Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “inverse prediction 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.

At least 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

The Principle of Maximum Conformality Correctly Resolves the Renormalization-Scheme-Dependence Problem

In this paper, we clarify a serious misinterpretation and consequent misuse of the Principle of Maximum Conformality (PMC), which also can serve as a mini-review of PMC. In a recently published article, P. M. Stevenson has claimed that “the PMC is ineffective and does nothing to resolve the renormalization-scheme-dependence problem”, concluding incorrectly that the success of PMC predictions is due to the PMC being a “laborious, ad hoc, and back-door” version of the Principle of Minimal Sensitivity (PMS). We show that such conclusions are incorrect, deriving from a misinterpretation of the PMC and an overestimation of the applicability of the PMS. The purpose of the PMC is to achieve precise fixed-order pQCD predictions, free from conventional renormalization schemes and scale ambiguities. We demonstrate that the PMC predictions satisfy all the self-consistency conditions of the renormalization group and standard renormalization-group invariance; the PMC predictions are thus independent of any initial choice of renormalization scheme and scale. The scheme independence of the PMC is also ensured by commensurate scale relations, which relate different observables to each other. Moreover, in the Abelian limit, the PMC dovetails into the well-known Gell-Mann–Low framework, a method universally revered for its precision in QED calculations. Due to the elimination of factorially divergent renormalon terms, the PMC series not only attains a convergence behavior far superior to that of its conventional counterparts but also deftly curtails any residual scale dependence caused by the unknown higher-order terms. This refined convergence, coupled with its robust suppression of residual uncertainties, furnishes a sound and reliable foundation for estimating the contributions from unknown higher-order terms. Anchored in the bedrock of standard renormalization-group invariance, the PMC simultaneously eradicates the factorial divergences and eliminates superfluous systematic errors, which inversely provides a good foundation for achieving high-precision pQCD predictions. Consequently, owing to its rigorous theoretical underpinnings, the PMC is eminently applicable to virtually all high-energy hadronic processes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

Bayesian Analysis of TRISO Fuel: Quantifying Model Inadequacy, Incorporating Lower-Length-Scale Effects, and Developing Parallel Active Learning Capabilities

The U.S. Department of Energy (DOE)’s Nuclear Energy Advanced Modeling and Simulation (NEAMS) program aims to develop predictive capabilities by applying computational methods to the analysis and design of advanced reactor and fuel-cycle systems. This program has been providing engineering-scale support for the continued development of BISON, a high-fidelity, high-resolution fuel performance tool. Fuel behavior in nuclear reactors is governed by a complex network of mechanisms that interact with various other physics aspects in the reactor system. Any model developed to represent fuel behavior will likely be idealized, resulting in uncertainties when comparing their predictions against the observed data. In Fiscal Year (FY)-23, we initiated the Uncertainty Quantification (UQ) work by using Bayesian methods to establish a level of model trustworthiness and further improve it, with a particular emphasis on TRI-Structural isOtropic (TRISO) nuclear fuel. This year, we further expanded on that UQ work by investigating an approach to quantifying model inadequacy and accounting for lower-length scale (LLS) effects in TRISO silver (Ag) release modeling. Furthermore, we are implementing parallel active learning capabilities to reduce the computational cost (i.e., required computational resources and elapsed time) of performing UQ. Specifically, we utilized The Kennedy O’Hagan framework for Bayesian uncertainty quantification (KOH) to account for model inadequacy in TRISO Ag release predictions made by BISON. The KOH framework represents an improvement over the standard Bayesian framework used in FY-23. Explicitly accounting for model inadequacy in the Bayesian framework helps establish the level of experimental noise uncertainty in the Advanced Gas Reactor (AGR) data. We compared the inverse UQ results obtained from both the standard Bayesian and KOH frameworks in light of the AGR-2/3/4 data, and also compared the predictive UQ results obtained from these two frameworks in light of the AGR-1 data. Next, we investigated the impact of considering LLS effects in the Ag release simulations. We developed an expanded database of LLS simulated effective diffusivities for Ag, covering a wide range of microstructures and temperatures. Using this database, we developed a framework for incorporating LLS effects into the engineering-scale Ag release UQ. We developed both parametric and non-parametric approaches for bridging the length scales. We then investigated the inverse UQ results in light of the AGR-2/3/4 data and the predictive UQ results in light of the AGR-1 data, and compared the LLS-informed approach and the Arrhenius equation, which does not include microstructure information. Finally, we discussed implementing parallel active learning capabilities in the Multiphysics Object Oriented Simulation Environment (MOOSE)/BISON to reduce the computational cost (i.e., computational resources and elapsed time) of Bayesian UQ. For verification purposes, we first tested these new capabil ities on a species interaction problem. We then demonstrated them on the TRISO Ag release application, showing that parallel active learning capabilities can enhance the accuracy of UQ while also substantially reducing the computational cost in comparison to the reference methods developed in FY-23.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗
DOE OSTI2025

Bayesian model mixing with multireference energy density functional

Reliably predicting nuclear properties across the entire chart of isotopes is important for applications ranging from nuclear astrophysics to superheavy science to nuclear technology. To this day, however, all the theoretical models that can scale at the level of the chart of isotopes remain semiphenomenological. Because they are fitted locally, their predictive power can vary significantly; different versions of the same theory provide different predictions. Bayesian model mixing takes advantage of such imperfect models to build a local mixture of a set of models to make improved predictions. Earlier attempts to use Bayesian model mixing for mass table calculations relied on models treated at single-reference energy density functional level, which fail to capture some of the correlations caused by configuration mixing or the restoration of broken symmetries. In this study we have applied Bayesian model mixing techniques within a multireference energy density functional (MR-EDF) framework. We considered predictions of two-particle separation energies from particle number projection or angular momentum projection with four different energy density functionals—a total of eight different MR-EDF models. We used a hierarchical Bayesian stacking framework with a Dirichlet prior distribution over weights together with an inverse log-ratio transform to enable positive correlations between different models. We found that Bayesian model mixing provides significantly improved predictions compared to the participating models. Published by the American Physical Society 2025

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Machine Learning Potentials with the Iterative Boltzmann Inversion: Training to Experiment

Methodologies for training machine learning potentials (MLPs) with quantum-mechanical simulation data have recently seen tremendous progress. Experimental data have a very different character than simulated data, and most MLP training procedures cannot be easily adapted to incorporate both types of data into the training process. Here, we investigate a training procedure based on iterative Boltzmann inversion that produces a pair potential correction to an existing MLP using equilibrium radial distribution function data. By applying these corrections to an MLP for pure aluminum based on density functional theory, we observe that the resulting model largely addresses previous overstructuring in the melt phase. Interestingly, the corrected MLP also exhibits improved performance in predicting experimental diffusion constants, which are not included in the training procedure. The presented method does not require autodifferentiating through a molecular dynamics solver and does not make assumptions about the MLP architecture. Our results suggest a practical framework for incorporating experimental data into machine learning models to improve the accuracy of molecular dynamics simulations.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Using Machine Learning to Predict Cloud Turbulent Entrainment–Mixing Processes

Different turbulent entrainment–mixing mechanisms between clouds and environment are essential to cloud–related processes; however, accurate representation of entrainment–mixing in weather/climate models still poses a challenge. This study exploits the use of machine learning (ML) to address this challenge. Four ML (Light Gradient Boosting Machine [LGB], eXtreme Gradient Boosting, Random Forest, and Support Vector Regression) are examined and compared. It is found that LGB performs best, and thus is selected to understand the impact of entrainment–mixing on microphysics using simulation data from Explicit Mixing Parcel Model. Compared with traditional parameterizations, the trained LGB provides more accurate microphysical properties (number concentration and cloud droplet spectral dispersion). The partial dependences of predicted microphysics on features exhibit a strong alignment with physical mechanisms and expectations, as determined by the interpreting method, thus overcoming the limitations of the “black box” scheme. The underlying mechanisms are that the smaller number concentration and larger spectral dispersion correspond to more inhomogeneous entrainment–mixing. Specifically, number concentration after entrainment–mixing is positively correlated with adiabatic number concentration and liquid water content affected by entrainment–mixing, and inversely correlated with adiabatic volume mean radius. Spectral dispersion after entrainment–mixing is negatively correlated with liquid water content affected by entrainment–mixing, turbulent dissipation rate and relative humidity of entrained air. Sensitivity analysis further suggests that number concentration is mainly determined by cloud microphysical properties whereas spectral dispersion is influenced by both cloud microphysical properties and environmental variables. The results indicate that the LGB scheme has the potential to enhance the representation of entrainment–mixing in weather/climate models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Deep learning multiphysics network for imaging CO 2 saturation and estimating uncertainty in geological carbon storage

Multiphysics inversion exploits different types of geophysical data that often complement each other and aims to improve overall imaging resolution and reduce uncertainties in geophysical interpretation. Despite the advantages, traditional multiphysics inversion is challenging because it requires a large amount of computational time and intensive human interactions for preprocessing data and finding trade-off parameters. These issues make it nearly impossible for traditional multiphysics inversion to be applied as a real-time monitoring tool for geological carbon storage. In this paper, we present a deep learning (DL) multiphysics network for imaging CO 2 saturation in real time. The multiphysics network consists of three encoders for analysing seismic, electromagnetic and gravity data and shares one decoder for combining imaging capabilities of the different geophysical data for better predicting CO 2 saturation. The network is trained on pairs of CO 2 label models and multiphysics data so that it can directly image CO 2 saturation. Here we use the bootstrap aggregating method to enhance the imaging accuracy and estimate uncertainties associated with CO 2 saturation images. Using realistic CO 2 label models and multiphysics data derived from the Kimberlina CO 2 storage model, we evaluate the performance of the deep learning multiphysics network and compare its imaging results to those from the deep learning single-physics networks. Our modelling experiments show that the deep learning multiphysics network for seismic, electromagnetic, and gravity data not only improves the imaging accuracy but also reduces uncertainties associated with CO 2 saturation images. Our results also suggest that the deep learning multiphysics network for the non-seismic data (i.e., electromagnetic and gravity) can be used as an effective low-cost monitoring tool in between regular seismic monitoring.

58 GEOSCIENCES↗
DOE OSTI2022

Physics constrained learning for data-driven inverse modeling from sparse observations

Deep neural networks (DNN) have been used to model nonlinear relations between physical quantities. Those DNNs are embedded in physical systems described by partial differential equations (PDE) and trained by minimizing a loss function that measures the discrepancy between predictions and observations in some chosen norm. This loss function often includes the PDE constraints as a penalty term when only sparse observations are available. As a result, the PDE is only satisfied approximately by the solution. However, the penalty term typically slows down the convergence of the optimizer for stiff problems. We present a new approach that trains the embedded DNNs while numerically satisfying the PDE constraints. We develop an algorithm that enables differentiating both explicit and implicit numerical solvers in reverse-mode automatic differentiation. This allows the gradients of the DNNs and the PDE solvers to be computed in a unified framework. We demonstrate that our approach enjoys faster convergence and better stability in relatively stiff problems compared to the penalty method. Furthermore, our approach allows for the potential to solve and accelerate a wide range of data-driven inverse modeling, where the physical constraints are described by PDEs and need to be satisfied accurately.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Basic Research Needs for Inverse Methods for Complex Systems under Uncertainty

Inverse problems, which aim to infer unknown properties of a system using experimental and observational data, are central to addressing many of the U.S. Department of Energy’s (DOE) most critical scientific and engineering challenges. Accurate, computationally efficient, and data-efficient solutions to inverse problems are essential for advancing DOE mission-critical science drivers, including analyzing data from large-scale experimental facilities, optimizing fusion reactor performance, accelerating materials discovery, enhancing geophysical imaging, improving wildfire predictions, and enabling autonomous systems and digital twins. However, these problems are becoming increasingly complex, often involving nonlinear, highdimensional, and interconnected systems and models that span multiple physics and scales, while relying on data with varying quantity, quality, and information content. Compounding these challenges is the uncertainty inherent in DOE-relevant systems, where errors in inputs, noise in data, incompleteness of data, and discrepancies between models and reality constrain the accuracy and precision of solutions. At the same time, the convergence of recent scientific computing trends—scientific machine learning, artificial intelligence, and computing advances such as exascale computing—is creating unprecedented opportunities for tackling these challenges. The cross-cutting nature of inverse problems, combined with their growing complexity and rapidly evolving data and algorithmic demands, strongly motivates the formulation of a prioritized research agenda to maximize their capabilities and impact. In response to this need, DOE’s Advanced Scientific Computing Research (ASCR) program in the Office of Science convened the Workshop on Basic Research Needs for Inverse Problems for Complex Systems Under Uncertainty in June 2025. This workshop brought together experts across disciplines to identify grand challenges and major opportunities in the field. Through collaborative discussions, the workshop defined transformative research directions aimed at addressing the mathematical, statistical, and computational challenges posed by inverse problems under uncertainty. As a result of these efforts, four priority research directions (PRDs) were identified to guide future research and development in this area. These PRDs, summarized below, represent a roadmap for advancing the foundational science and mathematics of inverse problems, enabling robust, scalable, and uncertainty-aware solutions that are critical for DOE applications.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Training calibration-based counterfactual explainers for deep learning models in medical image analysis

The rapid adoption of artificial intelligence methods in healthcare is coupled with the critical need for techniques to rigorously introspect models and thereby ensure that they behave reliably. This has led to the design of explainable AI techniques that uncover the relationships between discernible data signatures and model predictions. In this context, counterfactual explanations that synthesize small, interpretable changes to a given query while producing desired changes in model predictions have become popular. This under-constrained, inverse problem is vulnerable to introducing irrelevant feature manipulations, particularly when the model’s predictions are not well-calibrated. Hence, in this paper, we propose the TraCE (training calibration-based explainers) technique, which utilizes a novel uncertainty-based interval calibration strategy for reliably synthesizing counterfactuals. Given the wide-spread adoption of machine-learned solutions in radiology, our study focuses on deep models used for identifying anomalies in chest X-ray images. Using rigorous empirical studies, we demonstrate the superiority of TraCE explanations over several state-of-the-art baseline approaches, in terms of several widely adopted evaluation metrics. Our findings show that TraCE can be used to obtain a holistic understanding of deep models by enabling progressive exploration of decision boundaries, to detect shortcuts, and to infer relationships between patient attributes and disease severity.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2020

A Bayesian Approach to Regional Decadal Predictability: Sparse Parameter Estimation in High-Dimensional Linear Inverse Models of High-Latitude Sea Surface Temperature Variability

Stochastic reduced models are an important tool in climate systems whose many spatial and temporal scales cannot be fully discretized or underlying physics may not be fully accounted for. One form of reduced model, the linear inverse model (LIM), has been widely used for regional climate predictability studies—typically focusing more on tropical or midlatitude studies. However, most LIM fitting techniques rely on point estimation techniques deriving from fluctuation–dissipation theory. In this methodological study we explore the use of Bayesian inference techniques for LIM parameter estimation of sea surface temperature (SST), to quantify the skillful decadal predictability of Bayesian LIM models at high latitudes. We show that Bayesian methods, when compared to traditional point estimation methods for LIM-type models, provide better calibrated probabilistic skill, while simultaneously providing better point estimates due to the regularization effect of the prior distribution in high-dimensional problems. We compare the effect of several priors, as well as maximum likelihood estimates, on 1) estimating parameter values on a perfect model experiment and 2) producing calibrated 1-yr SST anomaly forecast distributions using a preindustrial control run of the Community Earth System Model (CESM). Finally, we employ a host of probabilistic skill metrics to determine the extent to which an LIM can forecast SST anomalies at high latitudes. We find that the choice of prior distribution has an appreciable impact on estimation outcomes, and priors that emphasize physically relevant properties enhance the model’s ability to capture variability of SST anomalies.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Bayesian Poroelastic Aquifer Characterization From InSAR Surface Deformation Data. Part I: Maximum A Posteriori Estimate

Characterizing the properties of groundwater aquifers is essential for predicting aquifer response and managing groundwater resources. In this work, we develop a high-dimensional scalable Bayesian inversion framework governed by a three-dimensional quasi-static linear poroelastic model to characterize lateral permeability variations in groundwater aquifers. In this work, we determine the maximum a posteriori (MAP) point of the posterior permeability distribution from centimeter-level surface deformation measurements obtained from Interferometric Synthetic Aperture Radar (InSAR). The scalability of our method to high parameter dimension is achieved through the use of adjoint-based derivatives, inexact Newton methods to determine the MAP point, and a Mat´ern class sparse prior precision operator. Together, these guarantee that the MAP point is found at a cost, measured in number of forward/adjoint poroelasticity solves, that is independent of the parameter dimension. We apply our methodology to a test case for a municipal well in Mesquite, Nevada, in which InSAR and GPS surface deformation data are available. We solve problems with up to 320,824 state variable degrees of freedom (DOFs) and 16,896 parameter DOFs. A consistent treatment of noise level is employed so that the aquifer characterization result does not depend on the pixel spacing of surface deformation data. Our results show that the use of InSAR data significantly improves characterization of lateral aquifer heterogeneity, and the InSAR-based aquifer characterization recovers complex lateral displacement trends observed by independent daily GPS measurements.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2020

Development of a Comparison Framework for Evaluating Environmental Contours of Extreme Sea States

Environmental contours of extreme sea states are often utilized for the purposes of reliability-based offshore design. Many methods have been proposed to estimate environmental contours of extreme sea states, including, but not limited to, the traditional inverse first-order reliability method (I-FORM) and subsequent modifications, copula methods, and Monte Carlo methods. These methods differ in terms of both the methodology selected for defining the joint distribution of sea state parameters and in the method used to construct the environmental contour from the joint distribution. It is often difficult to compare the results of proposed methods to determine which method should be used for a particular application or geographical region. The comparison of the predictions from various contour methods at a single site and across many sites is important to making environmental contours of extreme sea states useful in practice. The goal of this paper is to develop a comparison framework for evaluating methods for developing environmental contours of extreme sea states. This paper develops generalized metrics for comparing the performance of contour methods to one another across a collection of study sites, and applies these metrics and methods to develop conclusions about trends in the wave resource across geographic locations, as demonstrated for a pilot dataset. These proposed metrics and methods are intended to judge the environmental contours themselves relative to other contour methods, and are thus agnostic to a specific device, structure, or field of application. The metrics developed and applied in this paper include measures of predictive accuracy, physical validity, and aggregated temporal performance that can be used to both assess contour methods and provide recommendations for the use of certain methods in various geographical regions. The application and aggregation of the metrics proposed in this paper outline a comparison framework for environmental contour methods that can be applied to support design analysis workflows for offshore structures. This comparison framework could be extended in future work to include additional metrics of interest, potentially including those to address issues pertinent to a specific application area or analysis discipline, such as metrics related to structural response across contour methods or additional physics-based metrics based on wave dynamics.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Artificial intelligence driven laser parameter search: Inverse design of photonic surfaces using greedy surrogate-based optimization

Photonic surfaces designed with specific optical characteristics are becoming increasingly crucial for novel energy harvesting and storage systems. The design of these surfaces can be achieved by texturing materials using lasers. The optimal adjustment of laser fabrication parameters to achieve target surface optical properties is an open challenge. Thus, we develop a surrogate-based optimization approach. Our framework employs the Random Forest algorithm to model the forward relationship between the laser fabrication parameters and the resulting optical characteristics. During the optimization process, we use a greedy, prediction-based exploration strategy that iteratively selects batches of laser parameters to be used in experimentation by minimizing the predicted discrepancy between the surrogate model’s outputs and the user-defined target optical characteristics. This strategy allows for efficient identification of optimal fabrication parameters without the need to model the error landscape directly. We demonstrate the efficiency and effectiveness of our approach on two synthetic benchmarks and two specific experimental applications of photonic surface inverse design targets. By calculating the average performance of our algorithm compared to other state of the art optimization methods, we show that our algorithm performs, on average, twice as well across all benchmarks. Additionally, a warm starting inverse design technique for changed target optical characteristics enhances the performance of the introduced approach.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Quasielastic least-squares reverse time migration of PS reflections

Conventional analysis of amplitude variation with offset for elastic PS reflections is based on analytical reflection coefficients in a layered medium, and wave-equation-based PS migration is mainly used to produce a structural image. To overcome this problem, we have developed a least-squares reverse time migration (LSRTM) method for elastic PS reflections based on a quasielastic wave equation. Here, the quasielastic wave equation can accurately model PS reflections with elastic amplitudes under the first-order Born approximation. Our LSRTM method inverts for perturbations of the S-wave velocity and density by minimizing the L 2 norm of the difference between recorded and predicted PS reflections modeled using a quasielastic wave equation. We refer to our new method as quasielastic LSRTM of PS reflections. Numerical tests on synthetic and field data indicate that our method can properly handle the amplitudes of elastic PS reflections and provides an accurate estimate of the perturbations of S-wave velocity and density. Extending the method to the 3D case is not straightforward and might require incorporating certain PS data processing techniques into the inversion itself.

58 GEOSCIENCES↗
DOE OSTI2021

Methods and devices for preventing computationally explosive calculations in a computer for model parameters distributed on a hierarchy of geometric simplices

A computer-implemented method of preventing computationally explosive calculations. The method includes obtaining, by a processor of the computer, measured data of one of a physical process or a physical object; performing hierarchical numerical modeling of a physical process inclusive of an Earth model containing at least one of (a) infrastructure in the ground and (b) a formation feature in the ground, wherein predicted data is generated; comparing the measured data to the predicted data to calculate an estimated error; analyzing the estimated error via an inversion process to update the at least one of the Earth model and infrastructure model so as to reduce the estimated error and to determine a final composite Earth model of at least one of the infrastructure and the feature; and using the final composite Earth model to characterize at least one of the process and the physical object.

Source record↗
DOE OSTI2026

A new interatomic potential for mixed Mg-Al-Ga-In spinels

While density functional theory (DFT) has become the de facto approach for accurate simulation of materials at the atomic scale, there are many aspects of materials that are simply out of reach of DFT methods. In particular, finite temperature properties such as diffusivities, the structure and properties of grain boundaries and interfaces, and the study of defect properties in complex alloys are computationally challenging for DFT methods. Recently, a new class of spinels in which three cations order over two sublattices was discovered. In order to predict the properties of these types of structures, classical potentials are a must. Here, in this work, we derive a new classical potential for Mg-bearing spinels in which the B cations are Al, Ga, and/or In. The potential does well in describing the DFT energetics of various spinel structures as a function of chemistry and inversion. In particular, it reproduces the thermodynamically favorable MgAlGaO 4 structure while correctly predicting that neither MgAlInO 4 nor MgGaInO 4 are stable. Further, it reproduces physical trends in elastic properties as compared against experiment.

36 MATERIALS SCIENCE↗
DOE OSTI2020

Quadrupole topological photonic crystals

Quadrupole topological phases, exhibiting protected boundary states that are themselves topological insulators of lower dimensions, have recently been of great interest. Extensions of these ideas from current tight binding models to continuum theories for realistic materials require the identification of quantized invariants describing the bulk quadrupole order. Here we identify the analog of quadrupole order in Maxwell’s equations for a gyromagnetic photonic crystal (PhC) through a double-band-inversion process. The quadrupole moment is quantized by the simultaneous presence of crystalline symmetry and broken time-reversal symmetry, which is confirmed using three independent methods: analysis of symmetry eigenvalues, numerical calculations of the nested Wannier bands and the expectation value of the quadrupole operator. Furthermore, we reveal the boundary manifestations of quadrupole phases as quantized edge polarizations and fractional corner charges. The latter are the consequence of a filling anomaly of energy bands as first predicted in electronic systems.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2020

Learning and meta-learning of stochastic advection–diffusion–reaction systems from sparse measurements

Physics-informed neural networks (PINNs) were recently proposed in [18] as an alternative way to solve partial differential equations (PDEs). A neural network (NN) represents the solution, while a PDE-induced NN is coupled to the solution NN, and all differential operators are treated using automatic differentiation. Here, we first employ the standard PINN and a stochastic version, sPINN, to solve forward and inverse problems governed by a non-linear advection–diffusion–reaction (ADR) equation, assuming we have some sparse measurements of the concentration field at random or pre-selected locations. Subsequently, we attempt to optimise the hyper-parameters of sPINN by using the Bayesian optimisation method (meta-learning) and compare the results with the empirically selected hyper-parameters of sPINN. In particular, for the first part in solving the inverse deterministic ADR, we assume that we only have a few high-fidelity measurements, whereas the rest of the data is of lower fidelity. Hence, the PINN is trained using a composite multi-fidelity network, first introduced in [12], that learns the correlations between the multi-fidelity data and predicts the unknown values of diffusivity, transport velocity and two reaction constants as well as the concentration field. For the stochastic ADR, we employ a Karhunen–Loève (KL) expansion to represent the stochastic diffusivity, and arbitrary polynomial chaos (aPC) to represent the stochastic solution. Correspondingly, we design multiple NNs to represent the mean of the solution and learn each aPC mode separately, whereas we employ a separate NN to represent the mean of diffusivity and another NN to learn all modes of the KL expansion. For the inverse problem, in addition to stochastic diffusivity and concentration fields, we also aim to obtain the (unknown) deterministic values of transport velocity and reaction constants. The available data correspond to 7spatial points for the diffusivity and 20 space–time points for the solution, both sampled 2000 times. We obtain good accuracy for the deterministic parameters of the order of 1–2% and excellent accuracy for the mean and variance of the stochastic fields, better than three digits of accuracy. In the second part, we consider the previous stochastic inverse problem, and we use Bayesian optimisation to find five hyper-parameters of sPINN, namely the width, depth and learning rate of two NNs for learning the modes. Here, we obtain much deeper and wider optimal NNs compared to the manual tuning, leading to even better accuracy, i.e., errors less than 1% for the deterministic values, and about an order of magnitude less for the stochastic fields.

97 MATHEMATICS AND COMPUTING↗