Search NASA⌕ Search

SEARCH · Search NASA

Results for “Training Principles”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Learning dynamical systems from data: An introduction to physics-guided deep learning

Modeling complex physical dynamics is a fundamental task in science and engineering. Traditional physics-based models are first-principled, explainable, and sample-efficient. However, they often rely on strong modeling assumptions and expensive numerical integration, requiring significant computational resources and domain expertise. While deep learning (DL) provides efficient alternatives for modeling complex dynamics, they require a large amount of labeled training data. Furthermore, its predictions may disobey the governing physical laws and are difficult to interpret. Physics-guided DL aims to integrate first-principled physical knowledge into data-driven methods. It has the best of both worlds and is well equipped to better solve scientific problems. Recently, this field has gained great progress and has drawn considerable interest across discipline Here, we introduce the framework of physics-guided DL with a special emphasis on learning dynamical systems. We describe the learning pipeline and categorize state-of-the-art methods under this framework. We also offer our perspectives on the open challenges and emerging opportunities.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

SPIKANs: separable physics-informed Kolmogorov–Arnold networks

Physics-Informed Neural Networks (PINNs) have emerged as a promising method for solving partial differential equations (PDEs) in scientific computing. While PINNs typically use multilayer perceptrons (MLPs) as their underlying architecture, recent advancements have explored alternative neural network structures. One such innovation is the Kolmogorov–Arnold Network (KAN), which has demonstrated benefits over traditional MLPs, including faster neural scaling and better interpretability. The application of KANs to physics-informed learning has led to the development of Physics-Informed KANs (PIKANs), enabling the use of KANs to solve PDEs. However, despite their advantages, KANs often suffer from slower training speeds, particularly in higher-dimensional problems where the number of collocation points grows exponentially with the dimensionality of the system. To address this challenge, we introduce Separable Physics-Informed Kolmogorov–Arnold Networks (SPIKANs). This novel architecture applies the principle of separation of variables to PIKANs, decomposing the problem such that each dimension is handled by an individual KAN. This approach drastically reduces the computational complexity of training without sacrificing accuracy, facilitating their application to higher-dimensional PDEs. Through a series of benchmark problems, we demonstrate the effectiveness of SPIKANs, showcasing their superior scalability and performance compared to PIKANs and highlighting their potential for solving complex, high-dimensional PDEs in scientific computing.

Kolmogorov-Arnold networks↗
DOE OSTI2024

Sparsified time-dependent Fourier neural operators for fusion simulations

This paper presents a sparsified Fourier neural operator for coupled time-dependent partial differential equations (ST-FNO) as an efficient machine learning surrogate for fluid and particle-based fusion codes such as NIMROD (Non-Ideal Magnetohydrodynamics with Rotation - Open Discussion) and GTC (Gyrokinetic Toroidal Code). ST-FNO leverages the structures in the governing equations and utilizes neural operators to represent Green's function-like numerical operators in the corresponding numerical solvers. Once trained, ST-FNO can rapidly and accurately predict dynamics in fusion devices compared with first-principle numerical algorithms. In general, ST-FNO represents an efficient and accurate machine learning surrogate for numerical simulators for multi-variable nonlinear time-dependent partial differential equations, with the proposed architectures and loss functions. The efficacy of ST-FNO has been demonstrated using quiescent H-mode simulation data from NIMROD and kink-mode simulation data from GTC. The ST-FNO H-mode results show orders of magnitude reduction in memory and central processing unit usage in comparison with the numerical solvers in NIMROD when computing fields over a selected poloidal plane. The ST-FNO kink-mode results achieve a factor of 2 reduction in the number of parameters compared to baseline FNO models without accuracy loss.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2025

Unveiling X-ray absorption signatures of boron nitride via first-principles simulation and machine learning

Boron nitride (BN) allotropes hold great promise in many advanced applications ranging from optical and photonic devices to energy storage and battery systems to tribological components. The diverse functionalities of this material stem from BN’s highly tunable structural and electronic properties, which are governed by the versatile boron–nitrogen bonding configurations. Exploring the structural landscape of BN can unveil novel structures possessing unique properties suited for specific applications, therefore accelerating the design of next-generation advanced functional materials. In this work, we leverage boron K-edge X-ray absorption spectroscopy (XAS) as an effective probe for local structural features and chemical environments. A total of 210 BN crystal structures are generated via analogies to the extensive array of carbon allotropes, and XAS is simulated for each unique local motif within the resulting collection of structures. A mapping between structural features and spectral signatures was established by synergizing first-principle simulations with data-driven based post-analysis approaches. Specifically, we developed a neural network model that can satisfactorily predict spectra line shapes from local structural descriptors. Toward automatic spectroscopic interpretation of any new BN structures, supervised machine learning models, trained on this structure–spectrum dataset, can accurately infer local coordination environments from simulated XAS, highlighting the strength of this unique approach of combining high-fidelity first-principles simulation and machine-learning to accelerate target design of novel BN materials via rational understanding of local structure-spectrum correlations.

36 MATERIALS SCIENCE↗
DOE OSTI2026

SimLBR: Learning to Detect Fake Images by Learning to Detect Real Images

The rapid advancement of generative models has made the detection of AI-generated images a critical challenge for both research and society. Recent works have shown that most state-of-the-art fake image detection methods overfit to their training data and catastrophically fail when evaluated on curated hard test sets with strong distribution shifts. In this work, we argue that it is more principled to learn a tight decision boundary around the real image distribution and treat the fake category as a sink class. To this end, we propose SimLBR, a simple and efficient framework for fake image detection with Latent Blending Regularization (LBR). Our method significantly improves cross-generator generalization, achieving up to +24.85% accuracy and +69.62% recall on the challenging Chameleon benchmark. SimLBR is also highly efficient, training orders of magnitude faster than existing approaches. Furthermore, we emphasize the need for reliability-oriented evaluation in fake image detection, introducing risk-adjusted metrics and worst-case estimates to better assess model robustness. All the code and models are availabe at: https://github.com/mvrl/SimLBR

Dhakal, Aayush [Washington University, St. Louis]↗
DOE OSTI2026

Learning thermodynamic master equations for open quantum systems

The characterization of Hamiltonians and other components of open quantum dynamical systems plays a crucial role in quantum computing and other applications. Scientific machine learning techniques have been applied to this problem in a variety of ways, including by modeling with deep neural networks. However, the majority of mathematical models describing open quantum systems are linear, and the natural nonlinearities in learnable models have not been incorporated using physical principles. We present a data-driven model for open quantum systems that includes learnable, thermodynamically consistent terms. The trained model is interpretable, as it directly estimates the system Hamiltonian and linear components of coupling to the environment. We validate the model on synthetic two and three-level data, as well as experimental two-level data collected from a quantum device at Lawrence Livermore National Laboratory.

Mathematics and Computing↗
DOE OSTI2024

A high-throughput framework for lattice dynamics

We develop an automated high-throughput workflow for calculating lattice dynamical properties from first principles including those dictated by anharmonicity. The pipeline automatically computes interatomic force constants (IFCs) up to 4th order from perturbed training supercells, and uses the IFCs to calculate lattice thermal conductivity, coefficient of thermal expansion, and vibrational free energy and entropy. It performs phonon renormalization for dynamically unstable compounds to obtain real effective phonon spectra at finite temperatures and calculates the associated free energy corrections. The methods and parameters are chosen to balance computational efficiency and result accuracy, assessed through convergence testing and comparisons with experimental measurements. Deployment of this workflow at a large scale would facilitate materials discovery efforts toward functionalities including thermoelectrics, contact materials, ferroelectrics, aerospace components, as well as general phase diagram construction.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗
DOE OSTI2026

Towards Generalizable and Efficient Circuit Topology Design: A Graph-Transformer-based Surrogate Model with Curriculum Learning

Unlike circuit parameter and sizing optimizations, the automated design of analog circuit topologies poses significant challenges for learning-based approaches. One challenge arises from the combinatorial growth of the topology space with circuit size, which limits the topology optimization efficiency. Moreover, traditional circuit evaluation methods are time-consuming, while the presence of data discontinuity in the topology space makes the accurate prediction of circuit performance exceptionally difficult for unseen topologies. To tackle these challenges, we design a novel Graph-Transformer-based Network (GTN) as the surrogate model for circuit evaluation, offering a substantial acceleration in the speed of circuit topology optimization without sacrificing performance. Our GTN model architecture is designed to embed voltage changes in circuit loops and current flows in connected devices, enabling accurate performance predictions for circuits with unseen topologies. To address the cold start problem when scaling GTN to large-scale circuits, we further introduce a curriculum learning strategy that progressively trains GTN from small-scale to large-scale circuits. This approach enables the model to first learn fundamental physical principles from simpler topologies and gradually adapt to complex configurations, effectively bridging the circuit complexity gap and improving prediction accuracy. Taking the power converter circuit design as an experimental task, our GTN model significantly outperforms an analytical approach and baseline methods directly utilizing graph neural networks. Furthermore, GTN achieves less than 5% relative error and 196× speed-up compared with high-fidelity simulation. Notably, our GTN surrogate model empowers an automatic circuit design framework to discover circuits of comparable quality to those identified through high-fidelity simulation while reducing the time required by up to 98.2%. With curriculum learning, the enhanced GTN achieves a 51% improvement for performance prediction of large-scale circuits compared to the GTN model without this strategy. These advancements establish GTN as a scalable framework for automated analog circuit design across varying circuit complexity levels.

Lu, Haoshu [New Jersey Institute of Technology (NJ↗
DOE OSTI2025

Cyber-Informed Engineering (CIE) Guide for States

The Cyber-Informed Engineering (CIE) Guide for States provides state energy offices, public utility commissions, and partner organizations with a structured framework for integrating cyber-resilient engineering practices into energy planning, grantmaking, interconnection processes, and workforce development. As grid digitalization and the adoption of distributed energy resources accelerate, states face expanding cyber-physical risks that traditional cybersecurity measures alone cannot fully address. CIE offers a proactive, consequence-focused engineering methodology that emphasizes eliminating or mitigating high-impact failure modes through design, physical controls, and operational safeguards. The guide outlines the 12 core CIE principles, demonstrates their application through state-focused use cases—including grant evaluation rubrics, interconnection reviews, allow-list development, and training programs—and provides practical tools such as scoring frameworks, impact assessment methods, and implementation checklists. It also highlights pathways for state–utility collaboration and opportunities for technical assistance from national laboratories. By adopting CIE, states can enhance grid reliability, reduce lifecycle costs, strengthen supply-chain assurance, and foster a security-aware engineering culture that aligns with broader resilience and modernization goals. November 2025

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
DOE OSTI2025

Cyber-Informed Engineering (CIE) Guide for States

The Cyber-Informed Engineering (CIE) Guide for States provides state energy offices, public utility commissions, and partner organizations with a structured framework for integrating cyber-resilient engineering practices into energy planning, grantmaking, interconnection processes, and workforce development. As grid digitalization and the adoption of distributed energy resources accelerate, states face expanding cyber-physical risks that traditional cybersecurity measures alone cannot fully address. CIE offers a proactive, consequence-focused engineering methodology that emphasizes eliminating or mitigating high-impact failure modes through design, physical controls, and operational safeguards. The guide outlines the 12 core CIE principles, demonstrates their application through state-focused use cases—including grant evaluation rubrics, interconnection reviews, allow-list development, and training programs—and provides practical tools such as scoring frameworks, impact assessment methods, and implementation checklists. It also highlights pathways for state–utility collaboration and opportunities for technical assistance from national laboratories. By adopting CIE, states can enhance grid reliability, reduce lifecycle costs, strengthen supply-chain assurance, and foster a security-aware engineering culture that aligns with broader resilience and modernization goals. November 2025

29 - ENERGY PLANNING, POLICY AND ECONOMY↗
DOE OSTI2025

Generalizable Image Segmentation for Microstructure Characterization Through Integrated SEM and EBSD Analysis

We demonstrate generalizable semantic segmentation using minimal ground truth data. Correlated scanning electron microscopy (SEM) images and electron backscatter diffraction (EBSD) measurements of frictionstir processed 316L stainless steel plates were used to train deep learning models for grain boundary segmentation. Secondary electron (SE) imaging taken at an accelerating voltage of 10 keV correlated to EBSD-derived grain boundaries produced the best performing model. Notably, an ensemble of three models trained on a single SE image produced accurate segmentation over a series of BSE images of samples manufactured under different processing parameters, with a resultant mean absolute error in grain size of 0.34 µm. The striking generalizability of the models likely results from the similar escape depths of the SE training input and the EBSD training output and the reduced probability of dislocation artifacts appearing in the image. This finding highlights the importance of considering the physical principles behind imaging in the development of robust segmentation models for microstructure characterization.

Taufique, Mohammad Fuad Nur↗
DOE OSTI2026

Database and deep-learning scalability of anharmonic phonon properties by automated brute-force first-principles calculations

Understanding the anharmonic phonon properties of crystal compounds—such as phonon lifetimes and thermal conductivities—is essential for investigating and optimizing their thermal transport behaviors. These properties also impact optical, electronic, and magnetic characteristics through interactions between phonons and other quasiparticles and fields. In this study, we develop an automated first-principles workflow to calculate anharmonic phonon properties and build a comprehensive database encompassing more than 6500 inorganic compounds. Utilizing this dataset, we train a graph neural network model to predict thermal conductivity values and spectra from structural parameters, demonstrating a scaling law in which prediction accuracy improves with increasing training data size. High-throughput screening with the model enables the identification of materials exhibiting extreme thermal conductivities—both high and low. The resulting database offers valuable insights into the anharmonic behavior of phonons, thereby accelerating the design and development of advanced functional materials.

Ohnishi, Masato [University of Tokyo (Japan); Inst↗
DOE OSTI2025

Gaussian-process generative model for the QCD equation of state

We develop a generative model for the nuclear matter equation of state at zero net baryon density using the Gaussian process regression method. We impose first-principles theoretical constraints from lattice quantum chromodynamics and hadron resonance gas at high- and low-temperature regions, respectively. By allowing the trained Gaussian process regression model to vary freely near the phase transition region, we generate random smooth crossover equations of state with different speeds of sound that do not rely on specific parametrizations. Here, we explore a collection of experimental observable dependencies on the generated equations of state, which paves the groundwork for future Bayesian inference studies to use experimental measurements from relativistic heavy-ion collisions to constrain the nuclear matter equation of state.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2026

HARMONY: Large-Scale Architecture Search for Efficient Hybrid Language Models

As large language models scale to trillions of parameters, their computational and memory requirements present critical challenges for efficient training and deployment. While Mixture of Experts (MoE) architectures enable efficient scaling through sparse parameter activation, and state-space models like Mamba offer linear-time complexity, principled methods for combining these paradigms remain undeveloped. We introduce HARMONY (Hybrid Architecture Research for Mamba, Optimized with Neural efficiencY), a multi-objective evolutionary neural architecture search framework for discovering efficient hybrid language models that integrate Transformer attention mechanisms, Mixture-of-Experts routing, and Mamba state-space components. Through large-scale distributed search using 16,384 MI250X GPUs on the Frontier supercomputer, HARMONY explores a comprehensive design space encompassing six attention variants (MHA, MQA, GQA, MLA, SWA, and Mamba-2), variable MoE configurations with both routed and shared experts, and extensive Mamba hyperparameters. Our framework discovers heterogeneous architectures that balance training performance with computational efficiency through multi-objective optimization incorporating latency penalties and fitness-based selection. Analysis of discovered architectures reveals that optimal hybrid designs favor heterogeneous component mixing rather than homogeneous patterns, with Mamba-2 and Multi-Head Latent Attention (MLA) emerging as preferred mechanisms. Discovered architectures demonstrate superior training efficiency: our best configuration achieves a final perplexity of 1.0874 with 2.38B parameters while processing 4,320 tokens/second, outperforming significantly larger manually designed models. Full-scale evaluation shows HARMONY's top architectures achieve better loss trajectories than equivalently-sized models using state-of-the-art configurations including Mixtral, Jamba, and Samba. Additionally, we demonstrate 91% weak scaling efficiency when training discovered 36B-parameter models across 1,024 GPUs. HARMONY is released as an open framework with comprehensive tools for building and training hybrid models using expert-data-pipeline parallelism, democratizing access to automated architecture design for next-generation language models.

Herron, Emily [ORNL] (ORCID:0000000273008172)↗
DOE OSTI2025

Advanced Transmission Technologies – GETs and HPCs Session 1: ATT Foundations and Dynamic Line Ratings (DLRs)

The INL TADA GETs Cohort Session 1, held on November 4, 2025, convened experts to address the integration of advanced transmission technologies, including Grid-Enhancing Technologies (GETs) and High Performance Conductors (HPCs), with a focus on digital assurance challenges. The session highlighted the growing importance of cybersecurity, supply chain transparency, reliability, and business risk management in deploying GETs, especially Dynamic Line Ratings (DLRs). Participants examined how expanded attack surfaces, limited vendor pools, and new regulatory requirements—such as FERC Orders 881, 2023, and 1920—are influencing utilities and technology providers. The workshop underscored the need for cyber-informed engineering, secure-by-design principles, and practical risk management strategies, while fostering collaboration and knowledge sharing among industry peers. Technical discussions covered the evolution from static to dynamic line ratings, complexities of cloud-based architectures, and NERC CIP compliance challenges. The session concluded with a collaborative risk exercise and a preview of future workshops on advanced power flow control and transmission topology optimization, reinforcing the cohort’s commitment to advancing digital assurance in the energy sector.

24 - POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2025

OmniXAS: A universal deep-learning framework for materials x-ray absorption spectra

X-ray absorption spectroscopy (XAS) is a powerful characterization technique for probing the local chemical environment of absorbing atoms. However, analyzing XAS data presents significant challenges, often requiring extensive, computationally intensive simulations, as well as significant domain expertise. These limitations hinder the development of fast, robust XAS analysis pipelines that are essential in high-throughput studies and for autonomous experimentation. Here, we address these challenges with OmniXAS, a framework that contains a suite of transfer learning approaches for XAS prediction, each uniquely contributing to improved accuracy and efficiency, as demonstrated on the K-edge spectra database covering eight 3⁢d transition metals (Ti–Cu). The OmniXAS framework is built upon three distinct strategies. First, we use M3GNet [Nat. Comput. Sci. 2, 718 (2022)] to derive latent representations of the local chemical environment of absorption sites as input for XAS prediction, achieving significant improvements over conventional featurization techniques. Second, we employ a hierarchical transfer learning strategy, training a universal multitask model across elements before fine-tuning for element-specific predictions. Models based on this cascaded approach after elementwise fine-tuning outperform element-specific models by up to 69%. Third, we implement cross-fidelity transfer learning, adapting a universal model to predict spectra generated by simulation of a different fidelity with a much higher computational cost. This approach improves prediction accuracy by up to 11% over models trained on the target fidelity alone. Our approach significantly boosts the throughput of XAS modeling by orders of magnitude as compared to first-principles simulations and is extendable to XAS prediction for a broader range of elements. The proposed transfer learning framework is generalizable to enhance deep-learning models that target other properties in materials research.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Artificial-Intelligence Aided Design and Synthesis of Novel Layered 2D Multi-Principal Element Materials for Energy Storage (Final Report)

This DOE-EPSCoR project aimed to predict, synthesize, and characterize novel layered two-dimensional (2D) high-entropy materials (HEMs). These 2D-HEMs, composed of multiple principal elements in nearly equal concentrations, are distinct from traditional 2D materials (typically containing two or three elements) and conventional alloys (dominated by a single primary element with minor secondary additions). Their unique structural and compositional features enable significant lattice strain accommodation, resulting in enhanced electrode performance and potential applications in catalysis, hydrogen storage, sensing, quantum information technologies, and flexible electronics. The research focused on addressing four fundamental questions: (i) What combinations of elements can form stable and synthesizable 2D-HEMs? (ii) What mechanisms drive the stability and synthesizability of crystalline single-phase 2D-HEMs? (iii) How do local chemical disorder and defects influence the macroscopic electronic and mechanical properties? and (iv) What charge storage mechanisms are active in selectively synthesized 2D-HEMs for battery and supercapacitor electrode applications? To achieve these goals, the project employed an integrated theory-experiment approach, incorporating high-throughput first-principles calculations, theoretical modeling, data mining, experimental synthesis, and advanced characterization techniques. The advanced computing resources and state-of-the-art experimental characterization facilities at Oak Ridge National Laboratory (ORNL) were leveraged through collaboration. Beyond scientific advancements, the project contributed to workforce development. Two postdoctoral researchers and three graduate students at the University of Maine were trained through co-advising by ORNL scientists and collaborative interactions, strengthening their expertise in cutting-edge materials science.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Modeling and Optimization of Zeolites for Contaminant Removal from Coal Combustion Impoundment Leachates

Coal Combustion Residual (CCR) landfills in the U.S. can lead to elevated contaminant concentrations in groundwater and runoff, particularly of arsenic, boron, and selenium. Zeolites can be readily synthesized from materials contained within the coal combustion ash impoundments and can be cation-exchanged to enhance adsorption capacity, selectivity, and reactivity. However, the optimization of zeolites becomes a daunting task when accounting for the variety of Si:Al ratios, the species of extra-framework cations present, and the zeolite pore structure. Molecular simulations provide methods to study and guide the design of zeolites for the sorption of contaminants from aqueous solution. In this work, models that predict the sorption of arsenic, boron, and selenium from water using cation-exchanged zeolites. Because of the lack of experimental adsorption data for these species of contaminants, models were parametrized to reproduce the results of first-principles calculations and then used to predict sorption for zeolites for a dataset containing 6000 combinations of zeolites and sorption conditions. Machine learning was used to train a model to predict sorption for materials in this database based on the results of the molecular simulations. Next, a genetic algorithm was used to optimize zeolites for the removal of each contaminant from aqueous solution for individual impoundment sites based on contaminant concentrations reported by the Electric Power Research Institute.

Findley, John↗