SEARCH · Search NASA

Results for “Distributed computing, machine 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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 253 records · Page 14

All sources NASA NTRS DOE OSTI

DOE OSTI2024

Improving North American Wildfire Prediction by Integrating a Machine-Learning Fire Model in a Land Surface Model

Wildfires have shown increasing trends in both frequency and severity across the Contiguous United States (CONUS). However, process-based fire models have difficulties in accurately simulating the burned area over the CONUS due to a simplification of the physical process and cannot capture the interplay among fire, ignition, climate, and human activities. The deficiency of burned area simulation deteriorates the description of fire impact on energy balance, water budget, and carbon fluxes in the Earth System Models (ESMs). Alternatively, machine learning (ML) based fire models, which capture statistical relationships between the burned area and environmental factors, have shown promising burned area predictions and corresponding fire impact simulation. We develop a hybrid framework (ML4Fire-XGB) that integrates a pretrained eXtreme Gradient Boosting (XGBoost) wildfire model with the Energy Exascale Earth System Model (E3SM) land model (ELM) version 2.1. A Fortran-C-Python deep learning bridge is adapted to support online communication between ELM and the ML fire model. Specifically, the burned area predicted by the ML-based wildfire model is directly passed to ELM to adjust the carbon pool and vegetation dynamics after disturbance, which are then used as predictors in the ML-based fire model in the next time step. Evaluated against the historical burned area from Global Fire Emissions Database 5 from 2001-2020, the ML4Fire-XGB model outperforms process-based fire models in terms of spatial distribution and seasonal variations. Sensitivity analysis confirms that the ML4Fire-XGB well captures the responses of the burned area to rising temperatures. The ML4Fire-XGB model has proved to be a new tool for studying vegetation-fire interactions, and more importantly, enables seamless exploration of climate-fire feedback, working as an active component in E3SM.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2011

Collaborative Clustering for Sensor Networks

Traditionally, nodes in a sensor network simply collect data and then pass it on to a centralized node that archives, distributes, and possibly analyzes the data. However, analysis at the individual nodes could enable faster detection of anomalies or other interesting events, as well as faster responses such as sending out alerts or increasing the data collection rate. There is an additional opportunity for increased performance if individual nodes can communicate directly with their neighbors. Previously, a method was developed by which machine learning classification algorithms could collaborate to achieve high performance autonomously (without requiring human intervention). This method worked for supervised learning algorithms, in which labeled data is used to train models. The learners collaborated by exchanging labels describing the data. The new advance enables clustering algorithms, which do not use labeled data, to also collaborate. This is achieved by defining a new language for collaboration that uses pair-wise constraints to encode useful information for other learners. These constraints specify that two items must, or cannot, be placed into the same cluster. Previous work has shown that clustering with these constraints (in isolation) already improves performance. In the problem formulation, each learner resides at a different node in the sensor network and makes observations (collects data) independently of the other learners. Each learner clusters its data and then selects a pair of items about which it is uncertain and uses them to query its neighbors. The resulting feedback (a must and cannot constraint from each neighbor) is combined by the learner into a consensus constraint, and it then reclusters its data while incorporating the new constraint. A strategy was also proposed for cleaning the resulting constraint sets, which may contain conflicting constraints; this improves performance significantly. This approach has been applied to collaborative clustering of seismic and infrasonic data collected by the Mount Erebus Volcano Observatory in Antarctica. Previous approaches to distributed clustering cannot readily be applied in a sensor network setting, because they assume that each node has the same view of the data set. A view is the set of features used to represent each object. When a single data set is partitioned across several computational nodes, distributed clustering works; all objects have the same view. But when the data is collected from different locations, using different sensors, a more flexible approach is needed. This approach instead operates in situations where the data collected at each node has a different view (e.g., seismic vs. infrasonic sensors), but they observe the same events. This enables them to exchange information about the likely cluster membership relations between objects, even if they do not use the same features to represent the objects.

Wagstaff. Loro :/↗

DOE OSTI2024

A Computational Review of Privacy-Preserving Mechanisms for the Smart Grid

Smart grid technologies have rapidly become one of the largest and most comprehensive sources of data for the modern utility. For the most part, data streams are seen as an essential tool that enable utilities to carry their day-to-day business operations, but they also create the need for efficient and secure data management strategies. In the context of the smart grid, ensuring data privacy is becoming an increasing concern due to a combination of factors that range from shifts in operational paradigms and rapid technology evolution to changes in legislation. Furthermore, researchers have highlighted the risks associated with improperly protected energy records. For example, energy consumption data from homes could be used to infer the behaviors and habits of home occupants through activity recognition or user profiling (Fan, 2017), which may lead to unfair service pricing, targeted advertising, or other personal security violations. Similarly, Electric Vehicles’ (EVs) charging metadata could be used to reveal private information about the owner such as their payment methods, preferred charging stations, and other locational and timing information that could be used to reconstruct the vehicle owner’s behaviors. The privacy of user data, even when used for statistical analysis or machine learning training processes, also needs to be carefully considered, as an individual’s private traits may still be vulnerable if their inclusion/exclusion greatly impacts the result or could be linked to a public dataset through cross-reference. The breach of user privacy also has severe impacts for organizations that store, transmit, or work on the data in the form of diminishing the public’s trust in them while potentially incurring legal consequences (e.g., fines and suspensions under the European Union General Data Protection Regulation, Health Insurance Portability and Accountability Act, etc.). Because of these risks, several privacy-preserving mechanisms are available to help organizations comply with privacy legislations and prevent the unauthorized and malicious use of user data. In light of these concerns, this report focuses on performing a computational review of privacy-preserving mechanisms that have received a significant amount of interest in literature. It specifically focuses on 1) homomorphic encryption, 2) zero-knowledge proofs, 3) differential privacy, and 4) federated learning. It is worth noting that although many of the methods presented in this document rely on cryptographic primitives, their intent is not to provide perfect secrecy, but rather to enable users to maintain privacy, and thus they shall not be compared or equated to other constructs that are aimed to address cybersecurity constructs.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2024

Demystifying Cyberattacks: Potential for Securing Energy Systems With Explainable AI : Preprint

Modernization of energy systems has led to in- creased interactions among multiple critical infrastructures and diverse stakeholders making the challenge of operational decision making more complex and at times beyond cognitive capabilities of human operators. The state-of-the-art machine learning and deep learning approaches show promise of supporting users with complex decision-making challenges, such as those occurring in our rapidly transforming cyber-physical energy systems. However, successful adoption of data-driven decision support technology for critical infrastructure will be dependent on the ability of these technologies to be trustworthy and contextually interpretable. In this paper, we investigate the feasibility of implementing XAI for interpretable detection of cyberattacks in the energy system. Leveraging a proof-of-concept simulation use case of detection of a data falsification attack on a photovoltaic system using XGBoost algorithm, we demonstrate how Local Interpretable Model-Agnostic Explanations (LIME), a flavor XAI approach, can help provide contextual and actionable interpretation of cyberattack detection.

artificial intelligence↗

DOE OSTI2024

Side-by-Side Comparison of Subhourly Clipping Models

Over the past several years there have been numerous attempts at quantifying the inherent power clipping of inverters due to sub-hourly irradiance variability that is not captured in hourly PV performance models. Different models have been proposed to correct for these clipping losses in PV performance estimates, including matrix lookup models, distribution modeling of the PV power performance within a given hour, and machine learning methods. To date, there have been few comprehensive quantitative comparisons of these inverter clipping correction modeling approaches to evaluate the effectiveness of these approaches in predicting the actual behavior of PV system inverter clipping. In this study, we perform such a comparison, evaluating the Allen and Walker correction loss modeling approaches recently implemented in the System Advisor Model (SAM) against clipping losses modeled with 1-minute climate data. These comparisons were performed across a variety of climate locations and inverter loading ratios to thoroughly analyze the effectiveness of these modeling approaches relative to each other. Results from this analysis reveal that both clipping correction approaches improve annual energy accuracy to within 2% of 1-minute modeled energy yield. The two models predict annual clipping loss more accurately than simple hourly power limit clipping, with the Allen method typically being slightly more accurate at typical ILR values and the Walker method often being slightly more accurate at high ILR values The models can improve accuracy over the status quo clipping approach up to 3 percentage points in systems with ILR of 2.0, showing the importance of this modeling factor in energy yield estimates.

ENERGY PLANNING, POLICY, AND ECONOMY,MATHEMATICS A↗

DOE OSTI2026

Revealing the Hidden Third Dimension of Point Defects in Two-Dimensional MXenes

Point defects govern many important functional properties of two-dimensional (2D) materials. However, resolving the three-dimensional (3D) arrangement of these defects in multi-layer 2D materials remains a fundamental challenge, hindering rational defect engineering. Here, we overcome this limitation using an artificial intelligence-guided electron microscopy workflow to map the 3D topology and clustering of atomic vacancies in Ti3C2TX MXene. Our approach reconstructs the 3D coordinates of vacancies across hundreds of thousands of lattice sites, generating robust statistical insight into their distribution that can be correlated with specific synthesis pathways. This large-scale data enables us to classify a hierarchy of defect structures-from isolated vacancies to nanopores-revealing their preferred formation and interaction mechanisms, as corroborated by molecular dynamics simulations. This work provides a generalizable framework for understanding and ultimately controlling point defects across large volumes, paving the way for the rational design of defect-engineered functional 2D materials.

2D materials↗

DOE OSTI2025

Equation-Free Coarse Control of Distributed Parameter Systems via Local Neural Operators

The control of high-dimensional distributed parameter systems (DPS) remains a challenge when explicit coarse-grained equations are unavailable. Classical equation-free (EF) approaches rely on fine-scale simulators treated as black-box timesteppers. However, repeated simulations for steady-state computation, linearization, and control design are often computationally prohibitive, or the microscopic timestepper may not even be available, leaving us with data as the only resource. We propose a data-driven alternative that uses local neural operators, trained on spatiotemporal microscopic/mesoscopic data, to obtain efficient short-time solution operators. These surrogates are employed within Krylov subspace methods to compute coarse steady and unsteady-states, while also providing Jacobian information in a matrix-free manner. Krylov-Arnoldi iterations then approximate the dominant eigenspectrum, yielding reduced models that capture the open-loop slow dynamics without explicit Jacobian assembly. Both discrete-time Linear Quadratic Regulator (dLQR) and pole-placement (PP) controllers are based on this reduced system and lifted back to the full nonlinear dynamics, thereby closing the feedback loop.

93B52, 93C20, 47N70, 65J15, 65M32, 68T07, 68T20, 6↗

DOE OSTI2025

Conditioned quantum-assisted deep generative surrogate for particle-calorimeter interactions

Particle collisions at accelerators like the Large Hadron Collider (LHC), recorded by experiments such as ATLAS and CMS, enable precise standard model measurements and searches for new phenomena. Simulating these collisions significantly influences experiment design and analysis but incurs immense computational costs, projected at millions of CPU-years annually during the high luminosity LHC (HL-LHC) phase. Currently, simulating a single event with Geant4 consumes around 1000 CPU seconds, with calorimeter simulations especially demanding. To address this, we propose a conditioned quantum-assisted generative model, integrating a conditioned variational autoencoder (VAE) and a conditioned restricted Boltzmann machine (RBM). Our RBM architecture is tailored for D-Wave’s Pegasus-structured advantage quantum annealer for sampling, leveraging the flux bias for conditioning. This approach combines classical RBMs as universal approximators for discrete distributions with quantum annealing’s speed and scalability. We also introduce an adaptive method for efficiently estimating effective inverse temperature, and validate our framework on Dataset 2 of CaloChallenge.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2026

Ensemble Federated Machine Learning‐Based Cybersecurity Situational Awareness in Microgrid Network

Cyber-physical microgrids are vulnerable to stealthy cybersecurity threats that disguise their actions through the exploitation of system knowledge. Such actions can severely impacts microgrids deployed in defense bases, slowing the response time of military forces during national emergencies. Several machine-learning algorithms have been proposed to detect intrusions in the grid networks; however, these traditional machine-learning algorithms lack data privacy and are subject to several adversarial machine-learning threats. This paper proposes a novel federated machine learning (FML)-based three-model framework to detect and identify stealthy data-integrity attacks while ensuring data privacy in microgrid networks. The proposed architecture uses a variational mode decomposition technique to extract derived features from incoming measurement and control datasets. The extraction of these derived features allows FML models to learn minute variations in data patterns that allow them to perform significantly better than the models trained with generic datasets consisting of raw features. Our experimental results show the efficient performance of the proposed methodology against different types of data integrity attacks while considering primary and secondary controllers in microgrids. Further, the applied FML-integrated random forest ensemble algorithm outperforms the existing generic FML algorithms during noisy and noise-free datasets with prediction latencies of only 91–134 µs per sample within the 0.1 s sampling interval and requires communication bandwidth of around ∼8.25 KB/s at the control center and ∼2.7 KB/s per edge client for communication.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NASA NTRS2022

Optimal Estimation Framework for Ocean Color Atmospheric Correction and Pixel-level Uncertainty Quantification

Ocean color remote sensing requires compensation for atmospheric scattering and absorption (aerosol, Rayleigh, and trace gases), referred to as atmospheric correction (AC). AC allows inference of parameters such as spectrally resolved remote sensing reflectance ( R rs )(λ) ; sr 1 ) at the ocean surface from the top-of-atmosphere reflectance. Often, the uncertainty of this process is not fully explored. Bayesian inference techniques provide a simultaneous AC and uncertainty assessment via a full posterior distribution of the relevant variables, given the prior distribution of those variables and the radiative transfer (RT) likelihood function. Given uncertainties in the algorithm inputs, the Bayesian framework enables better constraints on the AC process by using the complete spectral information compared to traditional approaches that use only a subset of bands for AC. This paper investigates a Bayesian inference research method (Optimal Estimation, OE) for ocean color AC by simultaneously retrieving atmospheric and ocean properties using all visible and near-infrared spectral bands. The OE algorithm analytically approximates the posterior distribution of parameters based on normality assumptions and provides a potentially viable operational algorithm with a reduced computational expense. We developed a Neural Network (NN) RT forward model look-up-table-based emulator to increase algorithm efficiency further and thus speed up the likelihood computations. We then applied the OE algorithm to synthetic data and observations from the MODerate resolution Imaging Spectroradiometer (MODIS) on NASA’s Aqua spacecraft. We compared the R rs )(λ) retrieval and its uncertainty estimates from the OE method with in-situ validation data from the SeaWiFS Bio-optical Archive and Storage System (SeaBASS) and Aerosol Robotic Network Ocean Color (AERONET-OC) datasets. The OE algorithm improved R rs )(λ) estimates relative to the NASA standard operational algorithm by improving all statistical metrics at 443, 555, and 667 nm. Unphysical negative R rs )(λ) , which often appear in complex water conditions, was reduced by a factor of 3. The OE-derived pixel-level R rs )(λ) uncertainty estimates were also assessed relative to in-situ data and were shown to have skill.

Atmospheric correction↗

NASA NTRS1993

A pattern recognition system for locating small volvanoes in Magellan SAR images of Venus

The Magellan data set constitutes an example of the large volumes of data that today's instruments can collect, providing more detail of Venus than was previously available from Pioneer Venus, Venera 15/16, or ground-based radar observations put together. However, data analysis technology has not kept pace with data collection and storage technology. Due to the sheer size of the data, complete and comprehensive scientific analysis of such large volumes of image data is no longer feasible without the use of computational aids. Our progress towards developing a pattern recognition system for aiding in the detection and cataloging of small-scale natural features in large collections of images is reported. Combining classical image processing, machine learning, and a graphical user interface, the detection of the 'small-shield' volcanoes (less than 15km in diameter) that constitute the most abundant visible geologic feature in the more that 30,000 synthetic aperture radar (SAR) images of the surface of Venus are initially targeted. Our eventual goal is to provide a general, trainable tool for locating small-scale features where scientists specify what to look for simply by providing examples and attributes of interest to measure. This contrasts with the traditional approach of developing problem specific programs for detecting Specific patterns. The approach and initial results in the specific context of locating small volcanoes is reported. It is estimated, based on extrapolating from previous studies and knowledge of the underlying geologic processes, that there should be on the order of 10(exp 5) to 10(exp 6) of these volcanoes visible in the Magellan data. Identifying and studying these volcanoes is fundamental to a proper understanding of the geologic evolution of Venus. However, locating and parameterizing them in a manual manner is forbiddingly time-consuming. Hence, the development of techniques to partially automate this task were undertaken. The primary constraints for this particular problem are that the method must be reasonably robust and fast. Unlike most geological features, the small volcanoes of Venus can be ascribed to a basic process that produces features with a short list of readily defined characteristics differing significantly from other surface features on Venus. For pattern recognition purposes the relevant criteria include (1) a circular planimetric outline, (2) known diameter frequency distribution from preliminary studies, (3) a limited number of basic morphological shapes, and (4) the common occurrence of a single, circular summit pit at the center of the edifice.

Burl, M. C.↗

DOE OSTI2024

DiffLense: a conditional diffusion model for super-resolution of gravitational lensing data

Abstract Gravitational lensing data is frequently collected at low resolution due to instrumental limitations and observing conditions. Machine learning-based super-resolution techniques offer a method to enhance the resolution of these images, enabling more precise measurements of lensing effects and a better understanding of the matter distribution in the lensing system. This enhancement can significantly improve our knowledge of the distribution of mass within the lensing galaxy and its environment, as well as the properties of the background source being lensed. Traditional super-resolution techniques typically learn a mapping function from lower-resolution to higher-resolution samples. However, these methods are often constrained by their dependence on optimizing a fixed distance function, which can result in the loss of intricate details crucial for astrophysical analysis. In this work, we introduce DiffLense , a novel super-resolution pipeline based on a conditional diffusion model specifically designed to enhance the resolution of gravitational lensing images obtained from the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP). Our approach adopts a generative model, leveraging the detailed structural information present in Hubble space telescope (HST) counterparts. The diffusion model, trained to generate HST data, is conditioned on HSC data pre-processed with denoising techniques and thresholding to significantly reduce noise and background interference. This process leads to a more distinct and less overlapping conditional distribution during the model’s training phase. We demonstrate that DiffLense outperforms existing state-of-the-art single-image super-resolution techniques, particularly in retaining the fine details necessary for astrophysical analyses.

Computer Science↗

NASA NTRS2009

Understanding the Scalability of Bayesian Network Inference using Clique Tree Growth Curves

Bayesian networks (BNs) are used to represent and efficiently compute with multi-variate probability distributions in a wide range of disciplines. One of the main approaches to perform computation in BNs is clique tree clustering and propagation. In this approach, BN computation consists of propagation in a clique tree compiled from a Bayesian network. There is a lack of understanding of how clique tree computation time, and BN computation time in more general, depends on variations in BN size and structure. On the one hand, complexity results tell us that many interesting BN queries are NP-hard or worse to answer, and it is not hard to find application BNs where the clique tree approach in practice cannot be used. On the other hand, it is well-known that tree-structured BNs can be used to answer probabilistic queries in polynomial time. In this article, we develop an approach to characterizing clique tree growth as a function of parameters that can be computed in polynomial time from BNs, specifically: (i) the ratio of the number of a BN's non-root nodes to the number of root nodes, or (ii) the expected number of moral edges in their moral graphs. Our approach is based on combining analytical and experimental results. Analytically, we partition the set of cliques in a clique tree into different sets, and introduce a growth curve for each set. For the special case of bipartite BNs, we consequently have two growth curves, a mixed clique growth curve and a root clique growth curve. In experiments, we systematically increase the degree of the root nodes in bipartite Bayesian networks, and find that root clique growth is well-approximated by Gompertz growth curves. It is believed that this research improves the understanding of the scaling behavior of clique tree clustering, provides a foundation for benchmarking and developing improved BN inference and machine learning algorithms, and presents an aid for analytical trade-off studies of clique tree clustering using growth curves.

Mengshoel, Ole Jakob↗

DOE OSTI2024

Physics-informed heterogeneous graph neural networks for DC blocker placement

The threat of geomagnetic disturbances (GMDs) to the reliable operation of the bulk energy system has spurred the development of effective strategies for mitigating their impacts. One such approach involves placing transformer neutral blocking devices, which interrupt the path of geomagnetically induced currents (GICs) to limit their impact. The high cost of these devices and the sparsity of transformers that experience high GICs during GMD events, however, calls for a sparse placement strategy that involves high computational cost. To address this challenge, we developed a physics-informed heterogeneous graph neural network (PIHGNN) for solving the graph-based dc-blocker placement problem. Our approach combines a heterogeneous graph neural network (HGNN) with a physics-informed neural network (PINN) to capture the diverse types of nodes and edges in ac/dc networks and incorporates the physical laws of the power grid. We train the PIHGNN model using a surrogate power flow model and validate it using case studies. Results demonstrate that PIHGNN can effectively and efficiently support the deployment of GIC dc-current blockers, ensuring the continued supply of electricity to meet societal demands. Furthermore, our approach has the potential to contribute to the development of more reliable and resilient power grids capable of withstanding the growing threat that GMDs pose.

24 POWER TRANSMISSION AND DISTRIBUTION↗

NASA NTRS2009

Hybrid NN/SVM Computational System for Optimizing Designs

A computational method and system based on a hybrid of an artificial neural network (NN) and a support vector machine (SVM) (see figure) has been conceived as a means of maximizing or minimizing an objective function, optionally subject to one or more constraints. Such maximization or minimization could be performed, for example, to optimize solve a data-regression or data-classification problem or to optimize a design associated with a response function. A response function can be considered as a subset of a response surface, which is a surface in a vector space of design and performance parameters. A typical example of a design problem that the method and system can be used to solve is that of an airfoil, for which a response function could be the spatial distribution of pressure over the airfoil. In this example, the response surface would describe the pressure distribution as a function of the operating conditions and the geometric parameters of the airfoil. The use of NNs to analyze physical objects in order to optimize their responses under specified physical conditions is well known. NN analysis is suitable for multidimensional interpolation of data that lack structure and enables the representation and optimization of a succession of numerical solutions of increasing complexity or increasing fidelity to the real world. NN analysis is especially useful in helping to satisfy multiple design objectives. Feedforward NNs can be used to make estimates based on nonlinear mathematical models. One difficulty associated with use of a feedforward NN arises from the need for nonlinear optimization to determine connection weights among input, intermediate, and output variables. It can be very expensive to train an NN in cases in which it is necessary to model large amounts of information. Less widely known (in comparison with NNs) are support vector machines (SVMs), which were originally applied in statistical learning theory. In terms that are necessarily oversimplified to fit the scope of this article, an SVM can be characterized as an algorithm that (1) effects a nonlinear mapping of input vectors into a higher-dimensional feature space and (2) involves a dual formulation of governing equations and constraints. One advantageous feature of the SVM approach is that an objective function (which one seeks to minimize to obtain coefficients that define an SVM mathematical model) is convex, so that unlike in the cases of many NN models, any local minimum of an SVM model is also a global minimum.

Rai, Man Mohan↗

DOE OSTI2026

Multiscale modeling of packed-bed microwave reactors and estimation of intrinsic materials' permittivity

Modeling of packed-bed microwave reactors relies on an accurate representation of particle size, shape, and distribution within the bed, as well as the particles' dielectric properties. The measured permittivity of microwave susceptors (powders or structured materials) depends on the geometric features of the particles and the porosity of the bed, as well as the specific form factor of a structured material. These are effective properties and cannot be used to analyze other reactor configurations unless the geometric effects are removed. Therefore, we introduce a methodology for extracting the intrinsic particle permittivity from experimentally measured effective permittivity by combining cavity-based measurements with multiscale simulations and machine learning. Further, we develop the first multiscale model of packed-bed microwave reactors that incorporate particle effects (geometric features, random packing, and particle contact). This approach bridges macroscopic observables with mesoscopic physics, enabling analysis of local hotspots, arcing, and contact effects that control reactor performance. Using polymer-based spherical activated carbon (PBSAC) and silicon carbide (SiC) as examples, we demonstrate that the inferred particle permittivity is consistent with independent experimental heating profiles we collect from microwave reactors without adjustable parameters. Finally, this methodology establishes a foundation for predictive, multiscale design of microwave packed-bed reactors that explicitly accounts for particle-scale effects, enabling the estimation of intrinsic permittivity for the first time.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

Resolving turbulent magnetohydrodynamics: a hybrid operator-diffusion framework

We present a hybrid machine learning framework that combines physics-informed neural operators (PINOs) with score-based generative diffusion models to simulate the full spatio-temporal evolution of two-dimensional, incompressible, resistive magnetohydrodynamic turbulence across a broad range of Reynolds numbers (Re). The framework leverages the equation-constrained generalization capabilities of PINOs to predict coherent, low-frequency dynamics, while a conditional diffusion model stochastically corrects high-frequency residuals, enabling accurate modeling of fully developed turbulence. Trained on a comprehensive ensemble of high-fidelity simulations with Re ϵ {100, 250, 500, 750, 1000, 3000, 10000}, the approach achieves state-of-the-art accuracy in regimes previously inaccessible to deterministic surrogates. At Re = 1000 and 3000, the model faithfully reconstructs the full spectral energy distributions of both velocity and magnetic fields late into the simulation, capturing non-Gaussian statistics, intermittent structures, and cross-field correlations with high fidelity. At extreme turbulence levels (Re = 10 000), it remains the first surrogate capable of recovering the high-wavenumber evolution of the magnetic field, preserving large-scale morphology and enabling statistically meaningful predictions.

Diffusion-Integrated Neural Operators↗

DOE OSTI2025

FM4NPP: A Scaling Foundation Model for Nuclear and Particle Physics

Large language models have revolutionized artificial intelligence by enabling large, generalizable models trained through self-supervision. This paradigm has inspired the development of scientific foundation models (FMs). However, applying this capability to experimental particle physics is challenging due to the sparse, spatially distributed nature of detector data, which differs dramatically from natural language. This work addresses if an FM for particle physics can scale and generalize across diverse tasks. We introduce a new dataset with more than 11 million particle collision events and a suite of downstream tasks and labeled data for evaluation. We propose a novel self-supervised training method for detector data and demonstrate its neural scalability with models that feature up to 188 million parameters. With frozen weights and task-specific adapters, this FM consistently outperforms baseline models across all downstream tasks. The performance also exhibits robust data-efficient adaptation. Further analysis reveals that the representations extracted by the FM are task-agnostic but can be specialized via a single linear mapping for different downstream tasks.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗