Search NASA⌕ Search

SEARCH · Search NASA

Results for “gradient methods”

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 253 records · Page 14

Parameter Estimation and Uncertainty Quantification for Systems Biology Models

Mathematical models can provide quantitative insights into immunoreceptor signaling, and other biological processes, but require parameterization and uncertainty quantification before reliable predictions become possible. We review currently available methods and software tools to address these problems. We consider gradient-based and gradient-free methods for point estimation of parameter values, and methods of profile likelihood, bootstrapping, and Bayesian inference for uncertainty quantification. We consider recent and potential future applications of these methods to systems-level modeling of immune-related phenomena.

59 BASIC BIOLOGICAL SCIENCES↗

Natural evolutionary strategies for variational quantum computation

Abstract Natural evolutionary strategies (NES) are a family of gradient-free black-box optimization algorithms. This study illustrates their use for the optimization of randomly initialized parameterized quantum circuits (PQCs) in the region of vanishing gradients. We show that using the NES gradient estimator the exponential decrease in variance can be alleviated. We implement two specific approaches, the exponential and separable NES, for parameter optimization of PQCs and compare them against standard gradient descent. We apply them to two different problems of ground state energy estimation using variational quantum eigensolver and state preparation with circuits of varying depth and length. We also introduce batch optimization for circuits with larger depth to extend the use of ES to a larger number of parameters. We achieve accuracy comparable to state-of-the-art optimization techniques in all the above cases with a lower number of circuit evaluations. Our empirical results indicate that one can use NES as a hybrid tool in tandem with other gradient-based methods for optimization of deep quantum circuits in regions with vanishing gradients.

Anand, Abhinav (ORCID:0000000280812310)↗

Explaining machine-learning models for gamma-ray detection and identification

As more complex predictive models are used for gamma-ray spectral analysis, methods are needed to probe and understand their predictions and behavior. Recent work has begun to bring the latest techniques from the field of Explainable Artificial Intelligence (XAI) into the applications of gamma-ray spectroscopy, including the introduction of gradient-based methods like saliency mapping and Gradient-weighted Class Activation Mapping (Grad-CAM), and black box methods like Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP). In addition, new sources of synthetic radiological data are becoming available, and these new data sets present opportunities to train models using more data than ever before. In this work, we use a neural network model trained on synthetic NaI(Tl) urban search data to compare some of these explanation methods and identify modifications that need to be applied to adapt the methods to gamma-ray spectral data. We find that the black box methods LIME and SHAP are especially accurate in their results, and recommend SHAP since it requires little hyperparameter tuning. We also propose and demonstrate a technique for generating counterfactual explanations using orthogonal projections of LIME and SHAP explanations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

The Importance of Three-Body Interactions in Molecular Dynamics Simulations of Water with the Fragment Molecular Orbital Method

The analytic first derivative with respect to nuclear coordinates is formulated and implemented in the framework of the three-body fragment molecular orbital (FMO) method. The gradient has been derived and implemented for restricted Hartree-Fock, second-order Møller-Plesset perturbation, and density functional theories. The importance of the three-body fully analytic gradient is illustrated through the failure of the two-body FMO method during molecular dynamics simulations of a small water cluster. The parallel implementation of the fragment molecular orbital method, its parallel efficiency, and its scalability on the Blue Gene/Q architecture up to 262,144 CPU cores, are also discussed.

Pruitt, Spencer R.↗

Reconstructing small-scale lenses from the cosmic microwave background temperature fluctuations

Cosmic microwave background (CMB) lensing is a powerful probe of the matter distribution in the Universe. The standard quadratic estimator, which is typically used to measure the lensing signal, is known to be suboptimal for low-noise polarization data from next-generation experiments. In this paper, we explain why the quadratic estimator will also be suboptimal for measuring lensing on very small scales, even for measurements in temperature where this estimator typically performs well. Though maximum likelihood methods could be implemented to improve performance, we explore a much simpler solution, revisiting a previously proposed method to measure lensing that involves a direct inversion of the background gradient. An important application of this simple formalism is the measurement of cluster masses with CMB lensing. We find that directly applying a gradient inversion matched filter to simulated lensed images of the CMB can tighten constraints on cluster masses compared to the quadratic estimator. While the difference is not relevant for existing surveys, for future surveys it can translate to significant improvements in mass calibration for distant clusters, where galaxy lensing calibration is ineffective due to the lack of enough resolved background galaxies. Improvements can be as large as ~50 per cent for a cluster at z = 2 and a next-generation CMB experiment with 1 μK arcmin noise, and over an order of magnitude for lower noise levels. For future surveys, this simple matched filter or gradient inversion method approaches the performance of maximum likelihood methods, at a fraction of the computational cost.

79 ASTRONOMY AND ASTROPHYSICS↗

Filter diagonalization of shell-model calculations

We present a method of filter diagonalization for shell-model calculations. This method is based on the Sakurai and Sugiura (SS) method, but extended with the help of the shifted complex orthogonal conjugate gradient (COCG) method. A salient feature of this method is that it can calculate eigenvalues and eigenstates in a given energy interval. We show that this method can be an alternative to the Lanczos method for calculating ground and excited states, as well as spectral strength functions. With an application to the M-scheme shell-model calculations we demonstrate that several inherent problems in the widely used Lanczos method can be removed or reduced.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Application of the INSTANT-HPS PN Transport Code to the C5G7 Benchmark Problem

INSTANT is the INL's next generation neutron transport solver to support high-fidelity multi-physics reactor simulation INSTANT is in continuous development to extend its capability Code is designed to take full advantage of middle to large cluster (10-1000 processors) Code is designed to focus on method adaptation while also mesh adaptation will be possible. It utilizes the most modern computing techniques to generate a neutronics tool of full-core transport calculations for reactor analysis and design. It can perform calculations on unstructured 2D/3D triangular, hexagonal and Cartesian geometries. Calculations can be easily extended to more geometries because of the independent mesh framework coded with the model Fortran. This code has a multigroup solver with thermal rebalance and Chebyshev acceleration. It employs second-order PN and Hybrid Finite Element method (PNHFEM) discretization scheme. Three different in-group solvers - preconditioned Conjugate Gradient (CG) method, preconditioned Generalized Minimal Residual Method (GMRES) and Red-Black iteration - have been implemented and parallelized with the spatial domain decomposition in the code. The input is managed with extensible markup language (XML) format. 3D variables including the flux distributions are outputted into VTK files, which can be visualized by tools such as VisIt and ParaView. An extension of the code named INSTANTHPS provides the capability to perform 3D heterogeneous transport calculations within fuel pins. C5G7 is an OECD/NEA benchmark problem created to test the ability of modern deterministic transport methods and codes to treat reactor core problems without spatial homogenization. This benchmark problem had been widely analyzed with various code packages. In this transaction, results of the applying the INSTANT-HPS code to the C5G7 problem are summarized.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

The Effects of Compounded Model Size Reductions on Adversarial Robustness

Recent advances in Edge AI and Tiny Machine Learning (TinyML) have enabled the deployment of machine learning models on resource-constrained environments. However, deploying these models on edge devices, such as micro-controllers, requires significant model footprint reduction through a variety of techniques such as quantization, pruning, and clustering. While these optimization methods offer considerable advantages, they potentially introduce AI-related security vulnerabilities, particularly concerning model robustness with respect to adversarial AI attacks. Prior research has extensively examined the impact of quantization on adversarial robustness; however, the effects of alternative reduction techniques and their combinations remain understudied. This paper investigates the impact of model size reduction techniques on adversarial robustness, when applied individually and combined. We utilized Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) attacks to generate adversarial perturbations for both training and testing data, and then evaluated the models' accuracy under adversarial training conditions. Our findings revealed that reduction techniques generally diminished robustness; although, combining techniques was not found to make robustness any worse than when applied individually. Moreover, specific techniques can potentially enhance resistance to small size perturbations. This research provides insights into the trade-offs between model size reduction and security, establishing a foundation for future investigations into improving adversarial training techniques and methodologies for maintaining robustness while preserving memory footprint benefits.

Austria, Phillipe [ORNL] (ORCID:0000000236223973)↗

Chemically Amplified, Dry-Develop Poly(aldehyde) Photoresist

The catalytic decomposition of poly(phthalaldehyde) with a photoacid generator can be used as dry-develop photoresist, where the exposed film depolymerizes into small molecules to allow the development of features via controlled vaporization. Higher temperatures enabled shorter dry-development times, but also promoted faster photoacid diffusion that compromised pattern fidelity. Trihexylamine was used as a base quencher to counteract acid diffusion in a phthalaldehyde-propanal co-polymer photoresist. The propanal co-monomer in the polymer improves the vaporization rate because it has a higher vapor pressure than phthalaldehyde. Addition of the base quencher was found to improve the contrast, pattern fidelity, and ease-of-handling of the dry-develop resist in a direct-write UV lithography tool. The dry-development of 4 μ m features was achieved with no appreciable residue. For large area features, a spatially variable exposure method was used to direct the residue away from the exposed area. The gradient exposure method was used to produce 100 μ m features. Plasma etching after dry-development was also used to achieve residue-free dry-developed patterns. These results show the benefits of incorporating base additives into a dry-develop depolymerizable resist system and highlight the need for addressing residue formation.

Materials Science↗

Preliminary Theoretical Analysis of Mixed Precision Krylov Subspace Methods (Q3 Report)

The third quarter of the project was spent performing theoretical finite precision analysis of Krylov subspace method variants that use mixed precision. Our focus here is on the Conjugate Gradient (CG) method and the Lanczos method. We have performed an analysis of maximum attainable accuracy for the classical CG method in which 3 precisions are used: a working precision ε, a precision ε IP for the inner product computations, and a precision ε MV for the matrix-vector products. Our results show that performing inner product computations in lower precision does not affect the attainable accuracy. Further, we have performed a complete error analysis of the s-step Lanczos algorithm. In this case, we show that the numerical behavior of the algorithm can be significantly improved by using extra precision in a small part of the computation. We summarize the main theorems in the remainder of the document. Other activities include attending biweekly xSDK meetings. The subsequent quarter will be spent finalizing these results into technical reports and/or manuscripts for submission to journals, as well as identifying opportunities for future work.

97 MATHEMATICS AND COMPUTING↗

Contour Mapping Based Radiation Source Localization by UAS Swarm

The unmanned swarm of aerial systems (UAS) can be used for remote radiation sensing and search for source location. Recently, the cost of developing UAS platforms has dropped drastically partly due to the increase in their popularity. The radiation sensing using multiple UAS has becomes economically viable. Different algorithms for determining the location of a radiation source using simultaneous measurements taken from drones in formation have been studied including gradient-based methods. Generally in order to accomplish contour mapping, the swarm of drones with on board sensors use gradient estimation to determine the direction of steepest gradient which governs a bulk heading vector for the swarm to follow. However, some algorithms and test methods are useful only in a specific set of circumstances with particular initial conditions. Besides, implementation of drones in swarm formation was generally not tested in the context of radiation source detection. The objective of this research is to advance the state-of-the-art by conducting experiments using multiple UAS and radiative sources in an indoor flight test-bed. The dynamic flight characteristic of the UAS swarm were considered in the search and mapping algorithms. In this paper, the radiation source detection using algorithms pertaining to swarm formation of UAS, gradient estimation, contour mapping, and experimental flight test-bed are discussed. In addition, more effective and practical ways of constructing the contour maps are explored. (authors)

42 ENGINEERING↗

Significance of Low‐Velocity Zones on Solute Retention in Rough Fractures

Natural fractures are characterized by high internal heterogeneity. This internal variability is the cause of flow channeling, which in turn leads to contaminant transport taking place primarily along the high-velocity channels. Mass exchange between the high-velocity channels and the low-velocity zones has the potential to enhance contaminant retention, due to solute diffusion into the low-velocity zones and subsequent exposure to additional surface area for diffusion into the bordering rock matrix. Here, we derive a random walk particle tracking method for heterogeneous fractures, which includes an additional term to account for the aperture gradient. The method takes into account advection, diffusion in the fracture and matrix diffusion. The developed numerical framework is applied to assess the effect of low-velocity zones in rough self-affine fractures. The results show that diffusion into low-velocity zones has a visible but modest impact on contaminant retention. The magnitude of this impact does not change considerably, regardless of whether diffusion into the rock matrix is considered in the model, and increases for a decreasing average Péclet number of the fracture.

58 GEOSCIENCES↗

Structural and Electronic Effects on the Properties of Fe 2 (dobdc) upon Oxidation with N 2 O

Here, we report electronic, vibrational, and magnetic properties, together with their structural dependences, for the metal–organic framework Fe 2 (dobdc) (dobdc 4 – = 2,5-dioxido-1,4-benzenedicarboxylate) and its derivatives, Fe 2 (O) 2 (dobdc) and Fe 2 (OH) 2 (dobdc)—species arising in the previously proposed mechanism for the oxidation of ethane to ethanol using N 2 O as an oxidant. Magnetic susceptibility measurements reported for Fe 2 (dobdc) in an earlier study and reported in the current study for Fe II 0.26 [Fe III (OH)] 1.74 (dobdc)(DMF) 0.15 (THF) 0.22 , which is more simply referred to as Fe 2 (OH) 2 (dobdc), were used to confirm the computational results. Theory was also compared to experiment for infrared spectra and X-ray powder diffraction structures. Structural and magnetic properties were computed by using Kohn-Sham density functional theory both with periodic boundary conditions and with cluster models. In addition, we studied the effects of different treatments of the exchange interactions on the magnetic coupling parameters by comparing several approaches to the exchange-correlation functional: generalized gradient approximation (GGA), GGA with empirical Coulomb and exchange integrals for 3d electrons (the GGA+U method), nonseparable gradient approximation with empirical Coulomb and exchange integrals for 3d electrons (GAM+U), hybrid GGA, meta-GGA, and hybrid meta- GGA. We found the coupling between the metal centers to be ferromagnetic in the case of Fe 2 (dobdc) and antiferromagnetic in the cases of Fe 2 (O) 2 (dobdc) and Fe 2 (OH) 2 (dobdc). The shift in magnetic coupling behavior correlates with the changing electronic structure of the framework, which derives from both structural and electronic changes that occur upon metal oxidation and addition of the charge-balancing oxo and hydroxo ligands.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Magnetic method for stimulating transport in fluids

A method for producing mass and heat transport in fluids, wherein the method does not rely on conventional convection, that is, it does not require gravity, a thermal gradient, or a magnetic field gradient. This method gives rise to a unique class of vigorous, field-controllable flow patterns termed advection lattices. The advection lattices can be used to transport heat and/or mass in any desired direction using only magnetic fields.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

PCG method with Strang’s circulant preconditioner for Hermitian positive definite linear system in Riesz space fractional advection–dispersion equations

In this paper, preconditioned conjugate gradient (PCG) method with Strang’s circulant preconditioner is investigated to solve the Hermitian positive definite linear systems, which is result from the Crank–Nicolson (C-N) finite difference scheme with the weighted and shifted Grünwald difference (WSGD) operators to discretize the Riesz space fractional advection–dispersion equation (RSFADE). We show that the spectrum of the preconditioned matrix is clustered around 1, and the singular values of the preconditioned matrix are uniformly bounded away from zero under a certain condition, respectively; hence the PCG method, when applied to solving the preconditioned system, converges superlinearly. Moreover, the complexity in each iteration of the PCG method is O(Nlog N) via using the fast Fourier transforms, where N is the matrix size. Numerical experiments are included to demonstrate the effectiveness of our approach.

97 MATHEMATICS AND COMPUTING↗

Comparison of two-dimensional integration methods for shape reconstruction from gradient data

As a requisite and key step in some gradient-based measurement techniques, the reconstruction of the shape, more generally the scalar potential, from the measured gradient data has been studied for many years. Here, in this work, three types of two-dimensional integration methods are compared under various conditions. The merits and drawbacks of each integration method are consequently revealed to provide suggestions in selection of a proper integration method for a particular application.

36 MATERIALS SCIENCE↗

A Hierarchical OPF Algorithm with Improved Gradient Evaluation in Three-Phase Networks

Linear approximation commonly used in solving alternating-current optimal power flow (AC-OPF) simplifies the system models but incurs accumulated voltage errors in large power networks. Such errors will make the primal-dual type gradient algorithms converge to solutions with voltage violation. In this paper, we improve a recent hierarchical OPF algorithm that rested on primal-dual gradients evaluated with a linearized distribution power flow model. Specifically, we propose a more accurate gradient evaluation method based on an unbalanced three-phase nonlinear distribution power flow model to mitigate the errors arising from linearization. The resultant gradients feature a blocked structure that enables our development of an improved hierarchical primal-dual algorithm to solve the OPF problem. Numerical results on the IEEE 123-bus test feeder and a 4,518-node test feeder show that the proposed method can enhance voltage safety at comparable computational efficiency with the linearized algorithm.

approximation algorithms↗

High-order gradients with the shifted boundary method: An embedded enriched mixed formulation for elliptic PDEs

Here, we propose an extension of the embedded boundary method known as “shifted boundary method” to elliptic diffusion equations in mixed form (e.g., Darcy flow, heat diffusion problems with rough coefficients, etc.). Our aim is to obtain an improved formulation that, for linear finite elements, is at least second-order accurate for both flux and primary variable, when either Dirichlet or Neumann boundary conditions are applied. Following previous work of Nishikawa and Mazaheri in the context of residual distribution methods, we consider the mixed form of the diffusion equation (i.e., with Darcy-type operators), and introduce an enrichment of the primary variable. This enrichment is obtained exploiting the relation between the primary variable and the flux variable, which is explicitly available at nodes in the mixed formulation. The proposed enrichment mimics a formally quadratic pressure approximation, although only nodal unknowns are stored, similar to a linear finite element approximation. We consider both continuous and discontinuous finite element approximations and present two approaches: a non-symmetric enrichment, which, as in the original references, only improves the consistency of the overall method; and a symmetric enrichment, which enables a full error analysis in the classical finite element context. Combined with the shifted boundary method, these two approaches are extended to high-order embedded computations, and enable the approximation of both primary and flux (gradient) variables with second-order accuracy, independently on the type of boundary conditions applied. We also show that the primary variable is third-order accurate, when pure Dirichlet boundary conditions are embedded.

97 MATHEMATICS AND COMPUTING↗