Search NASA⌕ Search

SEARCH · Search NASA

Results for “functional approximation”

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 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Solving high-dimensional partial integral differential equations: The finite expression method

Partial integro-differential equations (PIDEs) have broad applications in the sciences, from electro-magnetism to options pricing. Here, in this paper, we introduce a new finite expression method (FEX) to solve PIDEs. This approach builds upon the original FEX and its inherent advantages with new advances: 1) A novel method of parameter grouping is proposed to reduce the number of coefficients in high-dimensional function approximation; 2) A Taylor series approximation method is implemented to significantly improve the computational efficiency and accuracy of the evaluation of the integral terms of PIDEs. The new FEX based method, denoted FEX-PG to indicate the addition of the parameter grouping (PG) step to the algorithm, provides both high accuracy and interpretable numerical solutions, with the outcome being an explicit equation that facilitates intuitive understanding of the underlying solution structures. These features are often absent in traditional methods, such as finite element methods (FEM) and finite difference methods, as well as in deep learning-based approaches. To benchmark our method against recent advances, we apply the new FEX-PG to solve benchmark PIDEs in the literature. In high-dimensional settings, FEX-PG exhibits strong and robust performance, achieving relative errors on the order of single precision machine epsilon, significantly outperforming existing approaches based on neural networks.

Combinatorial optimization↗
DOE OSTI2026

Thorium Monosilicide, ThSi: An Experimental and Theoretical Study

The present theoretical and experimental combination study investigates the ThSi molecule in detail. Computationally, we utilized high-level multireference and coupled-cluster levels of theory conjoined with large correlation consistent basis sets to study a series of electronic and spin–orbit states of ThSi. Here, we report potential energy curves (PECs), electron configurations at equilibrium distances, spectroscopic constants, energetics, and spin–orbit coupling effects for 16 electronic states of ThSi. The studied 16 electronic states are arranged tightly within 0.9 eV, highlighting the complexity of the electronic spectrum of ThSi. The ground electronic state of ThSi is a single-reference 1 1 Σ + state that derives from the 1σ 2 2σ 2 1π 4 electronic configuration. The Ω = 0 + spin–orbit ground state of ThSi is composed of 1 1 Σ + (47%) and 13Π (44%) electronic states. Our measured bond energy (D0) of ThSi, obtained using resonant two-photon ionization (R2PI) spectroscopy is 3.146(4) eV, where the assigned error limit is given in parentheses in units of the last quoted digits. The computed D0 of ThSi (Ω = 0 + ) at the CBS-C-CCSD(T)-δT(Q)-δDK-δSO level (3.181 eV) is in good agreement with the experimental value. Our derived enthalpy of formation for ThSi, Δ f H 0K o (ThSi(g)), is 971.8(6.0) kJ/mol. Finally, we have performed density functional theory (DFT) calculations for ThSi(1 1 Σ + ) using 16 exchange correlation functionals that span multiple rungs of “Jacob’s ladder” of density functional approximation (DFA) to assess the DFT errors on D 0 , r e , and ω e of ThSi with respect to experimental and ab initio coupled-cluster values.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

Insights into Nonelectroactive C–C Bond Formation on Cu(100) during Electrochemical CO 2 Reduction from Multiconfigurational Wavefunction Theory

Carbon–carbon (C–C) bond formation is necessary for hydrocarbon (and oxygenate) synthesis beyond methane (and formate/formic acid) during electrochemical CO and CO 2 reduction (ECOR and ECO 2 R). Cu has notable ability to form hydrocarbons compared to other pure metals. In particular, the (100) facet of face-centered cubic Cu forms ethylene competitively with H 2 and methane during both ECOR and ECO 2 R. Past simulations based on density functional theory (DFT) with standard exchange-correlation functional approximations predict fast nonelectroactive C–C bond formation channels involving adsorbed (*) CO together with another *CO, formyl (*CHO), or hydroxymethylidyne (*COH), forming OC*–*CO, OC*–CHO*, and OC*–*COH, respectively. Such simulations support the prevailing hypothesis that emergence of C 2 products is kinetically determined at the early stages of the reduction chemistry. Here we show, via simulations with more accurate many-body, i.e., “correlated”, wavefunction theory (enabled by an embedding scheme), that the coupling of *CO with a *CO or a *COH (previously predicted at the same level of theory to kinetically dominate over *CHO as the one-electron reduction product of *CO) is highly activated (kinetically impeded), with free energy barriers >1 eV, in contradiction to previous DFT-based simulations. Intriguingly, we find that the coupling of two adjacent *COHs incurs only a small barrier (<0.3 eV) and is exoergic (< –1 eV); however, given the predicted low surface mobility of *COH, the emergence of HOC*–*COH is also improbable, at least at low *COH coverages. We therefore conclude that it is highly unlikely for *CO to participate in nonelectroactive C–C bond formation on pristine Cu(100), contrary to conventional wisdom, and that the energetically favorable *COH dimerization may occur only after substantial buildup of *COH on the surface.

adsorption↗
DOE OSTI2025

Parametric matrix models

We present a general class of machine learning algorithms called parametric matrix models. In contrast with most existing machine learning models that imitate the biology of neurons, parametric matrix models use matrix equations that emulate physical systems. Similar to how physics problems are usually solved, parametric matrix models learn the governing equations that lead to the desired outputs. Parametric matrix models can be efficiently trained from empirical data, and the equations may use algebraic, differential, or integral relations. While originally designed for scientific computing, we prove that parametric matrix models are universal function approximators that can be applied to general machine learning problems. After introducing the underlying theory, we apply parametric matrix models to a series of different challenges that show their performance for a wide range of problems. For all the challenges tested here, parametric matrix models produce accurate results within an efficient and interpretable computational framework that allows for input feature extrapolation.

Computational science↗
DOE OSTI2025

Extending quantum-mechanical benchmark accuracy to biological ligand-pocket interactions

Predicting the binding affinity of ligands to protein pockets is key in the drug design pipeline. The flexibility of ligand-pocket motifs arises from a range of attractive and repulsive electronic interactions during binding. Accurately accounting for all interactions requires robust quantum-mechanical (QM) benchmarks, which are scarce for ligand-pocket systems. Additionally, disagreement between “gold standard” Coupled Cluster (CC) and Quantum Monte Carlo (QMC) methods casts doubt on many benchmarks for larger non-covalent systems. We introduce the “QUantum Interacting Dimer” (QUID) benchmark framework containing 170 non-covalent (non-)equilibrium systems modeling chemically and structurally diverse ligand-pocket motifs. Symmetry-adapted perturbation theory shows that QUID broadly covers non-covalent binding motifs and energetic contributions. Robust binding energies are obtained using complementary CC and QMC methods, achieving agreement of 0.5 kcal/mol. The benchmark data analysis reveals that several dispersion-inclusive density functional approximations provide accurate energy predictions, though their atomic van der Waals forces differ in magnitude and orientation. Contrarily, semiempirical methods and empirical force fields require improvements in capturing non-covalent interactions (NCIs) for out-of-equilibrium geometries. The wide span of NCIs, highly accurate interaction energies, and analysis of molecular properties take QUID beyond the “gold standard” for QM benchmarks of ligand-protein systems.

Puleva, Mirela [University of Luxembourg, Luxembou↗
DOE OSTI2026

Nonlinear thermodynamic computing out of equilibrium

We present the design for a thermodynamic computer that can perform arbitrary nonlinear calculations in or out of equilibrium. Simple thermodynamic circuits, fluctuating degrees of freedom in contact with a thermal bath and confined by a quartic potential, display an activity that is a nonlinear function of their input. Such circuits can therefore be regarded as thermodynamic neurons, and can serve as the building blocks of networked structures that act as thermodynamic neural networks, universal function approximators whose operation is powered by thermal fluctuations. We simulate a digital model of a thermodynamic neural network, and show that its parameters can be adjusted by genetic algorithm to perform nonlinear calculations at specified observation times, regardless of whether the system has attained thermal equilibrium. This work expands the field of thermodynamic computing beyond the regime of thermal equilibrium, enabling fully nonlinear computations, analogous to those performed by classical neural networks, at specified observation times.

Whitelam, Stephen [Lawrence Berkeley National Labo↗
DOE OSTI2025

Atomic ionization: sd energy imbalance and Perdew–Zunger self-interaction correction energy penalty in 3d atoms

To accurately describe the energetics of transition metal systems, density functional approximations (DFAs) must provide a balanced description of s- and d- electrons. One measure of this is the sd transfer error, which has previously been defined as E ( 3 d n − 1 4 s 1 ) − E ( 3 d n − 2 4 s 2 ) . Theoretical concerns have been raised about this definition due to its evaluation of excited-state energies using ground-state DFAs. A more serious concern appears to be strong correlation in the 4s 2 configuration. Here, we define a ground-state measure of the sd energy imbalance, based on the errors of s- and d-electron second ionization energies of the 3d atoms, that effectively circumvents the aforementioned problems. We find an improved performance as we move from the local spin density approximation (LSDA) to the Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA) to the regularized and restored Strongly Constrained and Appropriately Normed (r 2 SCAN) meta-GGA for first-row transition metal atoms. However, we find large (∼2 eV) ground-state sd energy imbalances when applying a Perdew–Zunger 1981 self-interaction correction. This is attributed to an “energy penalty” associated with the noded 3d orbitals. A local scaling of the self-interaction correction to LSDA results in a balance of s- and d-errors.

Science & Technology - Other Topics↗
DOE OSTI2026

Ab Initio Bulk Free Energy Surface of Proper Ferroelectrics

We report a systematic and accurate approach for deriving the bulk free energy surface (FES), a function of temperature, polarization, and strain, from the first-principles density functional theory (DFT) of proper ferroelectrics. The core of our approach is the metadynamics algorithm that extracts the polarization dependence of the FES from all-atom molecular dynamics simulations without an a priori ansatz. The rest of the FES is derived from the metadynamics trajectories that span the relevant phase space. We demonstrate our approach in the case of lead titanate. The errors across the phase transition, due to DFT numerics, all-atom molecular dynamics, and free energy evaluation by enhanced sampling, can be systematically controlled and are of the order of 1 meV/atom. The accuracy of the resulting ab initio FES is only limited by the adopted functional approximation of DFT.

Xie, Pinchen [Lawrence Berkeley National Laborator↗
DOE OSTI2025

First principles study of the Fermi surface topology of CeCu 2 ⁢Si 2

Since the discovery of heavy-fermion superconductivity in CeCu 2 ⁢Si 2 , the material has attracted great interest, particularly with regard to the nature of the superconducting pairing and its mechanism. Consequently, it is essential to better understand the electronic Fermi surface topology and its role in strong antiferromagnetic fluctuations. The standard density functional theory method is insufficient to model the interplay of strong on-site Coulomb repulsion in localized 4⁢𝑓 electrons and their hybridization with itinerant ligand-orbital electrons. We have performed electronic ground-state calculations on CeCu 2 ⁢Si 2 using the Gutzwiller wave function approximation. The Gutzwiller approximation captures the quasiparticle band renormalization from the strong on-site Coulomb repulsion. We have performed an analysis of this effect on the electronic structure and the Fermi surface topology by varying the interaction strength and taking into account the crystal-field splitting. Using the de Haas-van Alphen effect, the extremal Fermi surface cross-sectional areas were calculated to quantify the effects of quasiparticle mass renormalization on the Fermi surface. Our results confirm the presence of two Fermi surface sheets corresponding to the heavy (488⁢𝑚 𝑒 ) and light (4.35⁢𝑚 𝑒 ) quasiparticles when the crystal-field splitting is accounted for on equal footing with the electronic correlations. This method gives the best agreement with experimental measurements as well as the renormalized band method.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Model-form Error Correction using Universal Differential Equations for an Agent-Based Model of Infectious Disease

This report demonstrates universal differential equations (UDEs) as an approach to bridge the gap between ordinary differential equations (ODE) models and agent-based models (ABMs). Using UDE models as surrogates for ABMs allows us to preserve the foundational ODE that represents global disease dynamics while coupling it with a neural network model to approximate functions for the local behaviors of the ABM.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Anti-symmetric barron functions and their approximation with sums of determinants

A fundamental problem in quantum physics is to encode functions that are completely anti-symmetric under permutations of identical particles. The architecture of neural network models for the electron wave function typically comprises an equivariant component followed by a summation of determinants. The recently introduced Generic Antisymmetric (GA) block is designed to enhance the expressivity of such neural wave functions, and it was found that the 2-layer GA block achieved more accurate energies than the corresponding single-determinant FermiNet architecure, suggesting its promise as a way to improve the expressivity of neural wave functions. In this paper we show how the function expressed by the 2-layer GA block can be decomposed into a sum of determinants. We formalize this result by defining the antisymmetric Barron space as a generalized version of the 2-layer GA block and providing an appromation theorem for this function class. This result can be viewed as a negative result showing that the 2-layer GA block is not more expressive than using multiple determinants.

Abrahamsen, Nilin↗
DOE OSTI2026

Multiscale Neural Networks for Approximating Green’s Functions

Neural networks (NNs) have been widely used to solve partial differential equations (PDEs) in the applications of physics, biology, and engineering. One effective approach for solving PDEs with a fixed differential operator is learning Green’s functions. However, Green’s functions are notoriously difficult to learn due to their poor regularity, which typically requires larger NNs and longer training times. In this work, we address these challenges by leveraging multiscale NNs to learn Green’s functions. Through theoretical analysis using multiscale Barron space methods and experimental validation, we show that the multiscale approach significantly reduces the necessary NN size and accelerates training.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

A compendium of human gene functions derived from evolutionary modelling

A comprehensive, computable representation of the functional repertoire of all macromolecules encoded within the human genome is a foundational resource for biology and biomedical research. The Gene Ontology Consortium has been working towards this goal by generating a structured body of information about gene functions, which now includes experimental findings reported in more than 175,000 publications for human genes and genes in experimentally tractable model organisms 1,2 . Here, we describe the results of a large, international effort to integrate all of these findings to create a representation of human gene functions that is as complete and accurate as possible. Specifically, we apply an expert-curated, explicit evolutionary modelling approach to all human protein-coding genes. This approach integrates available experimental information across families of related genes into models that reconstruct the gain and loss of functional characteristics over evolutionary time. The models and the resulting set of 68,667 integrated gene functions cover approximately 82% of human protein-coding genes. The functional repertoire reveals a marked preponderance of molecular regulatory functions, and the models provide insights into the evolutionary origins of human gene functions. We show that our set of descriptions of functions can improve the widely used genomic technique of Gene Ontology enrichment analysis. The experimental evidence for each functional characteristic is recorded, thereby enabling the scientific community to help review and improve the resource, which we have made publicly available.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Deep Neural Networks are Adaptive to Function Regularity and Data Distribution in Approximation and Estimation

Deep learning has exhibited remarkable results across diverse areas. To understand its success, substantial research has been directed towards its theoretical foundations. Nev- ertheless, the majority of these studies examine how well deep neural networks can model functions with uniform regularities. In this paper, we explore a different angle: how deep neural networks can adapt to varying degrees of smoothness in functions and nonuni- form data distributions across different locations and scales. More precisely, we focus on a broad class of functions defined by nonlinear tree-based approximation methods. This class encompasses a range of function types, such as functions with uniform regularities and discontinuous functions. We develop nonparametric approximation and estimation theories for this class using deep ReLU networks. Our results show that deep neural networks are adaptive to the nonuniform smoothness of functions and nonuniform data distributions at different locations and scales. We apply our results to several function classes, and derive the corresponding approximation and generalization errors. The validity of our results is demonstrated through numerical experiments.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Chapter 7: Learning Stable Local Volt/Var Controllers in Distribution Grids

This chapter describes a framework to synthesize provably stable local Volt/Var controllers for distributed energy resources (DERs) in power distribution grids (DGs). The goal is to control the reactive power injections of DERs to improve the system performance as quantified by a generic optimal reactive power flow (ORPF) problem. To achieve this, we jointly design for each DER the control function, which prescribes the reactive power update rule, and the equilibrium function, which approximates the ORPF solutions from local measurements of voltages and powers. We provide conditions on the equilibrium functions and the control parameters ensuring the stability of the closed-loop system. In particular, we discuss the trade-offs between each set of conditions accounting for practical considerations, like fully exploiting the DERs' generation capabilities and reducing the optimality gap. These conditions are then translated into learning constraints on the neural networks' parameters that are enforced in the training phase. We validate our framework with numerical simulations on the IEEE 37-bus network and through a comparison with an optimized version of standard piece wise linear control rules.

closed-loop asymptotic stability↗
DOE OSTI2024

Anderson impurity mechanism for a multi-level model in δ-Pu

Abstract Electronic correlations and spin–orbit interactions in plutonium create variations in the bonding behavior of each of its allotropes. In δ -Pu, the 5f electrons lie at the tipping point between itinerant and localized behavior which has made the use of mixed-level models successful in describing its mechanical properties. The mechanism for the emergence of a mixed-level model has not yet been understood. We use a series of density functional theory approximations to understand the interactions that create a mixed-level description of δ -Pu which leads to accurate physical properties. With the intersite interactions present in the hybrid functional, we show that a single 5f electron engages in orbital-selective bonding that can be understood with an Anderson impurity picture. The Anderson model gives us a mechanism to understand how the bonding in δ -Pu evolves as a function of the interactions in the material such that we obtain both the accuracy and physics of the multi-level models from ab initio theory.

38 RADIATION CHEMISTRY, RADIOCHEMISTRY, AND NUCLEA↗