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 541 records · Page 30

Towards automating structural discovery in scanning transmission electron microscopy *

Abstract Scanning transmission electron microscopy is now the primary tool for exploring functional materials on the atomic level. Often, features of interest are highly localized in specific regions in the material, such as ferroelectric domain walls, extended defects, or second phase inclusions. Selecting regions to image for structural and chemical discovery via atomically resolved imaging has traditionally proceeded via human operators making semi-informed judgements on sampling locations and parameters. Recent efforts at automation for structural and physical discovery have pointed towards the use of ‘active learning’ methods that utilize Bayesian optimization with surrogate models to quickly find relevant regions of interest. Yet despite the potential importance of this direction, there is a general lack of certainty in selecting relevant control algorithms and how to balance a priori knowledge of the material system with knowledge derived during experimentation. Here we address this gap by developing the automated experiment workflows with several combinations to both illustrate the effects of these choices and demonstrate the tradeoffs associated with each in terms of accuracy, robustness, and susceptibility to hyperparameters for structural discovery. We discuss possible methods to build descriptors using the raw image data and deep learning based semantic segmentation, as well as the implementation of variational autoencoder based representation. Furthermore, each workflow is applied to a range of feature sizes including NiO pillars within a La:SrMnO 3 matrix, ferroelectric domains in BiFeO 3 , and topological defects in graphene. The code developed in this manuscript is open sourced and will be released at github.com/nccreang/AE_Workflows .

47 OTHER INSTRUMENTATION↗

Practical and Optimal Sequential Bayesian Experimental Design for Complex Systems Incorporating Human Experimenter Preferences (Final Scientific/Technical Report)

Experiments are indispensable for developing models of complex systems. Carefully designed experiments can provide substantial savings for these expensive data-acquisition opportunities. However, designs based on heuristics are often suboptimal for systems with multiphysics, nonlinear dynamics, and uncertain and noisy environments. Optimal experimental design, while leveraging predictive models, seeks to systematically quantify and maximize the value of experiments. In this project, we focused on the design of multiple experiments, where current approaches are largely suboptimal: batch-design does not adapt to new data acquired during the experiment campaign (no feedback), and greedy/myopic design ignores future dynamics and consequences (no lookahead). We developed the mathematical framework and computational methods for sequential optimal experimental design (sOED) for complex systems. We enabled tractable model-based sOED in a rigorous manner through novel algorithms based on reinforcement learning, and investigated the effects of human experimenters on the design process. Our methods are fully Bayesian, able to quantify and update uncertainty in a principled manner. The traits aimed by our approach—mathematical rigor and optimality, human effects and uncertainty quantification, computational practicality—are crucial for elevating the standards of artificial intelligence (AI) to support decision-making in scientific domains, and contribute toward trust and realistic adoption of AI in experimental design practice.

97 MATHEMATICS AND COMPUTING↗

Early Inference of Nuclear Technology-Directed Research Activities of Authors from Scientific Publications

Nuclear research articles can provide information about early nuclear proliferation indicators such as influential research entities and technology capability levels of a country, but detection of nuclear activities typically occurs after they have started. We investigate the extent to which nuclear research articles can be used to infer whether a research entity will acquire or develop a nuclear technology before it happens. Early detection of nuclear proliferation or technology development indicators from data is challenging due to partial observability, sparse and unlabeled information, and confounding signals from multiple concurrent activities. This paper presents the early detection problem as a sequential decision-making, goal inference problem, where the objective is to characterize and predict an individual’s, organization’s, or a country’s intent (unobserved goal-directed behavior) towards developing a nuclear capability from partially observed sequences of their research publications, using inverse reinforcement learning and Bayesian goal inference methods. A computational framework is presented, and its application demonstrated using 29,196 Scopus records for a case study related to a civil nuclear capability. The case study results serve as a proof-of-concept demonstration for inference of technology-directed research activity of authors who publish in the nuclear domain. The inference method, combined with advanced computing, may be used to assess and monitor activities pertaining to early developmental stages of a nuclear technology or capability, which in turn can help to identify and prioritize activities with nuclear proliferation potential for further investigation.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

The cosmic DANCe of Perseus

Context. Star-forming regions are excellent benchmarks for testing and validating theories of star formation and stellar evolution. The Perseus star-forming region, being one of the youngest (< 10 Myr), closest (280-320 pc), and most studied in the literature, is a fundamental benchmark. Aims. We aim to study the membership, phase-space structure, mass, and energy (kinetic plus potential) distribution of the Perseus star-forming region using public catalogues (Gaia, APOGEE, 2MASS, and Pan-STARRS). Methods. We used Bayesian methodologies that account for extinction to identify the Perseus physical groups in the phase-space, retrieve their candidate members, derive their properties (age, mass, 3D positions, 3D velocities, and energy), and attempt to reconstruct their origin. Results. We identify 1052 candidate members in seven physical groups (one of them new) with ages between 3 and 10 Myr, dynamical super-virial states, and large fractions of energetically unbounded stars. Their mass distributions are broadly compatible with that of Chabrier for masses ≳0.1 M ⊙ and do not show hints of over-abundance of low-mass stars in NGC 1333 with respect to IC 348. These groups’ ages, spatial structure, and kinematics are compatible with at least three generations of stars. Future work is still needed to clarify if the formation of the youngest was triggered by the oldest. Conclusions. The exquisite Gaia data complemented with public archives and mined with comprehensive Bayesian methodologies allow us to identify 31% more members than previous studies, discover a new physical group (Gorgophone: 7 Myr, 191 members, and 145 M ⊙ ), and confirm that the spatial, kinematic, and energy distributions of these groups support the hierarchical star formation scenario.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

SANE: strategic autonomous non-smooth exploration for multiple optima discovery in multi-modal and non-differentiable black-box functions

Both computational and experimental material discovery bring forth the challenge of exploring multidimensional and multimodal parameter spaces, such as phase diagrams of Hamiltonians with multiple interactions, composition spaces of combinatorial libraries, material structure image spaces, and molecular embedding spaces. Often these systems are black-boxes and time-consuming to evaluate, which resulted in strong interest towards active learning methods such as Bayesian optimization (BO). However, these systems are often noisy which make the black box function severely multi-modal and non-differentiable, where a vanilla BO can get overly focused near a single or faux optimum, deviating from the broader goal of scientific discovery. To address these limitations, here we developed Strategic Autonomous Non-Smooth Exploration (SANE) to facilitate an intelligent Bayesian optimized navigation with a proposed cost-driven probabilistic acquisition function to find multiple global and local optimal regions, avoiding the tendency to becoming trapped in a single optimum. To distinguish between a true and false optimal region due to noisy experimental measurements, a human (domain) knowledge driven dynamic surrogate gate is integrated with SANE. We implemented the gate-SANE into pre-acquired piezoresponse spectroscopy data of a ferroelectric combinatorial library with high noise levels in specific regions, and piezoresponse force microscopy (PFM) hyperspectral data. SANE demonstrated better performance than classical BO to facilitate the exploration of multiple optimal regions and thereby prioritized learning with higher coverage of scientific values in autonomous experiments. Our work showcases the potential application of this method to real-world experiments, where such combined strategic and human intervening approaches can be critical to unlocking new discoveries in autonomous research.

Biswas, Arpan [University of Tennessee, Knoxville,↗

Bayesian Spatially Varying Multi-Regularization Image Deblurring

Many scientific experiments such as those found in astronomy, geology, microbiology, and X-ray radiography require the use of high-energy instruments to capture images. Since blur and noise are inevitably present in any imaging system, the images must be \deblurred" to extract the full information content. Mathematically, image deblurring is an ill-posed inverse problem that requires regularization. The regularization, in turn, has a large effect on the deblurred image: different regularization strengths, and types, lead to drastically different reconstructions. Moreover, many images contain a mixture of smooth and sharp features which suggests the use of multi-regularization, i.e., varying the type of regularization (e.g. Tikhonov or total variation) across the image. We address these issues by formulating the image deblurring problem within a hierarchical Bayesian framework in which we spatially adapt the strength of the regularization and also vary the regularization type across the image. In this way, the image itself, along with corresponding regularization strength at each pixel, are described jointly by a posterior distribution which we can sample by Markov chain Monte Carlo (MCMC) methods. We illustrate our techniques on simplified test problems and apply them to high-energy X-ray images taken at the Nevada National Security Site. Numerical tests show that our new method is robustly applicable and increases the quality of the image reconstruction when compared to other (Bayesian) methods.

97 MATHEMATICS AND COMPUTING↗

A quantification of the effectiveness of EPID dosimetry and software-based plan verification systems in detecting incidents in radiotherapy

Purpose: Complex treatments in radiation therapy require robust verification in order to prevent errors that can adversely affect the patient. For this purpose, the authors estimate the effectiveness of detecting errors with a “defense in depth” system composed of electronic portal imaging device (EPID) based dosimetry and a software-based system composed of rules-based and Bayesian network verifications. Methods: The authors analyzed incidents with a high potential severity score, scored as a 3 or 4 on a 4 point scale, recorded in an in-house voluntary incident reporting system, collected from February 2012 to August 2014. The incidents were categorized into different failure modes. The detectability, defined as the number of incidents that are detectable divided total number of incidents, was calculated for each failure mode. Results: In total, 343 incidents were used in this study. Of the incidents 67% were related to photon external beam therapy (EBRT). The majority of the EBRT incidents were related to patient positioning and only a small number of these could be detected by EPID dosimetry when performed prior to treatment (6%). A large fraction could be detected by in vivo dosimetry performed during the first fraction (74%). Rules-based and Bayesian network verifications were found to be complimentary to EPID dosimetry, able to detect errors related to patient prescriptions and documentation, and errors unrelated to photon EBRT. Combining all of the verification steps together, 91% of all EBRT incidents could be detected. Conclusions: This study shows that the defense in depth system is potentially able to detect a large majority of incidents. The most effective EPID-based dosimetry verification is in vivo measurements during the first fraction and is complemented by rules-based and Bayesian network plan checking.

60 APPLIED LIFE SCIENCES↗

BAYESIAN ANALYSIS TO IDENTIFY NEW STAR CANDIDATES IN NEARBY YOUNG STELLAR KINEMATIC GROUPS

We present a new method based on a Bayesian analysis to identify new members of nearby young kinematic groups. The analysis minimally takes into account the position, proper motion, magnitude, and color of a star, but other observables can be readily added (e.g., radial velocity, distance). We use this method to find new young low-mass stars in the {beta} Pictoris and AB Doradus moving groups and in the TW Hydrae, Tucana-Horologium, Columba, Carina, and Argus associations. Starting from a sample of 758 mid-K to mid-M (K5V-M5V) stars showing youth indicators such as H{alpha} and X-ray emission, our analysis yields 214 new highly probable low-mass members of the kinematic groups analyzed. One is in TW Hydrae, 37 in {beta} Pictoris, 17 in Tucana-Horologium, 20 in Columba, 6 in Carina, 50 in Argus, 32 in AB Doradus, and the remaining 51 candidates are likely young but have an ambiguous membership to more than one association. The false alarm rate for new candidates is estimated to be 5% for {beta} Pictoris and TW Hydrae, 10% for Tucana-Horologium, Columba, Carina, and Argus, and 14% for AB Doradus. Our analysis confirms the membership of 58 stars proposed in the literature. Firm membership confirmation of our new candidates will require measurement of their radial velocity (predicted by our analysis), parallax, and lithium 6708 A equivalent width. We have initiated these follow-up observations for a number of candidates, and we have identified two stars (2MASSJ01112542+1526214, 2MASSJ05241914-1601153) as very strong candidate members of the {beta} Pictoris moving group and one strong candidate member (2MASSJ05332558-5117131) of the Tucana-Horologium association; these three stars have radial velocity measurements confirming their membership and lithium detections consistent with young age.

79 ASTRONOMY AND ASTROPHYSICS↗

Selecting the selector: Comparison of update rules for discrete global optimization

In this paper, we compare some well-known Bayesian global optimization methods in four distinct regimes, corresponding to high and low levels of measurement noise and to high and low levels of “quenched noise” (which term we use to describe the roughness of the function we are trying to optimize). We isolate the two stages of this optimization in terms of a “regressor,” which fits a model to the data measured so far, and a “selector,” which identifies the next point to be measured. Finally, the focus of this paper is to investigate the choice of selector when the regressor is well matched to the data.

97 MATHEMATICS AND COMPUTING↗

Autonomous convergence of STM control parameters using Bayesian optimization

Scanning tunneling microscopy (STM) is a widely used tool for atomic imaging of novel materials and their surface energetics. However, the optimization of the imaging conditions is a tedious process due to the extremely sensitive tip–surface interaction, thus limiting the throughput efficiency. In this paper, we deploy a machine learning (ML)-based framework to achieve optimal atomically resolved imaging conditions in real time. The experimental workflow leverages the Bayesian optimization (BO) method to rapidly improve the image quality, defined by the peak intensity in the Fourier space. The outcome of the BO prediction is incorporated into the microscope controls, i.e., the current setpoint and the tip bias, to dynamically improve the STM scan conditions. We present strategies to either selectively explore or exploit across the parameter space. As a result, suitable policies are developed for autonomous convergence of the control parameters. The ML-based framework serves as a general workflow methodology across a wide range of materials.

97 MATHEMATICS AND COMPUTING↗

PHOTOMETRIC REDSHIFTS AND SYSTEMATIC VARIATIONS IN THE SPECTRAL ENERGY DISTRIBUTIONS OF LUMINOUS RED GALAXIES FROM SDSS DR7

We describe the construction of a template set of spectral energy distributions (SEDs) for the estimation of photometric redshifts of luminous red galaxies (LRGs) with a Bayesian template fitting method. By examining the color properties of several publicly available SED sets within a redshift range of 0 < z {approx}< 0.5 and comparing them to Sloan Digital Sky Survey (SDSS) Data Release 7 data, we show that only some of the investigated SEDs approximately match the colors of the LRG data throughout the redshift range, however not at the quantitative level required for precise photometric redshifts. This is because the SEDs of galaxies evolve with time (and redshift) and because at fixed redshift the LRG colors have an intrinsic spread such that they cannot be matched by one SED only. We generate new SEDs by superposing model SEDs of composite stellar populations with a burst model, allowing both components to be reddened by dust, in order to match the data in five different redshift bins. We select a set of SEDs which represents the LRG data in color space within five redshift bins, thus defining our new SED template set for photometric redshift estimates. The results we obtain with the new template set and our Bayesian template fitting photometric redshift code (PhotoZ) are nearly unbiased, with a scatter of {sigma}{sub {Delta}z} = 0.027 (including outliers), a fraction of catastrophic outliers (|z{sub phot} - z{sub spec}|/(1 + z{sub spec}) > 0.15) of {eta} = 0.12%, and a normalized median absolute rest frame deviation (NMAD) of {sigma}{sub NMAD} = 1.48 Multiplication-Sign MAD = 0.017 for non-outliers. We show that templates that optimally describe the brightest galaxies (-24.5 {<=} M{sub R} {<=} -22.7) indeed vary from z = 0.1 to z = 0.5, consistent with aging of the stellar population. Furthermore, we find that templates that optimally describe galaxies at z < 0.1 strongly differ as a function of the absolute magnitude of the galaxies, indicating an increase in star formation activity for less luminous galaxies. Our findings based on the photometry of the SDSS LRGs and our SED template fitting are supported by comparison to the average SDSS LRG spectra in different luminosity and redshift bins.

79 ASTRONOMY AND ASTROPHYSICS↗

Assessing the Influence of Microphysical and Environmental Parameter Perturbations on Orographic Precipitation

Microphysical (MP) schemes contain parameters whose values can impact the amount and location of forecasted precipitation, and sensitivity is typically explored by perturbing one parameter at a time while holding the rest constant. Although much can be learned from these “one-at-a-time” studies, the results are limited as these methods do not allow for nonlinear interactions of multiple perturbed parameters. This study applies the Morris one-at-a-time (MOAT) method, a robust statistical tool allowing for simultaneous perturbation of numerous parameters, to explore orographic precipitation sensitivity to changes in microphysical and environmental parameters within an environment characteristic of an atmospheric river. Results show parameters associated with snow fall speed coefficient A s and density ρ s , ice-cloud water collection efficiency (ECI), rain accretion (WRA), relative humidity, zonal wind speed, and surface potential temperature cause the largest influence on simulated precipitation. MP and environmental parameter perturbations can cause precipitation changes of similar magnitude, but results vary by location on the mountain. Different environments are also tested, with A s being the most influential MP parameter regardless of environment. Fewer MP parameters influence precipitation in a faster-wind-speed environment, possibly due to the stronger dynamical forcing upwind and different wave dynamics downwind, compared to a slower-wind-speed environment. Finally, perturbing MP parameters within a single scheme can result in precipitation variations of similar magnitude compared to using entirely different microphysics schemes. MOAT results presented in this study have implications for Bayesian parameter estimation methods and stochastic parameterization within ensemble forecasting.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Scalable Algorithms for Inverse Problems With High-Dimensional Parameter Spaces

Inverse problems, which involve inferring unknown parameters from observed data, present significant computational challenges, especially in large-scale settings with high-dimensional unknown parameters and nonlinear relationships between the unknowns and observations. Bayesian inference provides an approach for addressing these problems, often relying on sequential sampling methods like Markov chain Monte Carlo (MCMC) to approximate the posterior distribution of the parameters. However, MCMC methods become computationally demanding as the dimensionality of the problem increases, particularly in large-scale systems where likelihood evaluations rely on solving partial differential equations (PDEs) on large spatial domains with finely resolved meshes. To overcome these limitations, recent advancements have focused on designing scalable computa tional techniques – for both PDE simulations and sampling strategies – to make Bayesian methods feasible for high-dimensional problems.

97 MATHEMATICS AND COMPUTING↗

Inference of reaction rate parameters based on summary statistics from experiments

Here, we present the results of an application of Bayesian inference and maximum entropy methods for the estimation of the joint probability density for the Arrhenius rate para meters of the rate coefficient of the H 2 /O 2 -mechanism chain branching reaction H + O 2 → OH + O. Available published data is in the form of summary statistics in terms of nominal values and error bars of the rate coefficient of this reaction at a number of temperature values obtained from shock-tube experiments. Our approach relies on generating data, in this case OH concentration profiles, consistent with the given summary statistics, using Approximate Bayesian Computation methods and a Markov Chain Monte Carlo procedure. The approach permits the forward propagation of parametric uncertainty through the computational model in a manner that is consistent with the published statistics. A consensus joint posterior on the parameters is obtained by pooling the posterior parameter densities given each consistent data set. To expedite this process, we construct efficient surrogates for the OH concentration using a combination of Pad'e and polynomial approximants. These surrogate models adequately represent forward model observables and their dependence on input parameters and are computationally efficient to allow their use in the Bayesian inference procedure. We also utilize Gauss-Hermite quadrature with Gaussian proposal probability density functions for moment computation resulting in orders of magnitude speedup in data likelihood evaluation. Despite the strong non-linearity in the model, the consistent data sets all res ult in nearly Gaussian conditional parameter probability density functions. The technique also accounts for nuisance parameters in the form of Arrhenius parameters of other rate coefficients with prescribed uncertainty. The resulting pooled parameter probability density function is propagated through stoichiometric hydrogen-air auto-ignition computations to illustrate the need to account for correlation among the Arrhenius rate parameters of one reaction and across rate parameters of different reactions.

97 MATHEMATICS AND COMPUTING↗

Improving Subsurface Stress Characterization for Carbon Dioxide Storage Projects by Incorporating Machine Learning Techniques

The overall objective of this project is to develop a framework for reliable characterization and prediction of the state of stress in the overburden and underburden (including the basement) in CO 2 storage reservoirs using machine learning and integrated geomechanics and geophysical methods. Specifically, we propose to develop workflow encompassing of technologies and/or methods to predict stress and pressure changes due to CO 2 injection in an active tertiary recovery site and their impacts on subtle fault activation, fractures and occurrence of microseismic events and compare responses to field observations. In this project, we anticipate using dataset from the Farnsworth field Unit (FWU) which is operated by Purdure Petroleum. A novel elastic-waveform VSP inversion technique will be used to estimate high-resolution spatial and temporal changes of elastic moduli in CO 2 storage reservoirs, which will be combined with velocity-stress relationship derived from laboratory tests to obtain subsurface pressure and stress. Clustered microseismic data will be jointly inverted for improved focal mechanisms. Least-squares reverse-time migration of microseismic waveform data will be performed to directly image fracture/fault zones. Additionally, a deep neural network machine learning technique with convolutional and recurrent layers will be used for learning the spectro-temporal structures in microseismic waveforms. The results of this geotechnical data analysis will be integrated to develop a high-resolution 3D mechanical earth model extending from the overburden sealing formations to the underburden including the basement. Mechanical properties will be derived through integration of mechanical logs, tests, available results from chemo-mechanical laboratory tests, and elastic inversion of seismic data using a combination of Bayesian and stochastic methods as well as machine learning technique. Failure features (faults/fractures) will be represented and/or modeled based on seismic and core data analysis. A transient hydrodynamic-geomechanical model will be developed through coupling with the calibrated FWU reservoir simulation model. The full physics coupled model will be used to train a reduced order proxy model using machine learning algorithm for estimating stress which will then be used with appropriate constitutive relationships and forward seismological models to simulate pressure changes and induced microseismicity. An advanced optimization framework will be developed to perform a history match to minimize error between field observations and simulated. The history matched proxy model will be verified against the full-physics equivalent. The field observations that will be used in the coupled model calibration process include pressure/stress inverted from VSP, moment magnitude from microseismic analysis, real time downhole pressure measurements, production and injection data. Parameter sensitivity and uncertainty analysis will be performed to characterize the impact of model parameter uncertainty on stress estimates. The proposed project will have significant impact on future field implementation of the proposed technology. Because the project field site is an ongoing CO 2 EOR development, the value of the new technology will be demonstrated in an operational context and evaluated as a viable risk mitigation strategy. Cost/benefit will be evaluated together with the various commercial incentives for CO 2 sequestration available to oil and gas operators. The extensive available dataset and ongoing data acquisition under the SWP Phase III work plan provides flexibility for investigation of multiple approaches and reduces technical risk.

58 GEOSCIENCES↗

Off-the-shelf deep learning is not enough, and requires parsimony, Bayesianity, and causality

Abstract Deep neural networks (‘deep learning’) have emerged as a technology of choice to tackle problems in speech recognition, computer vision, finance, etc. However, adoption of deep learning in physical domains brings substantial challenges stemming from the correlative nature of deep learning methods compared to the causal, hypothesis driven nature of modern science. We argue that the broad adoption of Bayesian methods incorporating prior knowledge, development of solutions with incorporated physical constraints and parsimonious structural descriptors and generative models, and ultimately adoption of causal models, offers a path forward for fundamental and applied research.

97 MATHEMATICS AND COMPUTING↗

Process damping identification using Bayesian learning and time domain simulation

Process damping can provide improved machining productivity by increasing the stability limit at low spindle speeds. However, existing methods for identifying process damping models experimentally require specialized setups and/or multiple cutting tests. While the phenomenon is well known, the modeling challenges limit pre-process parameter selection that leverages the potential increases in material removal rates. This paper proposes a physics-informed Bayesian method that can identify the cutting force and process damping models from a limited set of test cuts without requiring direct measurements of cutting force or vibration. The method uses time domain simulation to incorporate process damping and provide a basis for test selection. Here, new strategies for efficient sampling and dimensionality reduction are applied to lower computation time and minimize the effect of model error. The proposed method is demonstrated and the identified cutting and damping force coefficients are compared to values obtained using machining tests and least-squares fitting.

42 ENGINEERING↗

Differentiable Multiphysics Codes: A Breakthrough Technology for Simulation and Computing

This document summarizes the findings of a strategic planning exercise commissioned by the Weapons Simulation and Computing, Computational Physics (WSC/CP) program at the Lawrence Livermore National Laboratory (LLNL) in FY24. During the year, the committee met with multiple stakeholder communities to gather input, opinions, suggestions and concerns which have been incorporated throughout this document. The key findings from this exercise are summarized: • The development of multiphysics modelling and simulation (mod/sim) codes and software technologies, their deployment on exascale compute platforms, and their broad adoption across the NNSA is a major success of the Advanced Simulation and Computing (ASC) program and the Exascale Computing Project (ECP). Sustained investment in these core technologies is essential. • Today’s state of the art involves running ensembles of O(100K) simulations to perform uncertainty quantification (UQ) and design studies using multiple statistical methods such as Bayesian optimization to understand sensitivities of our models and explore parameterized design spaces. Even with exascale computing, we are practically limited to O(10) parameters in these studies since the number of simulations required to sample the space scales exponentially with the number of design parameters. • The data from these simulation ensembles is increasingly being used to train machine learned (ML) surrogates (or reduced order models, ROMs) which can then be used for optimization or real time design exploration. However, the trained surrogates are still limited in the number of parameters they can represent due to the sampling limitations previously noted. • Augmenting our suite of integrated multiphysics simulation codes, both current and emerging, with the ability to compute gradients (solution derivatives) of arbitrary simulation outputs with respect to (some or all) simulation inputs would be a breakthrough technology, opening the door to a new era of efficient and automated inverse design based on verified and validated mod/sim capabilities. • This capability, which we refer to as differentiable multiphysics codes (DMCs), would revolutionize both UQ and optimization studies by breaking the curse of dimensionality that presently limits our “gradient-free” ensemble based computing approach. A similar breakthrough occurred in the AI/ML community once the ability to compute gradients of arbitrary loss functions using back-propagation became commonplace. Gradient information from the multiphysics codes can also be used to dramatically improve the efficiency and scale of training of ML/ROM surrogates for rapid assessments. • Achieving this in our suite of codes will be a grand challenge, similar to the amount of effort that was required to transition from CPU to GPU computing. It will require buy-in from the entire WSC/CP program and beyond, including all integrated codes, physics and engineering models, third-party library dependencies and performance portability abstractions. It will also require investment in research and development of numerical methods for computing adjoints of coupled physics across multiple adaptively refined moving meshes and of stochastic (Monte Carlo) and mesh free (SPH) methods. • New software and numerical techniques, largely pioneered by the AI/ML community, make this feasible. Chief among these is automatic differentiation (AD), the ability to employ AD at point-wise locations in a physics calculation (instead of traditional black-box approaches) and the ability to perform “back-propagation in time” (or reverse mode AD) for non-linear partial differential equations (PDEs). Fundamentally, the conclusion of this strategic planning exercise is that the time is right to undertake a large scale effort in WSC, centered on the existing integrated codes, to continue the natural evolution of mod/sim in the age of AI/ML. Instead of attempting to replace mod/sim with purely data driven AI/ML models, we believe the key to success is to integrate AI/ML by building on top of the decades of hard-won knowledge and the verified/validated multiphysics modelling capability that is the hallmark of the ASC program.

97 MATHEMATICS AND COMPUTING↗