between samples when compared to distances inferred from differences in nucleotide sites. Thus, site-based distances, like NUCmer and extended MLST, are superior in performance, but accessing the computing resources necessary to perform them may be challenging when analyzing large databases.
Protein-ligand binding affinity prediction using multi-instance learning with docking structures
Recent advances in 3D structure-based deep learning approaches demonstrate improved accuracy in predicting protein-ligand binding affinity in drug discovery. These methods complement physics-based computational modeling such as molecular docking for virtual high-throughput screening. Despite recent advances and improved predictive performance, most methods in this category primarily rely on utilizing co-crystal complex structures and experimentally measured binding affinities as both input and output data for model training. Nevertheless, co-crystal complex structures are not readily available and the inaccurate predicted structures from molecular docking can degrade the accuracy of the machine learning methods. We introduce a novel structure-based inference method utilizing multiple molecular docking poses for each complex entity. Our proposed method employs multi-instance learning with an attention network to predict binding affinity from a collection of docking poses. We validate our method using multiple datasets, including PDBbind and compounds targeting the main protease of SARS-CoV-2. The results demonstrate that our method leveraging docking poses is competitive with other state-of-the-art inference models that depend on co-crystal structures. This method offers binding affinity prediction without requiring co-crystal structures, thereby increasing its applicability to protein targets lacking such data.
Spatially resolved X-ray emission measurements of the residual velocity during the stagnation phase of inertial confinement fusion implosion experiments
A technique for measuring residual motion during the stagnation phase of an indirectly driven inertial confinement experiment has been implemented. This method infers a velocity from spatially and temporally resolved images of the X-ray emission from two orthogonal lines of sight. This work investigates the accuracy of recovering spatially resolved velocities from the X-ray emission data. A detailed analytical and numerical modeling of the X-ray emission measurement shows that the accuracy of this method increases as the displacement that results from a residual velocity increase. For the typical experimental configuration, signal-to-noise ratios, and duration of X-ray emission, it is estimated that the fractional error in the inferred velocity rises above 50% as the velocity of emission falls below 24 μm/ns. By inputting measured parameters into this model, error estimates of the residual velocity as inferred from the X-ray emission measurements are now able to be generated for experimental data. Details of this analysis are presented for an implosion experiment conducted with an unintentional radiation flux asymmetry. The analysis shows a bright localized region of emission that moves through the larger emitting volume at a relatively higher velocity towards the location of the imposed flux deficit. This technique allows for the possibility of spatially resolving velocity flows within the so-called central hot spot of an implosion. This information would help to refine our interpretation of the thermal temperature inferred from the neutron time of flight detectors and the effect of localized hydrodynamic instabilities during the stagnation phase. Across several experiments, along a single line of sight, the average difference in magnitude and direction of the measured residual velocity as inferred from the X-ray and neutron time of flight detectors was found to be ∼13 μm/ns and ∼14°, respectively.
Approximate Bayesian Model Inversion for PDEs with Heterogeneous and State-Dependent Coefficients
We present two approximate Bayesian inference methods for parameter estimation in partial differential equation (PDE) models with space-dependent and state-dependent parameters. We demonstrate that these methods provide accurate and cost-effective alternatives to Markov Chain Monte Carlo simulation. We assume a parameterized Gaussian prior on the unknown functions, and approximate the posterior density by a parameterized multivariate Gaussian density. The parameters of the prior and posterior are estimated from sparse observations of the PDE model's states and the unknown functions themselves by maximizing the evidence lower bound (ELBO), a lower bound on the log marginal likelihood of the observations. The first method, Laplace-EM, employs the expectation maximization algorithm to maximize the ELBO, with a Laplace approximation of the posterior on the E-step, and minimization of a Kullback-Leibler divergence on the M-step. The second method, DSVI-EB, employs the doubly stochastic variational inference (DSVI) algorithm, in which the ELBO is maximized via gradient-based stochastic optimization, with nosiy gradients computed via simple Monte Carlo sampling and Gaussian backpropagation. We apply these methods to identifying diffusion coefficients in linear and nonlinear diffusion equations, and we find that both methods provide accurate estimates of posterior densities and the hyperparameters of Gaussian priors. While the Laplace-EM method is more accurate, it requires computing Hessians of the physics model. The DSVI-EB method is found to be less accurate but only requires gradients of the physics model.
Filtering and Domain Decomposition Techniques for Intrusive and Non-intrusive Reduced Order Models of Convection-Dominated Problems
These chapters contain methods to regularize and generally improve reduced order models using Operator Inference methods. This is accomplished using the Schwarz method (chapter 5) as well as spatial filtering methods (Chapter 6)
Analytical methods for superresolution dislocation identification in dark-field X-ray microscopy
In this work, we develop several inference methods to estimate the position of dislocations from images generated using dark-field X-ray microscopy (DFXM)—achieving superresolution accuracy and principled uncertainty quantification. Using the framework of Bayesian inference, we incorporate models of the DFXM contrast mechanism and detector measurement noise, along with initial position estimates, into a statistical model coupling DFXM images with the dislocation position of interest. We motivate several position estimation and uncertainty quantification algorithms based on this model. We then demonstrate the accuracy of our primary estimation algorithm on synthetic realistic DFXM images of edge dislocations in single-crystal aluminum. We conclude with a discussion of our methods’ impact on future dislocation studies and possible future research avenues.
Geotagging one hundred million Twitter accounts with total variation minimization
Geographically annotated social media is extremely valuable for modern information retrieval. However, when researchers can only access publicly-visible data, one quickly finds that social media users rarely publish location information. In this work, we provide a method which can geolocate the overwhelming majority of active Twitter users, independent of their location sharing preferences, using only publicly-visible Twitter data. Our method infers an unknown user's location by examining their friend's locations. We frame the geotagging problem as an optimization over a social network with a total variation-based objective and provide a scalable and distributed algorithm for its solution. Furthermore, we show how a robust estimate of the geographic dispersion of each user's ego network can be used as a per-user accuracy measure which is effective at removing outlying errors. Leave-many-out evaluation shows that our method is able to infer location for 101, 846, 236 Twitter users at a median error of 6.38 km, allowing us to geotag over 80% of public tweets.
Deep Koopman operators for causal discovery
Causal discovery aims to identify cause-effect mechanisms for better scientific understanding, explainable decision-making, and more accurate modeling. Standard statistical frameworks, such as Granger causality, lack the ability to quantify causal relationships in nonlinear dynamics due to the presence of complex feedback mechanisms, timescale mixing, and nonstationarity. Thus, applying these methods to study causal dynamics in real-world systems, such as the Earth, is a major challenge. Addressing this shortcoming, we leverage deep learning and a Koopman operator-theoretic formalism to present a class of causal discovery algorithms. Kausal uses deep Koopman operator methods to approximate nonlinear dynamics in a linearized vector space in which traditional causal inference methods such as Granger causality can be more easily applied. Our idealized experiments demonstrate Kausal’s superior ability in discovering and characterizing causal signals compared to existing deep learning and non-deep learning state-of-the-art approaches. Finally, the successful identification of major El Niño and La Niña events in observations showcases Kausal’s skill to handle real-world applications.
Learning Latent Interactions for Event Identification via Graph Neural Networks and PMU Data
Phasor measurement units (PMUs) are being widely installed on power systems, providing a unique opportunity to enhance wide-area situational awareness. One essential application is the use of PMU data for real-time event identification. However, how to take full advantage of all PMU data in event identification is still an open problem. Thus, we propose a novel method that performs event identification by mining interaction graphs among different PMUs. The proposed interaction graph inference method follows an entirely data-driven manner without knowing the physical topology. Moreover, unlike previous works that treat interactive learning and event identification as two different stages, our method learns interactions jointly with the identification task, thereby improving the accuracy of graph learning and ensuring seamless integration between the two stages. Moreover, to capture multi-scale event patterns, a dilated inception-based method is investigated to perform feature extraction of PMU data. To test the proposed data-driven approach, a large real-world dataset from tens of PMU sources and the corresponding event logs have been utilized in this work. We report numerical results validate that our method has higher classification accuracy compared to previous methods.
PyCBC Inference: A Python-based Parameter Estimation Toolkit for Compact Binary Coalescence Signals
We introduce new modules in the open-source PyCBC gravitational-wave astronomy toolkit that implement Bayesian inference for compact-object binary mergers. We review the Bayesian inference methods implemented and describe the structure of the modules. Here, we demonstrate that the PyCBC Inference modules produce unbiased estimates of the parameters of a simulated population of binary black hole mergers. We show that the parameters' posterior distributions obtained using our new code agree well with the published estimates for binary black holes in the first Advanced LIGO–Virgo observing run.
Polynomial Chaos–Based Bayesian Inference of K -Profile Parameterization in a General Circulation Model of the Tropical Pacific
The authors present a polynomial chaos (PC)–based Bayesian inference method for quantifying the uncertainties of the K-profile parameterization (KPP) within the MIT general circulation model (MITgcm) of the tropical Pacific. The inference of the uncertain parameters is based on a Markov chain Monte Carlo (MCMC) scheme that utilizes a newly formulated test statistic taking into account the different components representing the structures of turbulent mixing on both daily and seasonal time scales in addition to the data quality, and filters for the effects of parameter perturbations over those as a result of changes in the wind. To avoid the prohibitive computational cost of integrating the MITgcm model at each MCMC iteration, a surrogate model for the test statistic using the PC method is built. Because of the noise in the model predictions, a basis-pursuit-denoising (BPDN) compressed sensing approach is employed to determine the PC coefficients of a representative surrogate model. The PC surrogate is then used to evaluate the test statistic in the MCMC step for sampling the posterior of the uncertain parameters. Results of the posteriors indicate good agreement with the default values for two parameters of the KPP model, namely the critical bulk and gradient Richardson numbers; while the posteriors of the remaining parameters were barely informative.
Identifying strengths and weaknesses of methods for computational network inference from single-cell RNA-seq data
Single-cell RNA-sequencing (scRNA-seq) offers unparalleled insight into the transcriptional programs of different cellular states by measuring the transcriptome of thousands of individual cells. An emerging problem in the analysis of scRNA-seq is the inference of transcriptional gene regulatory networks and a number of methods with different learning frameworks have been developed to address this problem. Here, we present an expanded benchmarking study of eleven recent network inference methods on seven published scRNA-seq datasets in human, mouse, and yeast considering different types of gold standard networks and evaluation metrics. We evaluate methods based on their computing requirements as well as on their ability to recover the network structure. We find that, while most methods have a modest recovery of experimentally derived interactions based on global metrics such as Area Under the Precision Recall curve, methods are able to capture targets of regulators that are relevant to the system under study. Among the top performing methods that use only expression were SCENIC, PIDC, MERLIN or Correlation. Addition of prior biological knowledge and the estimation of transcription factor activities resulted in the best overall performance with the Inferelator and MERLIN methods that use prior knowledge outperforming methods that use expression alone. We found that imputation for network inference did not improve network inference accuracy and could be detrimental. Comparisons of inferred networks for comparable bulk conditions showed that the networks inferred from scRNA-seq datasets are often better or at par with the networks inferred from bulk datasets. Our analysis should be beneficial in selecting methods for network inference. At the same time, this highlights the need for improved methods and better gold standards for regulatory network inference from scRNAseq datasets.
MO-FG-CAMPUS-IeP2-01: Characterization of Beam Shaping Filters and Photon Spectra From HVL Profiles in CT
Purpose: Advanced dosimetry in CT (e.g. the Monte Carlo method) requires an accurate characterization of the shaped filter and radiation quality used during a scan. The purpose of this work was to develop a method where half value layer (HVL) profiles along shaped filters could be made. From the HVL profiles the beam shaping properties and effective photon spectrum for a particular scan can be inferred. Methods: A measurement rig was developed to allow determinations of the HVL under a scatter-free narrow-beam geometry and constant focal spot to ionization chamber distance for different fan angles. For each fan angle the HVL is obtained by fitting the transmission of radiation through different thicknesses of an Al absorber (type 1100) using an appropriate model. The effective Al thickness of shaped filters and effective photon spectra are estimated using a model of photon emission from a Tungsten anode. This method is used to obtain the effective photon spectra and effective Al thickness of shaped filters for a CT scanner recently introduced to the market. Results: This study resulted in a set of effective photon spectra (central ray) for each kVp along with effective Al thicknesses of the different shaped filters. The effective photon spectra and effective Al thicknesses of shaped filters were used to obtain numerically approximated HVL profiles and compared to measured HVL profiles (mean absolute percentage error = 0.02). The central axis HVL found in the vendor’s technical documentation were compared to approximated HVL values (mean absolute percentage error = 0.03). Conclusion: This work has resulted in a unique method of measuring HVL profiles along shaped filters in CT. Further the effective photon spectra and the effective Al thicknesses of shaped filters that were obtained can be incorporated into Monte Carlo simulations.
Monitoring the State-of-Charge of a Vanadium Redox Flow Battery with the Acoustic Attenuation Coefficient: An In Operando Noninvasive Method
Redox flow battery technology has been increasingly recognized as a promising option for large-scale grid energy storage. Access to high-fidelity information on the health status of the electrolyte, including the state-of-charge (SOC), is vital to maintaining optimal and economical battery operation. In this study, an ultrasonic probing cell that can be used to measure SOC in real time is designed. This unprecedented, new measurement approach overcomes the influence of varying temperatures by measuring the acoustic attenuation coefficient of the redox flow battery electrolyte online and noninvasively. The new approach is used to estimate the SOC of a vanadium redox flow battery (VRFB) in operando from measured acoustic properties. The accuracy of the SOC estimated from the acoustic properties is validated against SOC calculated by the titration method. The results show that the acoustic attenuation coefficient is a robust parameter for SOC monitoring, with a maximum error of 4.8% and extremely low sensitivity to temperature, while sound speed appears to be less accurate in the benchmark-inference method, with a maximum error of 22.5% and high sensitivity to temperature. The acoustic measurement approach has great potential for inexpensive real-time SOC monitoring of redox flow battery operations.
Joint state-parameter estimation for the reduced fracture model via the united filter
Here, in this paper, we introduce an effective United Filter method for jointly estimating the solution state and physical parameters in flow and transport problems within fractured porous media. Fluid flow and transport in fractured porous media are critical in subsurface hydrology, geophysics, and reservoir geomechanics. Reduced fracture models, which represent fractures as lower-dimensional interfaces, enable efficient multi-scale simulations. However, reduced fracture models also face accuracy challenges due to modeling errors and uncertainties in physical parameters such as permeability and fracture geometry. To address these challenges, we propose a United Filter method, which integrates the Ensemble Score Filter (EnSF) for state estimation with the Direct Filter for parameter estimation. EnSF, based on a score-based diffusion model framework, produces ensemble representations of the state distribution without deep learning. Meanwhile, the Direct Filter, a recursive Bayesian inference method, estimates parameters directly from state observations. The United Filter combines these methods iteratively: EnSF estimates are used to refine parameter values, which are then fed back to improve state estimation. Numerical experiments demonstrate that the United Filter method surpasses the state-of-the-art Augmented Ensemble Kalman Filter, delivering more accurate state and parameter estimation for reduced fracture models. This framework also provides a robust and efficient solution for PDE-constrained inverse problems with uncertainties and sparse observations.
Bayesian operator inference for data-driven reduced-order modeling
This work proposes a Bayesian inference method for the reduced-order modeling of time-dependent systems. Informed by the structure of the governing equations, the task of learning a reduced-order model from data is posed as a Bayesian inverse problem with Gaussian prior and likelihood. The resulting posterior distribution characterizes the operators defining the reduced-order model, hence the predictions subsequently issued by the reduced-order model are endowed with uncertainty. The statistical moments of these predictions are estimated via a Monte Carlo sampling of the posterior distribution. Since the reduced models are fast to solve, this sampling is computationally efficient. Furthermore, the proposed Bayesian framework provides a statistical interpretation of the regularization term that is present in the deterministic operator inference problem, and the empirical Bayes approach of maximum marginal likelihood suggests a selection algorithm for the regularization hyperparameters. The proposed method is demonstrated on two examples: the compressible Euler equations with noise-corrupted observations, and a single-injector combustion process.
Functional Data Analysis for Extracting the Intrinsic Dimensionality of Spectra: Application to Chemical Homogeneity in the Open Cluster M67
High-resolution spectroscopic surveys of the Milky Way have entered the Big Data regime and have opened avenues for solving outstanding questions in Galactic archeology. However, exploiting their full potential is limited by complex systematics, whose characterization has not received much attention in modern spectroscopic analyses. In this work, we present a novel method to disentangle the component of spectral data space intrinsic to the stars from that due to systematics. Using functional principal component analysis on a sample of 18,933 giant spectra from APOGEE, we find that the intrinsic structure above the level of observational uncertainties requires ≈10 functional principal components (FPCs). Our FPCs can reduce the dimensionality of spectra, remove systematics, and impute masked wavelengths, thereby enabling accurate studies of stellar populations. To demonstrate the applicability of our FPCs, we use them to infer stellar parameters and abundances of 28 giants in the open cluster M67. We employ Sequential Neural Likelihood, a simulation-based Bayesian inference method that learns likelihood functions using neural density estimators, to incorporate non-Gaussian effects in spectral likelihoods. By hierarchically combining the inferred abundances, we limit the spread of the following elements in M67: Fe ≲ 0.02 dex; C ≲ 0.03 dex; O, Mg, Si, Ni ≲ 0.04 dex; Ca ≲ 0.05 dex; N, Al ≲ 0.07 dex (at 68% confidence). Our constraints suggest a lack of self-pollution by core-collapse supernovae in M67, which has promising implications for the future of chemical tagging to understand the star formation history and dynamical evolution of the Milky Way.