Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “randomized algorithms”

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 1,117 records · Page 62
All sourcesNASA NTRSDOE OSTI
DOE OSTI2012

STAR-GALAXY CLASSIFICATION IN MULTI-BAND OPTICAL IMAGING

Ground-based optical surveys such as PanSTARRS, DES, and LSST will produce large catalogs to limiting magnitudes of r {approx}> 24. Star-galaxy separation poses a major challenge to such surveys because galaxies-even very compact galaxies-outnumber halo stars at these depths. We investigate photometric classification techniques on stars and galaxies with intrinsic FWHM <0.2 arcsec. We consider unsupervised spectral energy distribution template fitting and supervised, data-driven support vector machines (SVMs). For template fitting, we use a maximum likelihood (ML) method and a new hierarchical Bayesian (HB) method, which learns the prior distribution of template probabilities from the data. SVM requires training data to classify unknown sources; ML and HB do not. We consider (1) a best-case scenario (SVM{sub best}) where the training data are (unrealistically) a random sampling of the data in both signal-to-noise and demographics and (2) a more realistic scenario where training is done on higher signal-to-noise data (SVM{sub real}) at brighter apparent magnitudes. Testing with COSMOS ugriz data, we find that HB outperforms ML, delivering {approx}80% completeness, with purity of {approx}60%-90% for both stars and galaxies. We find that no algorithm delivers perfect performance and that studies of metal-poor main-sequence turnoff stars may be challenged by poor star-galaxy separation. Using the Receiver Operating Characteristic curve, we find a best-to-worst ranking of SVM{sub best}, HB, ML, and SVM{sub real}. We conclude, therefore, that a well-trained SVM will outperform template-fitting methods. However, a normally trained SVM performs worse. Thus, HB template fitting may prove to be the optimal classification method in future surveys.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Deep learning of parameterized equations with applications to uncertainty quantification

We propose a learning algorithm for discovering unknown parameterized dynamical systems by using observational data of the state variables. Our method is built upon and extends the recent work of discovering unknown dynamical systems, in particular those using deep neural network (DNN). We propose a DNN structure, largely based upon the residual network (ResNet), to not only learn the unknown form of the governing equation but also take into account the random effect embedded in the system, which is generated by the random parameters. Once the DNN model is successfully constructed, it is able to produce system prediction over longer term and for arbitrary parameter values. For uncertainty quantification, it allows us to conduct uncertainty analysis by evaluating solution statistics over the parameter space.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2011

Automated Planning of Tangential Breast Intensity-Modulated Radiotherapy Using Heuristic Optimization

Purpose: To present an automated technique for two-field tangential breast intensity-modulated radiotherapy (IMRT) treatment planning. Method and Materials: A total of 158 planned patients with Stage 0, I, and II breast cancer treated using whole-breast IMRT were retrospectively replanned using automated treatment planning tools. The tools developed are integrated into the existing clinical treatment planning system (Pinnacle{sup 3}) and are designed to perform the manual volume delineation, beam placement, and IMRT treatment planning steps carried out by the treatment planning radiation therapist. The automated algorithm, using only the radio-opaque markers placed at CT simulation as inputs, optimizes the tangential beam parameters to geometrically minimize the amount of lung and heart treated while covering the whole-breast volume. The IMRT parameters are optimized according to the automatically delineated whole-breast volume. Results: The mean time to generate a complete treatment plan was 6 min, 50 s {+-} 1 min 12 s. For the automated plans, 157 of 158 plans (99%) were deemed clinically acceptable, and 138 of 158 plans (87%) were deemed clinically improved or equal to the corresponding clinical plan when reviewed in a randomized, double-blinded study by one experienced breast radiation oncologist. In addition, overall the automated plans were dosimetrically equivalent to the clinical plans when scored for target coverage and lung and heart doses. Conclusion: We have developed robust and efficient automated tools for fully inversed planned tangential breast IMRT planning that can be readily integrated into clinical practice. The tools produce clinically acceptable plans using only the common anatomic landmarks from the CT simulation process as an input. We anticipate the tools will improve patient access to high-quality IMRT treatment by simplifying the planning process and will reduce the effort and cost of incorporating more advanced planning into clinical practice.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2021

Pre-conditioned BFGS-based uncertainty quantification in elastic full-waveform inversion

SUMMARY Full-waveform inversion has become an essential technique for mapping geophysical subsurface structures. However, proper uncertainty quantification is often lacking in current applications. In theory, uncertainty quantification is related to the inverse Hessian (or the posterior covariance matrix). Even for common geophysical inverse problems its calculation is beyond the computational and storage capacities of the largest high-performance computing systems. In this study, we amend the Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm to perform uncertainty quantification for large-scale applications. For seismic inverse problems, the limited-memory BFGS (L-BFGS) method prevails as the most efficient quasi-Newton method. We aim to augment it further to obtain an approximate inverse Hessian for uncertainty quantification in FWI. To facilitate retrieval of the inverse Hessian, we combine BFGS (essentially a full-history L-BFGS) with randomized singular value decomposition to determine a low-rank approximation of the inverse Hessian. Setting the rank number equal to the number of iterations makes this solution efficient and memory-affordable even for large-scale problems. Furthermore, based on the Gauss–Newton method, we formulate different initial, diagonal Hessian matrices as pre-conditioners for the inverse scheme and compare their performances in elastic FWI applications. We highlight our approach with the elastic Marmousi benchmark model, demonstrating the applicability of pre-conditioned BFGS for large-scale FWI and uncertainty quantification.

58 GEOSCIENCES↗
DOE OSTI2022

Deep learning methods for obtaining photometric redshift estimations from images

ABSTRACT Knowing the redshift of galaxies is one of the first requirements of many cosmological experiments, and as it is impossible to perform spectroscopy for every galaxy being observed, photometric redshift (photo-z) estimations are still of particular interest. Here, we investigate different deep learning methods for obtaining photo-z estimates directly from images, comparing these with ‘traditional’ machine learning algorithms which make use of magnitudes retrieved through photometry. As well as testing a convolutional neural network (CNN) and inception-module CNN, we introduce a novel mixed-input model that allows for both images and magnitude data to be used in the same model as a way of further improving the estimated redshifts. We also perform benchmarking as a way of demonstrating the performance and scalability of the different algorithms. The data used in the study comes entirely from the Sloan Digital Sky Survey (SDSS) from which 1 million galaxies were used, each having 5-filtre (ugriz) images with complete photometry and a spectroscopic redshift which was taken as the ground truth. The mixed-input inception CNN achieved a mean squared error (MSE) =0.009, which was a significant improvement ($30{{\ \rm per\ cent}}$) over the traditional random forest (RF), and the model performed even better at lower redshifts achieving a MSE = 0.0007 (a $50{{\ \rm per\ cent}}$ improvement over the RF) in the range of z < 0.3. This method could be hugely beneficial to upcoming surveys, such as Euclid and the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST), which will require vast numbers of photo-z estimates produced as quickly and accurately as possible.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2018

Characterizing quantum supremacy in near-term devices

A critical question for quantum computing in the near future is whether quantum devices without error correction can perform a well-defined computational task beyond the capabilities of supercomputers. Such a demonstration of what is referred to as quantum supremacy requires a reliable evaluation of the resources required to solve tasks with classical approaches. Here, we propose the task of sampling from the output distribution of random quantum circuits as a demonstration of quantum supremacy. We extend previous results in computational complexity to argue that this sampling task must take exponential time in a classical computer. We introduce cross-entropy benchmarking to obtain the experimental fidelity of complex multiqubit dynamics. This can be estimated and extrapolated to give a success metric for a quantum supremacy demonstration. We study the computational cost of relevant classical algorithms and conclude that quantum supremacy can be achieved with circuits in a two-dimensional lattice of 7 × 7 qubits and around 40 clock cycles. This requires an error rate of around 0.5% for two-qubit gates (0.05% for one-qubit gates), and it would demonstrate the basic building blocks for a fault-tolerant quantum computer.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Computing Nonequilibrium Responses with Score-Shifted Stochastic Differential Equations

Using equilibrium fluctuations to understand the response of a physical system to an externally imposed perturbation is the basis for linear response theory, which is widely used to interpret experiments and shed light on microscopic dynamics. For nonequilibrium systems, perturbations cannot be interpreted simply by monitoring fluctuations in a conjugate observable and general response results rely on path ensemble averaging. Furthermore, these techniques do not apply to perturbations that affect the diffusion tensor in a stochastic system. Here, we introduce an “effective” physical process that represents the diffusion perturbed dynamics and enables accurate calculations of responses to a change in the diffusion. Interestingly, the effective dynamics contain an additional drift involving the instantaneous “score” of the system, and we leverage score matching algorithms to carry out nonequilibrium response calculations on systems for which the exact stationary distribution is unknown.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2013

Experimental validation of the van Herk margin formula for lung radiation therapy

Purpose: To validate the van Herk margin formula for lung radiation therapy using realistic dose calculation algorithms and respiratory motion modeling. The robustness of the margin formula against variations in lesion size, peak-to-peak motion amplitude, tissue density, treatment technique, and plan conformity was assessed, along with the margin formula assumption of a homogeneous dose distribution with perfect plan conformity.Methods: 3DCRT and IMRT lung treatment plans were generated within the ORBIT treatment planning platform (RaySearch Laboratories, Sweden) on 4DCT datasets of virtual phantoms. Random and systematic respiratory motion induced errors were simulated using deformable registration and dose accumulation tools available within ORBIT for simulated cases of varying lesion sizes, peak-to-peak motion amplitudes, tissue densities, and plan conformities. A detailed comparison between the margin formula dose profile model, the planned dose profiles, and penumbra widths was also conducted to test the assumptions of the margin formula. Finally, a correction to account for imperfect plan conformity was tested as well as a novel application of the margin formula that accounts for the patient-specific motion trajectory.Results: The van Herk margin formula ensured full clinical target volume coverage for all 3DCRT and IMRT plans of all conformities with the exception of small lesions in soft tissue. No dosimetric trends with respect to plan technique or lesion size were observed for the systematic and random error simulations. However, accumulated plans showed that plan conformity decreased with increasing tumor motion amplitude. When comparing dose profiles assumed in the margin formula model to the treatment plans, discrepancies in the low dose regions were observed for the random and systematic error simulations. However, the margin formula respected, in all experiments, the 95% dose coverage required for planning target volume (PTV) margin derivation, as defined by the ICRU; thus, suitable PTV margins were estimated. The penumbra widths calculated in lung tissue for each plan were found to be very similar to the 6.4 mm value assumed by the margin formula model. The plan conformity correction yielded inconsistent results which were largely affected by image and dose grid resolution while the trajectory modified PTV plans yielded a dosimetric benefit over the standard internal target volumes approach with up to a 5% decrease in the V20 value.Conclusions: The margin formula showed to be robust against variations in tumor size and motion, treatment technique, plan conformity, as well as low tissue density. This was validated by maintaining coverage of all of the derived PTVs by 95% dose level, as required by the formal definition of the PTV. However, the assumption of perfect plan conformity in the margin formula derivation yields conservative margin estimation. Future modifications to the margin formula will require a correction for plan conformity. Plan conformity can also be improved by using the proposed trajectory modified PTV planning approach. This proves especially beneficial for tumors with a large anterior–posterior component of respiratory motion.

61 RADIATION PROTECTION AND DOSIMETRY↗
DOE OSTI2015

Reexamining classical and quantum models for the D-Wave One processor: The role of excited states and ground state degeneracy

We revisit the evidence for quantum annealing in the D-Wave One device (DW1) based on the study of random Ising instances. Using the probability distributions of finding the ground states of such instances, previous work found agreement with both simulated quantum annealing (SQA) and a classical rotor model. Thus the DW1 ground state success probabilities are consistent with both models, and a different measure is needed to distinguish the data and the models. Here in this paper we consider measures that account for ground state degeneracy and the distributions of excited states, and present evidence that for these new measures neither SQA nor the classical rotor model correlate perfectly with the DW1 experiments. We thus provide evidence that SQA and the classical rotor model, both of which are classically efficient algorithms, do not satisfactorily explain all the DW1 data. A complete model for the DW1 remains an open problem. Using the same criteria we find that, on the other hand, SQA and the classical rotor model correlate closely with each other. To explain this we show that the rotor model can be derived as the semiclassical limit of the spin-coherent states path integral. We also find differences in which set of ground states is found by each method, though this feature is sensitive to calibration errors of the DW1 device and to simulation parameters.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2017

Computational Redesign of Acyl-ACP Thioesterase with Improved Selectivity toward Medium-Chain-Length Fatty Acids

Enzyme and metabolic engineering offer the potential to develop biocatalysts for converting natural resources to a wide range of chemicals. To broaden the scope of potential products beyond natural metabolites, methods of engineering enzymes to accept alternative substrates and/or perform novel chemistries must be developed. DNA synthesis can create large libraries of enzyme-coding sequences, but most biochemistries lack a simple assay to screen for promising enzyme variants. Our solution to this challenge is structure-guided mutagenesis, in which optimization algorithms select the best sequences from libraries based on specified criteria (i.e., binding selectivity). We demonstrate this approach by identifying medium-chain (C8–C12) acyl-ACP thioesterases through structure-guided mutagenesis. Medium-chain fatty acids, which are products of thioesterase-catalyzed hydrolysis, are limited in natural abundance, compared to long-chain fatty acids; the limited supply leads to high costs of C6–C10 oleochemicals such as fatty alcohols, amines, and esters. Here, we applied computational tools to tune substrate binding of the highly active ‘TesA thioesterase in Escherichia coli. We used the IPRO algorithm to design thioesterase variants with enhanced C12 or C8 specificity, while maintaining high activity. After four rounds of structure-guided mutagenesis, we identified 3 variants with enhanced production of dodecanoic acid (C12) and 27 variants with enhanced production of octanoic acid (C8). The top variants reached up to 49% C12 and 50% C8 while exceeding native levels of total free fatty acids. A comparably sized library created by random mutagenesis failed to identify promising mutants. The chain length-preference of ‘TesA and the best mutant were confirmed in vitro using acyl-CoA substrates. Molecular dynamics simulations, confirmed by resolved crystal structures, of ‘TesA variants suggest that hydrophobic forces govern ‘TesA substrate specificity. Finally, we expect the design rules that we uncovered and the thioesterase variants that we identified will be useful to metabolic engineering projects aimed at sustainable production of medium-chain-length oleochemicals.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2014

High Energy Electromagnetic Particle Transportation on the GPU

We present massively parallel high energy electromagnetic particle transportation through a finely segmented detector on a Graphics Processing Unit (GPU). Simulating events of energetic particle decay in a general-purpose high energy physics (HEP) detector requires intensive computing resources, due to the complexity of the geometry as well as physics processes applied to particles copiously produced by primary collisions and secondary interactions. The recent advent of hardware architectures of many-core or accelerated processors provides the variety of concurrent programming models applicable not only for the high performance parallel computing, but also for the conventional computing intensive application such as the HEP detector simulation. The components of our prototype are a transportation process under a non-uniform magnetic field, geometry navigation with a set of solid shapes and materials, electromagnetic physics processes for electrons and photons, and an interface to a framework that dispatches bundles of tracks in a highly vectorized manner optimizing for spatial locality and throughput. Core algorithms and methods are excerpted from the Geant4 toolkit, and are modified and optimized for the GPU application. Program kernels written in C/C++ are designed to be compatible with CUDA and OpenCL and with the aim to be generic enough for easy porting to future programming models and hardware architectures. To improve throughput by overlapping data transfers with kernel execution, multiple CUDA streams are used. Issues with floating point accuracy, random numbers generation, data structure, kernel divergences and register spills are also considered. Performance evaluation for the relative speedup compared to the corresponding sequential execution on CPU is presented as well.

Canal, P.↗
DOE OSTI2015

SU-E-T-106: An Institutional Review of Using Commercially Available Software to Evaluate Treatment Plan Quality for Various Treatment Sites and Beam Deliveries

Purpose: Use Sun Nuclear Quality Reports™ with PlanIQ™ to evaluate different treatment delivery techniques for various treatment sites. Methods: Fifteen random patients with different treatment sites were evaluated. These include the Head/Neck, prostate, pelvis, lung, esophagus, axilla, bladder and abdomen. Initially, these sites were planned on the Pinnacle {sup 3} V9.6 treatment planning system and utilized nine 6MV step-n-shoot IMRT fields. The RT plan, dose and structure sets were sent to Quality Reports™ where a DVH was recreated and the plans were compared to a unique Plan Algorithm for each treatment site. Each algorithm has its own plan quality metrics and objectives, which include the PTV coverage, PTV maximum dose, the prescription dose outside the target, doses to the critical structures, and the global maximum dose and its location. Each plan was scored base on meeting each objective. Plans may have been reoptimized and reevaluated with Quality Reports™ based on the initial score. PlanIQ™ was used to evaluate if any objective not met was achievable or difficult to obtain. A second plan using VMAT delivery was created for each patient and scored with Quality Reports™. Results: There were a wide range of scores for the different treatment sites with some scoring better for IMRT plans and some better for the VMAT deliveries. The variation in the scores could be attributed to the treatment site, location, and shape of the target. Most deliveries were chosen for the VMAT due to the short treatment times and quick patient throughput with acceptable plan scores. Conclusion: The tools are provided for both physician and dosimetrist to objectively evaluate the use of VMAT delivery versus the step-n-shoot IMRT delivery for various sites. PlanIQ validates if objectives can be met. For the physicist, a concise pass/fail report is created for plan evaluation.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Bias-Variance Trade-Off in Physics-Informed Neural Networks with Randomized Smoothing for High-Dimensional PDEs

Physics-Informed Neural Networks (PINNs) have triggered a paradigm shift in scientific computing, leveraging mesh-free properties and robust approximation capabilities. While proving effective for low-dimensional partial differential equations (PDEs), the computational cost of PINNs remains a hurdle in high-dimensional scenarios. This is particularly pronounced when computing high-order and high-dimensional derivatives in the physics-informed loss. Randomized Smoothing PINN (RS-PINN) introduces Gaussian noise for stochastic smoothing of the original neural net model, enabling the use of Monte Carlo methods for derivative approximation, which eliminates the need for costly automatic differentiation. Despite its computational efficiency, especially in the approximation of high-dimensional derivatives, RS-PINN introduces biases in both loss and gradients, negatively impacting convergence, especially when coupled with stochastic gradient descent (SGD) algorithms. We present a comprehensive analysis of biases in RS-PINN, attributing them to the nonlinearity of the Mean Squared Error (MSE) loss as well as the intrinsic nonlinearity of the PDE itself. We propose tailored bias correction techniques, delineating their application based on the order of PDE nonlinearity. The derivation of an unbiased RS-PINN allows for a detailed examination of its advantages and disadvantages compared to the biased version. Specifically, the biased version has a lower variance and runs faster than the unbiased version, but it is less accurate due to the bias. To optimize the bias-variance trade-off, we combine the two approaches in a hybrid method that balances the rapid convergence of the biased version with the high accuracy of the unbiased version. In addition to methodological contributions, we present an enhanced implementation of RS-PINN. Extensive experiments on diverse high-dimensional PDEs, including Fokker-Planck, Hamilton-Jacobi-Bellman (HJB), viscous Burgers’, Allen-Cahn, and Sine-Gordon equations, illustrate the bias-variance trade-off and highlight the effectiveness of the hybrid RS-PINN. Empirical guidelines are provided for selecting biased, unbiased, or hybrid versions, depending on the dimensionality and nonlinearity of the specific PDE problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2014

MO-C-17A-03: A GPU-Based Method for Validating Deformable Image Registration in Head and Neck Radiotherapy Using Biomechanical Modeling

Purpose: Deformable image registration (DIR) plays a pivotal role in head and neck adaptive radiotherapy but a systematic validation of DIR algorithms has been limited by a lack of quantitative high-resolution groundtruth. We address this limitation by developing a GPU-based framework that provides a systematic DIR validation by generating (a) model-guided synthetic CTs representing posture and physiological changes, and (b) model-guided landmark-based validation. Method: The GPU-based framework was developed to generate massive mass-spring biomechanical models from patient simulation CTs and contoured structures. The biomechanical model represented soft tissue deformations for known rigid skeletal motion. Posture changes were simulated by articulating skeletal anatomy, which subsequently applied elastic corrective forces upon the soft tissue. Physiological changes such as tumor regression and weight loss were simulated in a biomechanically precise manner. Synthetic CT data was then generated from the deformed anatomy. The initial and final positions for one hundred randomly-chosen mass elements inside each of the internal contoured structures were recorded as ground truth data. The process was automated to create 45 synthetic CT datasets for a given patient CT. For instance, the head rotation was varied between +/− 4 degrees along each axis, and tumor volumes were systematically reduced up to 30%. Finally, the original CT and deformed synthetic CT were registered using an optical flow based DIR. Results: Each synthetic data creation took approximately 28 seconds of computation time. The number of landmarks per data set varied between two and three thousand. The validation method is able to perform sub-voxel analysis of the DIR, and report the results by structure, giving a much more in depth investigation of the error. Conclusions: We presented a GPU based high-resolution biomechanical head and neck model to validate DIR algorithms by generating CT equivalent 3D volumes with simulated posture changes and physiological regression.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2020

Single-frame far-field diffractive imaging with randomized illumination

We introduce a single-frame diffractive imaging method called randomized probe imaging (RPI). In RPI, a sample is illuminated by a structured probe field containing speckles smaller than the sample’s typical feature size. Quantitative amplitude and phase images are then reconstructed from the resulting far-field diffraction pattern. The experimental geometry of RPI is straightforward to implement, requires no near-field optics, and is applicable to extended samples. When the resulting data are analyzed with a complimentary algorithm, reliable reconstructions which are robust to missing data are achieved. To realize these benefits, a resolution limit associated with the numerical aperture of the probe-forming optics is imposed. RPI therefore offers an attractive modality for quantitative X-ray phase imaging when temporal resolution and reliability are critical but spatial resolution in the tens of nanometers is sufficient. We discuss the method, introduce a reconstruction algorithm, and present two proof-of-concept experiments: one using visible light, and one using soft X-rays.

77 NANOSCIENCE AND NANOTECHNOLOGY↗
DOE OSTI2024

GrainGNN: A dynamic graph neural network for predicting 3D grain microstructure

We propose GrainGNN, a surrogate model for the evolution of polycrystalline grain structure under rapid solidification conditions in metal additive manufacturing. High fidelity simulations of solidification microstructures are typically performed using multicomponent partial differential equations (PDEs) with moving interfaces. The inherent randomness of the PDE initial conditions (grain seeds) necessitates ensemble simulations to predict microstructure statistics, e.g., grain size, aspect ratio, and crystallographic orientation. Here, currently such ensemble simulations are prohibitively expensive and surrogates are necessary.In GrainGNN, we use a dynamic graph to represent interface motion and topological changes due to grain coarsening. We use a reduced representation of the microstructure using hand-crafted features; we combine pattern finding and altering graph algorithms with two neural networks, a classifier (for topological changes) and a regressor (for interface motion). Both networks have an encoder-decoder architecture; the encoder has a multi-layer transformer long-short-term-memory architecture; the decoder is a single layer perceptron.We evaluate GrainGNN by comparing it to high-fidelity phase field simulations for in-distribution and out-of-distribution grain configurations for solidification under laser power bed fusion conditions. GrainGNN results in 80%–90% pointwise accuracy; and nearly identical distributions of scalar quantities of interest (QoI) between phase field and GrainGNN simulations compared using Kolmogorov-Smirnov test. GrainGNN's inference speedup (PyTorch on single x86 CPU) over a high-fidelity phase field simulation (CUDA on a single NVIDIA A100 GPU) is 150×–2000× for 100-initial grain problem. Further, using GrainGNN, we model the formation of 11,600 grains in 220 seconds on a single CPU core.

36 MATERIALS SCIENCE↗
DOE OSTI2022

Global Sensitivity Analysis Using the Ultra‐Low Resolution Energy Exascale Earth System Model

Abstract For decades, Arctic temperatures have increased twice as fast as average global temperatures. As a first step toward quantifying parametric uncertainty in Arctic climate, we performed a variance‐based global sensitivity analysis (GSA) using a fully coupled, ultra‐low resolution (ULR) configuration of version 1 of the U.S. Department of Energy's Energy Exascale Earth System Model (E3SMv1). Specifically, we quantified the sensitivity of six quantities of interests (QOIs), which characterize changes in Arctic climate over a 75 year period, to uncertainties in nine model parameters spanning the sea ice, atmosphere, and ocean components of E3SMv1. Sensitivity indices for each QOI were computed with a Gaussian process emulator using 139 random realizations of the random parameters and fixed preindustrial forcing. Uncertainties in the atmospheric parameters in the Cloud Layers Unified by Binormals (CLUBB) scheme were found to have the most impact on sea ice status and the larger Arctic climate. Our results demonstrate the importance of conducting sensitivity analyses with fully coupled climate models. The ULR configuration makes such studies computationally feasible today due to its low computational cost. When advances in computational power and modeling algorithms enable the tractable use of higher‐resolution models, our results will provide a baseline that can quantify the impact of model resolution on the accuracy of sensitivity indices. Moreover, the confidence intervals provided by our study, which we used to quantify the impact of the number of model evaluations on the accuracy of sensitivity estimates, have the potential to inform the computational resources needed for future sensitivity studies.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2026

Two datasets are better than one: method of double moments for 3D reconstruction in cryo-EM

Cryo-electron microscopy is a powerful imaging technique for reconstructing three-dimensional molecular structures from noisy tomographic projection images of randomly oriented particles. We introduce a new data fusion framework, termed the method of double moments, which reconstructs molecular structures from two instances of the second-order moment of projection images obtained under distinct orientation distributions: one uniform, the other non-uniform and unknown. We prove that these moments generically uniquely determine the underlying structure, up to a global rotation and reflection, and we develop a convex-relaxation-based algorithm that achieves accurate recovery using only second-order statistics. Our results demonstrate the advantage of collecting and modeling multiple datasets under different experimental conditions, illustrating that leveraging dataset diversity can substantially enhance reconstruction quality in computational imaging tasks.

Kam’s method↗