Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “constrained learning”

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 163 records · Page 9
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Investigating Fission Reaction Rate Ratio Sensitivities [Abstract]

Reaction rate ratios are a measurable parameter for reactor and criticality applications. A number of foil irradiations and fission chamber measurements have been performed for critical assemblies at Los Alamos National Laboratory starting in the 1950’s including (i) Godiva, a bare HEU spherical assembly; (ii) Flattop-25, a spherical assembly consisting of an HEU core and a natural uranium reflector; (iii) Jezebel, a bare 239 Pu assembly; and (iv) Flattop-Pu, a spherical assembly consisting of a 239Pu core and a natural uranium reflector. Fission ratio data for 238 U(n,f)/ 235 U(n,f), 237 Np(n,f)/ 235 U(n,f), 233 U(n,f)/ 235 U(n,f) and 239 Pu(n,f)/ 235 U(n,f) were obtained and reported. The EUCLID (Experiments Underpinned by Computational Learning for Improvements in nuclear Data) project at Los Alamos National Laboratory (LANL) aims to constrain nuclear data by using a suite of measurement types beyond k-effective. Recent investigations include the use of pulsed spheres for nuclear data validation and other measurement methods of interest. One focus of the work is to determine if other methods are complimentary to the critical experiments utilized for nuclear data validation. It is anticipated the investigations will help inform methods that may be utilized in machine learning algorithms for nuclear validation. In order to use a measurement type for nuclear validation, it is necessary to obtain cross-section sensitivities for parameters. This work looks at reaction rate ratio sensitivities with SENSMG and Monte Carlo N-Particle R Code Version 6.21.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2025

Extratropical Cloud Feedback Constrained by Cloud Sources and Sinks in Cyclones

Constraining cloud feedback in global climate models (GCMs) using observations is important for establishing accurate predictions of future climate. Uncertainty in shortwave cloud feedback (SW FB ) dominates uncertainty in total cloud feedback. Recent studies show a shift toward more positive extratropical SW FB in the latest generations of GCMs leading to the emergence of very high equilibrium climate sensitivity (ECS). In this study, we use precipitation efficiency and albedo susceptibility to constrain liquid water path (LWP) response to warming and SW FB in the Southern Ocean (SO; 50°–80°S). We analyze precipitation in extratropical cyclones (ECs) to learn about extratropical condensed water sink processes, combined with observations of clouds and moisture convergence, and use the analysis to better understand and constrain SW FB . We utilize a perturbed parameter ensemble (PPE) hosted in the Community Atmosphere Model, version 6 (CAM6), to provide a constraint on SW FB based on observations from Clouds and the Earth’s Radiant Energy System (CERES) and Multisensor Advanced Climatology of LWP (MAC-LWP). We apply Gaussian process regression to emulate the model response to all parameters perturbed in the PPE. Confronting the emulator output with observations provides a new estimated response of Earth to global warming. Furthermore, our new estimates of SO LWP reduce the PPE range by 66%–72%, which results in a shortwave cloud radiative effect estimated range that is 27%–34% less than the PPE range. Observations suggest a more positive SO SW FB than the Community Earth System Model, version 2 (CESM2), and consequently do not reject the high climate sensitivity GCMs emerging from the Coupled Model Intercomparison Project phase 6 (CMIP6).

Atmosphere↗
DOE OSTI2026

A generative machine learning model for designing metal hydrides applied to hydrogen storage

Developing new metal hydrides is a critical step toward efficient hydrogen storage in carbon-neutral energy systems. However, existing materials databases, such as the Materials Project, contain a limited number of well-characterized hydrides, which constrains the discovery of optimal candidates. This work presents a framework that integrates causal discovery with a lightweight generative machine learning model to generate novel metal hydride candidates that may not exist in current databases. Using a dataset of 450 samples (270 training, 90 validation, and 90 testing), the model generates 1000 candidates. After ranking and filtering, six previously unreported chemical formulas and crystal structures are identified, four of which are validated by density functional theory simulations and show strong potential for future experimental investigation. Overall, the proposed framework provides a scalable and time-efficient approach for expanding hydrogen storage datasets and accelerating materials discovery.

generative model↗
DOE OSTI2025

The Artificial Scientist: in-Transit Machine Learning of Plasma Simulations

Large-scale simulations or scientific experiments produce petabytes of data per run. This poses massive challenges for I/O and storage when scientific analysis workflows are run manually offline. Unsupervised deep learning-based techniques to extract patterns and non-linear relations from these large amounts of data provide a way to build scientific understanding from raw data, reducing the need for manual pre-selection of analysis steps, but require exascale compute and memory to process the full dataset available. In this paper, we demonstrate a heterogeneous streaming workflow in which plasma simulation data is streamed directly to a Machine Learning (ML) application training a model on the simulation data in-transit, completely circumventing the capacity-constrained filesystem bottleneck. This workflow employs openPMD to provide a high level interface to describe scientific data and also uses ADIOS2, to transfer volumes of data that exceed the capabilities of the filesystem. We employ experience replay to avoid catastrophic forgetting in learning from this non-steady state process in a continual manner and adapt it to improve model convergence while learning in-transit. As a proof-of-concept, we approach the ill-posed inverse problem of predicting particle dynamics from radiation in a particle-incell (PIConGPU) simulation of the Kelvin-Helmholtz instability (KHI). We detail hardware-software co-design challenges as we scale PIConGPU to full Frontier, the Top-1 system as of June 2024 Top500 list.

Kelling, Jeffrey [Helmholtz-Zentrum Dresden Rossen↗
DOE OSTI2024

Photometric redshifts probability density estimation from recurrent neural networks in the DECam local volume exploration survey data release 2

Photometric wide-field surveys are imaging the sky in unprecedented detail. These surveys face a significant challenge in efficiently estimating galactic photometric redshifts while accurately quantifying associated uncertainties. In this work, we address this challenge by exploring the estimation of Probability Density Functions (PDFs) for the photometric redshifts of galaxies across a vast area of 17,000 square degrees, encompassing objects with a median 5 σ point-source depth of g = 24.3, r = 23 . 9 , i = 23.5, and z = 22.8 mag. Our approach uses deep learning, specifically integrating a Recurrent Neural Network architecture with a Mixture Density Network, to leverage magnitudes and colors as input features for constructing photometric redshift PDFs across the whole DECam Local Volume Exploration (DELVE) survey sky footprint. Subsequently, we rigorously evaluate the reliability and robustness of our estimation methodology, gauging its performance against other well-established machine learning methods to ensure the quality of our redshift estimations. Our best results constrain photometric redshifts with the bias of − 0 . 0013 , a scatter of 0.0293, and an outlier fraction of 5.1%. These point estimates are accompanied by well-calibrated PDFs evaluated using diagnostic tools such as Probability Integral Transform and Odds distribution. We also address the problem of the accessibility of PDFs in terms of disk space storage and the time demand required to generate their corresponding parameters.We present a novel Autoencoder model that reduces the size of PDF parameter arrays to one-sixth of their original length, significantly decreasing the time required for PDF generation to one-eighth of the time needed when generating PDFs directly from the magnitudes.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2024

Coarse-Graining Conformational Dynamics with Multidimensional Generalized Langevin Equation: How, When, and Why

A data-driven ab initio generalized Langevin equation (AIGLE) approach is developed to learn and simulate high-dimensional, heterogeneous, coarse-grained (CG) conformational dynamics. Constrained by the fluctuation–dissipation theorem, the approach can build CG models in dynamical consistency (DC) with all-atom molecular dynamics. Here, we also propose practical criteria for AIGLE to enforce long-term DC. Case studies of a toy polymer, with 20 CG sites, and the alanine dipeptide, with two dihedral angles, elucidate why one should adopt AIGLE or its Markovian limit for modeling CG conformational dynamics in practice.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2023

EPiC-GAN: Equivariant point cloud generation for particle jets

With the vast data-collecting capabilities of current and future high-energy collider experiments, there is an increasing demand for computationally efficient simulations. Generative machine learning models enable fast event generation, yet so far these approaches are largely constrained to fixed data structures and rigid detector geometries. In this paper, we introduce EPiC-GAN - equivariant point cloud generative adversarial network - which can produce point clouds of variable multiplicity. This flexible framework is based on deep sets and is well suited for simulating sprays of particles called jets. The generator and discriminator utilize multiple EPiC layers with an interpretable global latent vector. Crucially, the EPiC layers do not rely on pairwise information sharing between particles, which leads to a significant speed-up over graph- and transformer-based approaches with more complex relation diagrams. We demonstrate that EPiC-GAN scales well to large particle multiplicities and achieves high generation fidelity on benchmark jet generation tasks.

Buhmann, Erik↗
DOE OSTI2026

Improving thermodynamic nudging in the E3SM Atmosphere Model version 2 (EAMv2): strategy and hindcast skills on weather systems

Nudging techniques are commonly employed to constrain atmospheric simulations toward observed states, facilitating model evaluation and sensitivity studies. However, if applied improperly – particularly to thermodynamic variables such as temperature and humidity – nudging can distort physical processes and introduce spurious biases, undermining the credibility of the simulations. This study presents an improved nudging implementation that applies vertically modulated tendencies to reduce adverse impacts on model physics. The framework is tested in version 2 of the Energy Exascale Earth System Model (EAMv2) using a suite of hindcast simulations nudged toward ERA5 reanalysis. We systematically evaluate the individual and combined effects of nudging wind, temperature, and humidity fields on the model's ability to represent large-scale atmospheric states and high-impact weather systems. Results show that the revised strategy – particularly when nudging temperature and humidity at selected levels – enhances hindcast skill by improving agreement with ERA5 without degrading the hydrological cycle or precipitation processes. Additional improvements in surface temperature, outgoing longwave radiation, and precipitation biases are achieved through targeted nudging of land-surface variables. The proposed approach strengthens the representation of large-scale conditions relevant to tropical cyclones, atmospheric rivers, and extratropical cyclones in the low-resolution EAMv2. These findings demonstrate that carefully designed thermodynamic nudging, especially of temperature and humidity, improves the realism of constrained simulations and broadens the utility of nudged EAMv2 for atmospheric modeling, machine learning, and high-impact weather research.

Atmospheric river↗
DOE OSTI2023

Automated Energy-Dispersive X-ray Spectroscopy Analysis for Multi-Modal Few-Shot Learning

Scanning transmission electron microscopy (STEM) is a powerful tool that allows for the atomic-scale analysis of a materials’ structure, chemistry, and defect domains (Akers et al. 2021). The current generation of microscopes generate vast amounts of data, surpassing the limits of effective manual analysis traditionally performed by domain experts (Spurgeon et al. 2021). While recent strides in machine learning have significantly enhanced the processing of large and intricate datasets acquired through electron microscopy, the prevalent use of proprietary software packages for initial data collection poses a challenge. In many cases, these software packages act as a ‘black box’, constraining user functionality and hindering the output of data in a format that is conducive to seamless integration into machine learning models. This work addresses these challenges by adapting HyperSpy, an open-source Python library, for the analysis and quantification of raw energy dispersive spectroscopy (EDS) data acquired through STEM. The modified HyperSpy code successfully facilitates user-defined segmentation of the data, enabling the integration of atomic %, weight %, and raw EDS spectra for each segmented region into an existing few-shot machine learning model. While initial results reveal discrepancies in quantified atomic and weight percentages when compared to proprietary software, ongoing efforts aim to rectify this issue by refining the fit of the HyperSpy model to the EDS spectra. Overall, this research underscores the potential of open-source tools like HyperSpy to enhance the accessibility of analytical tools, fostering a transparent and user-friendly environment for seamlessly incorporating electron microscopy data into machine learning models.

36 MATERIALS SCIENCE↗
DOE OSTI2021

Investigation of Delayed Neutron Sensitivities for Several ICSBEP Benchmarks using MCNP

The effective delayed neutron (β$_{eff}$) is a very important parameter for reactor and criticality applications. This parameter is equal to the difference in reactivity between delayed critical ($k_{eff}$ = 1, which requires both prompt and delayed neutrons to achieve criticality) and prompt critical ($k_p$ = 1, which requires only prompt neutrons to achieve criticality). This is often referred to as the delayed critical "window" (the region of criticality between delayed and prompt critical). β$_{eff}$ is a reactor kinetics parameters and depends on the nuclides in the system that undergo fission as well as the spectral characteristics of the system. Measurements of β$_{eff}$ have been performed for many criticality experiments. The EUCLID (Experiments Underpinned by Computational Learning for Improvements in nuclear Data) project at Los Alamos National Laboratory (LANL) aims to constrain nuclear data by using a suite of measurement types beyond $k_{eff}$. Our team has recently investigated the use of pulsed spheres for nuclear data validation. Several other measurement methods are also of interest, including β$_{eff}$ (investigated here) and reactivity coefficients (investigated in a separate work at this same meeting). One focus of our work is to determine if other methods are complimentary to the critical experiments already utilized for nuclear data validation. This is important because if a method has similar sensitivities then it will not be particularly useful for nuclear data validation as it will provide the same information as the critical experiments already being used. Here "similar" could refer to several characteristics, one being the shape of a sensitivity profile over energy. In the future, these methods will be utilized (with both existing and new experiments) in machine learning algorithms for nuclear validation, similar to what is currently done for criticality experiments. In order to use a measurement type for nuclear validation, it is necessary to obtain cross-section sensitivities for that parameter. This work looks at one approach to estimate β$_{eff}$ sensitivities by utilizing $k_{eff}$ sensitivities within Monte Carlo N-Particle ® Code Version 6.2. This is applied to several criticality benchmarks. Results are compared and the benefits and limitations of this approach are discussed.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

AEOLUS: Advances in Experimental Design, Optimal Control, and Learning for Uncertain Complex Systems

Sustained advances in the mathematics of modeling and simulation have resulted in the capability today for routine simulation of a number of large scale complex DOE-relevant systems. As remarkable as this capability for solving the so-called forward problem is, it is typically only the first step-an inner loop within an outer loop that explores the simulation model's parameter space and decision space to characterize uncertainty in the model's predictions, learn unknown model parameters from data, design the most informative experiments, determine optimal control strategies, and create optimal designs. Broadly, what unifies all of these outer loop problems is that they are, in one form or another, optimization problems over parameter/control/design space that are constrained by complex uncertain models. To fully realize the power of scientific simulation as a basis for scientific discovery, technological innovation, and rational decision-making, it is imperative to move beyond simulation to tackle the outer loop of optimization for learning from data, experimental design, and control with complex uncertain models. When the models under consideration are large-scale and complex, and when the optimization variable and uncertain parameter spaces are high (or infinite) dimensional, this constitutes a grand challenge of the highest order, and is intractable with conventional methods. To overcome these challenges, the AEOLUS Center was established to develop a unified mathematical, computational, and statistical framework for (1) Learning predictive models from complex data via Bayesian inference and optimization, and (2) Optimizing experiments, processes, and designs using the resulting uncertain models. These problems are intractable with conventional methods, for several reasons: (1) The simulation problems that govern the inner loops of the optimization problems are expensive to execute (due to severe nonlinearity, heterogeneity, multiphysics/multiscale coupling); (2) The optimization variable and uncertain parameter spaces are high dimensional, often stemming from discretizations of infinite dimensional fields such as initial conditions, sources, or material properties. We argue that the key to overcoming these challenges is to develop new mathematical, computational, and statistical methods that exploit the structure of the Bayesian inference and optimization problems mediated by their underlying complex uncertain models. This structure includes the regularity, sparsity, geometry, low intrinsic dimensionality, and multifidelity nature of the maps from uncertain parameter/optimization variable spaces to the specific objectives targeted: Bayesian inference, optimal experimental design, and optimal control design. Black box methods developed as generic tools are incapable of exploiting this structure. To be successful, we must create, integrate, and cross-fertilize ideas across multiple areas of applied math--including approximation theory, Bayesian inference, data science, experimental design, information theory, machine learning, model reduction, optimal control theory, parallel algorithms, PDE-constrained optimization, randomized algorithms, stochastic optimization, and uncertainty quantification--all while exploiting the structure of the problems at hand. With this goal in mind, we have marshaled a team of leading authorities in these areas. While the methods we develop will be broadly applicable across a wide spectrum of DOE problems in which experiments inform models and the systems those models describe must be optimized under uncertainty, we have chosen a specific area, advanced manufacturing and materials, to drive our work. AMM is characterized by complex models across multiple scales, and is a rich source of challenging problems in inference, experimental design, and optimal control, requiring multifaceted and integrated advances in applied mathematics. As such, AMM serves as an excellent vehicle to motivate and demonstrate the advances in applied mathematics developed by our center.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Accelerating defect predictions in semiconductors using graph neural networks

First-principles computations reliably predict the energetics of point defects in semiconductors but are constrained by the expense of using large supercells and advanced levels of theory. Machine learning models trained on computational data, especially ones that sufficiently encode defect coordination environments, can be used to accelerate defect predictions. Here, we develop a framework for the prediction and screening of native defects and functional impurities in a chemical space of group IV, III–V, and II–VI zinc blende semiconductors, powered by crystal Graph-based Neural Networks (GNNs) trained on high-throughput density functional theory (DFT) data. Using an innovative approach of sampling partially optimized defect configurations from DFT calculations, we generate one of the largest computational defect datasets to date, containing many types of vacancies, self-interstitials, anti-site substitutions, impurity interstitials and substitutions, as well as some defect complexes. We applied three types of established GNN techniques, namely crystal graph convolutional neural network, materials graph network, and Atomistic Line Graph Neural Network (ALIGNN), to rigorously train models for predicting defect formation energy (DFE) in multiple charge states and chemical potential conditions. We find that ALIGNN yields the best DFE predictions with root mean square errors around 0.3 eV, which represents a prediction accuracy of 98% given the range of values within the dataset, improving significantly on the state-of-the-art. We further show that GNN-based defective structure optimization can take us close to DFT-optimized geometries at a fraction of the cost of full DFT. The current models are based on the semi-local generalized gradient approximation-Perdew–Burke–Ernzerhof (PBE) functional but are highly promising because of the correlation of computed energetics and defect levels with higher levels of theory and experimental data, the accuracy and necessity of discovering novel metastable and low energy defect structures at the PBE level of theory before advanced methods could be applied, and the ability to train multi-fidelity models in the future with new data from non-local functionals. The DFT-GNN models enable prediction and screening across thousands of hypothetical defects based on both unoptimized and partially optimized defective structures, helping identify electronically active defects in technologically important semiconductors.

Rahman, Md Habibur (ORCID:000000027705984X)↗
DOE OSTI2025

JGI Plant Transformation Workshop, May 20-21, 2025

Domestic biomass crops such as sorghum, switchgrass, Miscanthus, and poplar can provide United States industries with renewable feedstocks while also supporting low-input farming systems and strengthening supply chains for biofuels, biochemicals and biomaterials. The U.S. leads globally in biomass crop genomics, yet progress in engineering traits is constrained by slow, genotype-dependent transformation methods and lengthy Design-Build-Test-Learn (DBTL) cycles. At a May 2025 workshop, a panel of experts recommended establishing a DOE Plant Transformation Capability (PTC) to overcome these barriers. The PTC would unite two missions: advancing research to achieve genotype-independent, automated methods, and delivering scalable transformation services through a user-facility model. With expected gains of 10–100x in efficiency, including transformation and cost reduction, the PTC would accelerate the path from discovery to engineered plants, expand community access and training, and support downstream applications and workflows including field trials and regulatory navigation. By enabling rapid and predictable crop engineering, the PTC would strengthen U.S. supply chains, enhance industrial competitiveness, and ensure that DOE’s genomic investments deliver national impact.

09 BIOMASS FUELS↗
DOE OSTI2024

Quantifying Drivers of Methane Hydrobiogeochemistry in a Tidal River Floodplain System

The influence of coastal ecosystems on global greenhouse gas (GHG) budgets and their response to increasing inundation and salinization remains poorly constrained. In this study, we have integrated an uncertainty quantification (UQ) and ensemble machine learning (ML) framework to identify and rank the most influential processes, properties, and conditions controlling methane behavior in a freshwater floodplain responding to recently restored seawater inundation. Our unique multivariate, multiyear, and multi-site dataset comprises tidal creek and floodplain porewater observations encompassing water level, salinity, pH, temperature, dissolved oxygen (DO), dissolved organic carbon (DOC), total dissolved nitrogen (TDN), partial pressure of carbon dioxide (pCO 2 ), nitrous oxide (pN 2 O), methane (pCH 4 ), and the stable isotopic composition of methane (δ 13 CH 4 ). Additionally, we incorporated topographical data, soil porosity, hydraulic conductivity, and water retention parameters for UQ analysis using a previously developed 3D variably saturated flow and transport floodplain model for a physical mechanistic understanding of factors influencing groundwater levels and salinity and, therefore, CH 4 . Principal component analysis revealed that groundwater level and salinity are the most significant predictors of overall biogeochemical variability. The ensemble ML models and UQ analyses identified DO, water level, salinity, and temperature as the most influential factors for porewater methane levels and indicated that approximately 80% of the total variability in hourly water levels and around 60% of the total variability in hourly salinity can be explained by permeability, creek water level, and two van Genuchten water retention function parameters: the air-entry suction parameter α and the pore size distribution parameter m. These findings provide insights on the physicochemical factors in methane behavior in coastal ecosystems and their representation in local- to global-scale Earth system models.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Leveraging transfer learning and leaf spectroscopy for leaf trait prediction with broad spatial, species, and temporal applicability

Accurate and reliable prediction of leaf traits is crucial for understanding plant adaptations to environmental variation, monitoring terrestrial ecosystems, and enhancing comprehension of functional diversity and ecosystem functioning. Currently, various approaches (e.g., statistical, physical models) have been developed to estimate leaf traits through hyperspectral remote sensing and leaf spectroscopy. However, the absence of high-performing, transferable, and stable models across various domains of space, plant functional types (PFTs) and seasons hinder our ability to quantify and comprehend spatiotemporal variations in leaf traits. This study proposes robust and highly transferable models for better predicting leaf traits with hyperspectral reflectance. Initially, three datasets were assembled, pairing common leaf traits — chlorophyll (Chla+b), carotenoids (Ccar), leaf mass per area (LAM), equivalent water thickness (EWT) — with leaf spectra measurements collected across diverse geographic locations in the U.S. and Europe, PFTs, and seasons. Measurements were acquired using spectroradiometers (e.g., ASD FieldSpec 3/4/Pro and SVC HR-1024i) with integrating spheres, leaf clips, and contact probes. Here, we then developed transfer learning-based hybrid models that incorporated the domain knowledge of radiative transfer models (RTMs) through pretraining processes and were well-constrained by fine-tuning with field measurements. Through comparison with other state-of-the-art statistical models, including partial-least squares regression (PLSR) and Gaussian Process Regression (GPR), as well as pure physical models, we found that the proposed transfer learning models achieved better predictive performance and higher transferability. Specifically, compared to other statistical models and pure RTMs, the transfer learning model exhibited higher coefficient of determination (R 2 ) values with range of 0.01 to 0.79, lower normalized root mean square error (NRMSE) with range of 0.06 % to 33.25 % in model performance. Additionally, the models exhibited improved transferability, with higher R 2 values range from 0.04 to 0.32, lower NRMSE range from 0.08 % to 30.81 %. The findings underscore that transfer learning models through integrating domain knowledge from RTMs and limited observations, can harness the advantages of both RTMs and statistical models and serve as a promising approach for effectively predicting leaf traits.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

A safe reinforcement learning algorithm for supervisory control of power plants

Traditional control theory-based methods require tailored engineering for each system and constant fine-tuning. In power plant control, one often needs to obtain a precise representation of the system dynamics and carefully design the control scheme accordingly. Model-free Reinforcement learning (RL) has emerged as a promising solution for control tasks due to its ability to learn from trial-and-error interactions with the environment. It eliminates the need for explicitly modeling the environment’s dynamics, which is potentially inaccurate. However, the direct imposition of state constraints in power plant control raises challenges for standard RL methods. To address this, we propose a chance-constrained RL algorithm based on Proximal Policy Optimization for supervisory control. Our method employs Lagrangian relaxation to convert the constrained optimization problem into an unconstrained objective, where trainable Lagrange multipliers enforce the state constraints. In conclusion, our approach achieves the smallest distance of violation and violation rate in a load-follow maneuver for an advanced Nuclear Power Plant design.

constrained optimization↗
DOE OSTI2021

Communication-Avoiding and Memory-Constrained Sparse Matrix-Matrix Multiplication at Extreme Scale

Sparse matrix-matrix multiplication (SpGEMM) is a widely used kernel in various graph, scientific computing and machine learning algorithms. In this paper, we consider SpGEMMs performed on hundreds of thousands of processors generating trillions of nonzeros in the output matrix. Distributed SpGEMM at this extreme scale faces two key challenges: (1) high communication cost and (2) inadequate memory to generate the output. Furthermore, we address these challenges with an integrated communication-avoiding and memory-constrained SpGEMM algorithm that scales to 262,144 cores (more than 1 million hardware threads) and can multiply sparse matrices of any size as long as inputs and a fraction of output fit in the aggregated memory. As we go from 16,384 cores to 262,144 cores on a Cray XC40 supercomputer, the new SpGEMM algorithm runs 10x faster when multiplying large-scale protein-similarity matrices.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Deep learning enhanced super-resolution x-ray fluorescence microscopy by a dual-branch network

X-ray fluorescence (XRF) microscopy is a powerful technique for quantifying the distribution of elements in complex materials, which makes it a crucial imaging technique across a wide range of disciplines in physical and biological sciences, including chemistry, materials science, microbiology, and geosciences. However, as a scanning microscopy technique, the spatial resolution of XRF imaging is inherently constrained by the x-ray probe profile and scanning step size. Here we propose a dual-branch machine learning (ML) model, which can extract scale-variant features and bypass abundant low-frequency information separately, to enhance the spatial resolution of the XRF images by mitigating the effects of blurring from the probe profile. The model is trained by simulated natural images, and a two-stage training strategy is used to overcome the domain gap between the natural images and experimental data. The tomography reconstruction from enhanced XRF projections shows an improvement in resolution by a scale factor of four and reveals distinct internal features invisible in low-resolution XRF within a battery sample. This study offers a promising approach for obtaining high-resolution XRF imaging from its low-resolution version, paving the way for future investigations in a broader range of disciplines and materials.

36 MATERIALS SCIENCE↗