SEARCH · Search NASA

Results for “stochastic 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.

SourceResearch connectionFrom yearThrough year

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

At least 163 records · Page 9

All sources NASA NTRS DOE OSTI

DOE OSTI2021

Bayesian inference of heterogeneous epidemic models: Application to COVID-19 spread accounting for long-term care facilities

Here we propose a high dimensional Bayesian inference framework for learning heterogeneous dynamics of a COVID-19 model, with a specific application to the dynamics and severity of COVID-19 inside and outside long-term care (LTC) facilities. We develop a heterogeneous compartmental model that accounts for the heterogeneity of the time-varying spread and severity of COVID-19 inside and outside LTC facilities, which is characterized by time-dependent stochastic processes and time-independent parameters in ~ 1500 dimensions after discretization. To infer these parameters, we use reported data on the number of confirmed, hospitalized, and deceased cases with suitable post-processing in both a deterministic inversion approach with appropriate regularization as a first step, followed by Bayesian inversion with proper prior distributions. To address the curse of dimensionality and the ill-posedness of the high-dimensional inference problem, we propose use of a dimension-independent projected Stein variational gradient descent method, and demonstrate the intrinsic low-dimensionality of the inverse problem. We present inference results with quantified uncertainties for both New Jersey and Texas, which experienced different epidemic phases and patterns. Moreover, we also present forecasting and validation results based on the empirical posterior samples of our inference for the future trajectory of COVID-19.

42 ENGINEERING↗

NASA NTRS1975

Computation of output feedback gains for linear stochastic systems using the Zangnill-Powell Method

Because conventional optimal linear regulator theory results in a controller which requires the capability of measuring and/or estimating the entire state vector, it is of interest to consider procedures for computing controls which are restricted to be linear feedback functions of a lower dimensional output vector and which take into account the presence of measurement noise and process uncertainty. To this effect a stochastic linear model has been developed that accounts for process parameter and initial uncertainty, measurement noise, and a restricted number of measurable outputs. Optimization with respect to the corresponding output feedback gains was then performed for both finite and infinite time performance indices without gradient computation by using Zangwill's modification of a procedure originally proposed by Powell. Results using a seventh order process show the proposed procedures to be very effective.

Kaufman, H.↗

NASA NTRS1977

Computation of output feedback gains for linear stochastic systems using the Zangwill-Powell method

Kaufman, H.↗

DOE OSTI2024

Hutchinson Trace Estimation for high-dimensional and high-order Physics-Informed Neural Networks

Physics-Informed Neural Networks (PINNs) have proven effective in solving partial differential equations (PDEs), especially when some data are available by seamlessly blending data and physics. However, extending PINNs to high-dimensional and even high-order PDEs encounters significant challenges due to the computational cost associated with automatic differentiation in the residual loss function calculation. Herein, we address the limitations of PINNs in handling high-dimensional and high-order PDEs by introducing the Hutchinson Trace Estimation (HTE) method. Starting with the second-order high-dimensional PDEs, which are ubiquitous in scientific computing, HTE is applied to transform the calculation of the entire Hessian matrix into a Hessian vector product (HVP). This approach not only alleviates the computational bottleneck via Taylor-mode automatic differentiation but also significantly reduces memory consumption from the Hessian matrix to an HVP’s scalar output. We further showcase HTE’s convergence to the original PINN loss and its unbiased behavior under specific conditions. Comparisons with the Stochastic Dimension Gradient Descent (SDGD) highlight the distinct advantages of HTE, particularly in scenarios with significant variability and variance among dimensions. We further extend the application of HTE to higher-order and higher-dimensional PDEs, specifically addressing the biharmonic equation. By employing tensor-vector products (TVP), HTE efficiently computes the colossal tensor associated with the fourth-order high-dimensional biharmonic equation, saving memory and enabling rapid computation. The effectiveness of HTE is illustrated through experimental setups, demonstrating comparable convergence rates with SDGD under memory and speed constraints. Additionally, HTE proves valuable in accelerating the Gradient-Enhanced PINN (gPINN) version as well as the Biharmonic equation. Overall, HTE opens up a new capability in scientific machine learning for tackling high-order and high-dimensional PDEs.

Curse of dimensionality↗

DOE OSTI2018

Confronting Models with Regional CO 2 Observations

This research addressed an important DOE science area: the role of natural disturbances in carbon cycling. Our project included both (a) modeling and field measurements and (b) strengthened ties between DOE national laboratories and universities which continue after the project was completed. Our focus was on understanding the consequences of drought and emissions (fire and urban) to atmospheric trace gas composition (carbon dioxide, carbon monoxide, and methane), including both concentration and stable isotope composition. Measurements of these atmospheric gases were used to infer the spatial and temporal patterns of both sources and sinks in the carbon cycle. As part of this effort, we maintained two long-term monitoring networks along a geographic gradient in Utah and Colorado that spans montane forests, urban regions, and oil/gas fields. In this research, we expanded on a well-accepted atmospheric model (STILT-WRF) as a tool to evaluate fluxes in CLM/CESM-related models. We used the STILT model to provide a strong linkage between point atmospheric measurements and the surface parameterizations/emissions that are part of the CLM/CESM models. Overall, this project benefited and supported the DOE Long Term Mission and Goals in four distinct ways: (a) testing of carbon cycle models, (b) testing and evaluating model mechanisms whereby factors in CLM influence trace gas composition in CESM, (c) acquiring high-quality, long-term data on concentrations of carbon dioxide in western USA ecosystems, and (d) by reducing uncertainties associated with the representation of climate-carbon feedbacks in Earth System models through the development of new methods for evaluating model performance. One postdoctoral associate (Henrique Duarte) and one Ph.D. student (Darek Malia) were trained with this support. Nine peer-reviewed publications and 14 poster presentations were produced with support from this project. This proposal addresses a specific science area in the RFP: the role of natural disturbances in carbon cycling. In particular, we will focus on the consequences of drought and fire at landscape to regional scales for atmospheric trace gas concentration and stable isotope composition (CO 2 , CO, and CH 4 ). We seek to expand on the well accepted atmospheric Stochastic Time-Inverted Lagrangian Transport (STILT) model coupled to the Weather Research and Forecasting (WRF) model as a tool to evaluate stable isotope fluxes in CLM/CESM-related models through existing, long-term data sets and field campaigns where high-precision trace gas concentration and isotope measurements can be made along gradients and transects. Given the confounding impacts of anthropogenic trace gas emissions on landscape-to-regional atmospheric observations, we will also measure and parameterize isotopic characteristics of CO 2 , CO, and CH 4 from anthropogenic or forest sources in order to allow us the opportunity to partition atmospheric trace gases into natural and anthropogenic sources.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2025

A Novel LDPP-MADDPG Approach for Distributed Power Allocation in mmWave Cellular Networks

This paper considers the problem of distributed beam scheduling and power allocation problem in millimeter- Wave (mmWave) cellular networks, in which multiple Base Stations (BSs) operate as individual operators over a shared spectrum. We propose a novel learning-aided approach that integrates the Lyapunov Drift-Plus-Penalty (LDPP) framework and Multi-agent Deep Deterministic Policy Gradient (MADDPG) reinforcement learning algorithms. This offers a powerful approach to learning stable and constraint-aware policies, reaping the joint benefit of both LDPP and MADDPG, in complex multiagent environments. The major challenge for this approach is to integrate these two approaches in a meaningful and effective manner. The key idea to solve this problem is to introduce a novel feature of local observation that incorporates potential negative value of the reward function due to the stochastic constraints introduced by the LDPP framework. Empirical results demonstrate that our proposed scheme outperforms the baseline methods under various conditions.

99 - GENERAL AND MISCELLANEOUS↗

DOE OSTI2021

Stochastic Gradient-Based Distributed Bayesian Estimation in Cooperative Sensor Networks

Distributed Bayesian inference provides a full quantification of uncertainty offering numerous advantages over point estimates that autonomous sensor networks are able to exploit. However, fully-decentralized Bayesian inference often requires large communication overheads and low network latency, resources that are not typically available in practical applications. In this paper, we propose a decentralized Bayesian inference approach based on stochastic gradient Langevin dynamics, which produces full posterior distributions at each of the nodes with significantly lower communication overhead. We provide analytical results on convergence of the proposed distributed algorithm to the centralized posterior, under typical network constraints. Finally, we also provide extensive simulation results to demonstrate the validity of the proposed approach.

42 ENGINEERING↗

DOE OSTI2023

Temperature uncertainty modelling with proxy structural data as geostatistical constraints for well siting: an example applied to Granite Springs Valley, NV, USA

Utilizing existing temperature and structural geology information around Granite Springs Valley, Nevada, we build 3D stochastic temperature models with the aims of evaluating the 3D uncertainty of temperature and choosing between candidate exploration well locations. The data used to support the modelling are measured temperatures and structural proxies from 3D geologic modelling (distance to fault, distance to fault intersections and terminations, Coulomb stress change and dilation tendency), the latter considered ‘secondary’ data. Two stochastic geostatistical techniques are explored for incorporating the structural proxies: cosimulation and local varying mean. With both the cosimulation and local varying mean methods, many equally-likely temperature models (i.e. realizations) are produced, from which temperature probability profiles are calculated at candidate well locations. To aid in choosing between the candidate locations, two quantities summarize the temperature probabilities: V prior and entropy. V prior quantifies the likelihood for economic temperatures at each candidate location, whereas entropy identifies where new information has the most potential to reduce uncertainty. In general, the cosimulation realizations have smoother spatial structure, and extrapolate high temperatures at candidate locations that are located along the direction of the longest spatial correlation, which are down dip from existing temperature logs. The smooth realizations result in tight temperature probability profiles that are easier to interpret, but they have unrealistic temperature reversals in some locations because of the dipping ellipsoid shape created and that the cosimulation technique does not enforce a conductive geothermal gradient as a baseline (i.e. linearly increasing temperature with depth). The local varying mean results produce realizations with more realistic geothermal gradients, with temperatures increasing downward since a depth-temperature relationship is included. However, because they have much noisier spatial nature compared to cosimulation, it is harder to interpret the temperature probability profiles. The different local varying mean results allow the geologist to determine which proxy (e.g. dilation v. distance to fault termination) should be used given the specific geothermal system. In general, V prior from local varying mean results identify locations that are close to high values for the structural proxies: areas with higher probabilities for higher temperatures. The entropy results identify where uncertainty is greatest and therefore new drilling information could be most useful. Though these techniques provide useful information, even when applied to areas of sparse data, our comparison of these two techniques demonstrates the need for new geothermal geostatistics techniques that combine the advantages of these two methods and that are tailored to the spatial uncertainty issues inherent in geothermal exploration.

15 GEOTHERMAL ENERGY↗

DOE OSTI2023

MrHyDE v.1.0

SAND2024-01324O MrHyDE, which stands for Multi-resolution Hybridized Differential Equations, is a general-purpose C++ package for the solution of coupled multiphysics and multiscale systems on massively parallel computing systems. MrHyDE is designed to enable moving beyond forward simulation for multiscale applications which includes optimization, control, uncertainty quantification, and stochastic inversion. The framework provides interfaces to several packages within the Trilinos framework and leverages automatic differentiation to enable adjoint capabilities for large-scale, gradient-based optimization. MrHyDE provides automated multiscale capabilities through a subgrid model interface and multiscale Dirichlet-to-Neumann maps. For extreme-scale applications, MrHyDE provides in situ data-compression algorithms to reduce memory requirements while maintaining performance. MrHyDE is a general-purpose, computational framework for the solution of multiscale and multiphysics applications. It uses a combination of structure-preserving, physics-compatible discretizations, fully implicit methods, multi-resolution schemes, or fully explicit methods. Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525.

SciDAC↗

DOE OSTI2018

ISAP-MATLAB Package for Sensitivity Analysis of High-Dimensional Stochastic Chemical Networks

Stochastic simulation and modeling play an important role to elucidate the fundamental mechanisms in complex biochemical networks. The parametric sensitivity analysis of reaction networks becomes a powerful mathematical and computational tool, yielding information regarding the robustness and the identifiability of model parameters. However, due to overwhelming computational cost, parametric sensitivity analysis is a extremely challenging problem for stochastic models with a high-dimensional parameter space and for which existing approaches are very slow. Here we present an information-theoretic sensitivity analysis in path-space (ISAP) MATLAB package that simulates stochastic processes with various algorithms and most importantly implements a gradient-free approach to quantify the parameter sensitivities of stochastic chemical reaction network dynamics using the pathwise Fisher information matrix (PFIM; Pantazis, Katsoulakis, and Vlachos 2013). The sparse, block-diagonal structure of the PFIM makes its computational complexity scale linearly with the number of model parameters. As a result of the gradientfree and the sparse nature of the PFIM, it is highly suitable for the sensitivity analysis of stochastic reaction networks with a very large number of model parameters, which are typical in the modeling and simulation of complex biochemical phenomena. Finally, the PFIM provides a fast sensitivity screening method (Arampatzis, Katsoulakis, and Pantazis 2015) which allows it to be combined with any existing sensitivity analysis software.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

End-to-End Differentiable Modeling and Management of the Environment

Focal Area: (2) Data acquisition and assimilation enabled by machine learning, AI, and advanced methods including experimental/network design/optimization. We emphasize the importance of leveraging optimization techniques from AI/machine learning (ML) to solve challenging problems in Earth system modeling. Science Challenge: Automatic differentiation has had a transformative effect on ML by allowing the calculation of gradients of arbitrary functions in an incredibly large class of models. We can potentially realize similar improvements in parameter estimation and control for Earth system models (ESMs) by reimplementing them in computational frameworks from ML. Practitioners working with large (>10 7 parameters) models in ML and AI can obtain good predictive performance in a range of spatiotemporal tasks by making use of optimization via stochastic gradient descent and incorporating prior knowledge at multiple levels. We propose writing ESMs in open-source computational frameworks such as Torch, Tensorflow, and JAX to greatly expand the scope of environmental forecasting and management challenges, which can be addressed by leveraging automatic differentiation and gradient descent-like algorithms. We do not call for a wholesale replacement of physical models with data-driven surrogates, but rather advocate for interleaving physical and empirical equations in a manner that is most faithful to the extent of our scientific knowledge and observational data. Central to this topic is the merging of differentiable physical simulations with differentiable optimization layers, which are now both beginning to come to the forefront.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

An improved stochastic weighted particle method for boundary driven flows

Here, the stochastic weighted particle method (SWPM) is a generalization of the Direct Simulation Monte Carlo (DSMC) method where particle weights are variable and dynamic. SWPM is backed by a strong theoretical foundation but has not been critically evaluated for problems of practical interest. A thorough assessment of SWPM for boundary-driven flows reveals significant numerical artifacts near the boundary, notably a diverging heat flux. To correct the boundary heat flux, two modifications to SWPM are proposed: separated grouping and a spatially-dependent weight transfer function. To gauge the relative efficiency of SWPM in comparison to DSMC, a high-Mach-number wheel flow which forms a strong density gradient is also simulated.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Differentiable Multiphysics Codes: A Breakthrough Technology for Simulation and Computing

This document summarizes the findings of a strategic planning exercise commissioned by the Weapons Simulation and Computing, Computational Physics (WSC/CP) program at the Lawrence Livermore National Laboratory (LLNL) in FY24. During the year, the committee met with multiple stakeholder communities to gather input, opinions, suggestions and concerns which have been incorporated throughout this document. The key findings from this exercise are summarized: • The development of multiphysics modelling and simulation (mod/sim) codes and software technologies, their deployment on exascale compute platforms, and their broad adoption across the NNSA is a major success of the Advanced Simulation and Computing (ASC) program and the Exascale Computing Project (ECP). Sustained investment in these core technologies is essential. • Today’s state of the art involves running ensembles of O(100K) simulations to perform uncertainty quantification (UQ) and design studies using multiple statistical methods such as Bayesian optimization to understand sensitivities of our models and explore parameterized design spaces. Even with exascale computing, we are practically limited to O(10) parameters in these studies since the number of simulations required to sample the space scales exponentially with the number of design parameters. • The data from these simulation ensembles is increasingly being used to train machine learned (ML) surrogates (or reduced order models, ROMs) which can then be used for optimization or real time design exploration. However, the trained surrogates are still limited in the number of parameters they can represent due to the sampling limitations previously noted. • Augmenting our suite of integrated multiphysics simulation codes, both current and emerging, with the ability to compute gradients (solution derivatives) of arbitrary simulation outputs with respect to (some or all) simulation inputs would be a breakthrough technology, opening the door to a new era of efficient and automated inverse design based on verified and validated mod/sim capabilities. • This capability, which we refer to as differentiable multiphysics codes (DMCs), would revolutionize both UQ and optimization studies by breaking the curse of dimensionality that presently limits our “gradient-free” ensemble based computing approach. A similar breakthrough occurred in the AI/ML community once the ability to compute gradients of arbitrary loss functions using back-propagation became commonplace. Gradient information from the multiphysics codes can also be used to dramatically improve the efficiency and scale of training of ML/ROM surrogates for rapid assessments. • Achieving this in our suite of codes will be a grand challenge, similar to the amount of effort that was required to transition from CPU to GPU computing. It will require buy-in from the entire WSC/CP program and beyond, including all integrated codes, physics and engineering models, third-party library dependencies and performance portability abstractions. It will also require investment in research and development of numerical methods for computing adjoints of coupled physics across multiple adaptively refined moving meshes and of stochastic (Monte Carlo) and mesh free (SPH) methods. • New software and numerical techniques, largely pioneered by the AI/ML community, make this feasible. Chief among these is automatic differentiation (AD), the ability to employ AD at point-wise locations in a physics calculation (instead of traditional black-box approaches) and the ability to perform “back-propagation in time” (or reverse mode AD) for non-linear partial differential equations (PDEs). Fundamentally, the conclusion of this strategic planning exercise is that the time is right to undertake a large scale effort in WSC, centered on the existing integrated codes, to continue the natural evolution of mod/sim in the age of AI/ML. Instead of attempting to replace mod/sim with purely data driven AI/ML models, we believe the key to success is to integrate AI/ML by building on top of the decades of hard-won knowledge and the verified/validated multiphysics modelling capability that is the hallmark of the ASC program.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Prediction of the energetic particle redistribution by an improved critical gradient model and analysis of the transport threshold

Based on the theory of critical gradient model (CGM) and following the simulation method proposed by Waltz et al., a combination of TGLFEP and EPtran code is employed to predict the energetic particle (EP) transport induced by Alfvén eigenmodes (AEs). To be consistent with the experiment, recent improvements to the simulation method include consideration of threshold evolution and orbit loss due to finite orbit width. The revised CGM is applied to simulate two DIII-D experimental discharges (#142111 and #153071). It well reproduces the experimental profiles with multiple unstable AEs and large-scale EP transport. Discharge #142111 had previously been simulated using a nonlinear MHD-kinetic code MEGA with a transport mechanism based on stochasticity induced by overlapping AE. By comparing the simulated EP profiles, we find that the AE transport threshold is approximated by both the MEGA nonlinear stability threshold and the proposed CGM threshold (error <5% for single n and <17% for multiple n simulation). Both of them are larger than the linear stability threshold of the most unstable AE mode by a quantity of the order of the flux needed to sustain EP transport by the background turbulence. Additionally, we have also applied the improved CGM to simulate the α particle redistribution for a China Fusion Engineering Test Reactor steady state scenario. Because of the clear separation between the AE unstable region and the loss cone, only a moderate α particle loss of ~9.6% is predicted.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2019

Capacity assessment and co-optimization of CO 2 storage and enhanced oil recovery in residual oil zones

Residual oil zones (ROZs) are increasingly being commercially exploited using CO 2 -enhanced oil recovery (CO 2 -EOR) method. In this study, CO 2 storage potential, long-term CO 2 fate and oil recovery potential in ROZs are characterized based on a reservoir model for Goldsmith-Landreth San Andres Unit in the Permian Basin. The effects of CO 2 injection rates, well patterns (five-spot and line-drive), well spacings, injection modes (continuous CO 2 injection and water-alternatinggas injection) on the CO 2 retention in the reservoir and the oil production are investigated. After the preliminary assessment of CO 2 storage and EOR potentials in ROZs, we next develop a novel approach based on a newly developed optimization algorithm-Stochastic Simplex Approximate Gradient (StoSAG) and predictive empirical models constructed using machine learning technique to co-optimize CO 2 storage and oil recovery in ROZs. The performance of co-optimization of CO 2 storage and oil recovery is compared with the performance of optimization of only CO 2 storage.

58 GEOSCIENCES↗

DOE OSTI2019

The Fragment Molecular Orbital Method Based on Long-Range Corrected Density-Functional Tight-Binding

The presently available linear scaling approaches to density-functional tight-binding (DFTB) based on the fragment molecular orbital (FMO) method are severely impacted by the problem of artificial charge transfer due to the self-interaction error (SIE), which hampers the simulation of zwitterionic systems such as biopolymers or ionic liquids. Here we report an extension of FMO-DFTB where we included a long-range corrected (LC) functional designed to mitigate the DFTB SIE, called the FMO-LC-DFTB method, resulting in a robust method which succeeds in simulating zwitterionic systems. Both energy and analytic gradient are developed for the gas phase and the polarizable continuum model of solvation. The scaling of FMO-LC-DFTB with system size N is shown to be almost linear, O( N 1.13–1.28 ), and its numerical accuracy is established for a variety of representative systems including neutral and charged polypeptides. It is shown that pair interaction energies between fragments for two mini-proteins are in excellent agreement with results from long-range corrected density functional theory. Here the new method was employed in long time scale (1 ns) molecular dynamics simulations of the tryptophan cage protein (PDB: 1L2Y) in the gas phase for four different protonation states and in stochastic global minimum structure searches for 1-ethyl-3-methylimidazolium nitrate ionic liquid clusters containing up to 2300 atoms.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

Rapid Optimization of Total Variation with Applications in Imaging, Additive Manufacturing, and Qualification

Total Variation optimization penalizes the gradient of a control variable or state. While this work focuses on image processing in particular, it has also found applications in inverse problems and topology optimization. In image processing, the goal is to maintain faithfulness to the original image while denoising and/or deblurring. Additionally, bilevel optimization over the spatially varying regularization weights can illuminate interfaces such as damage regions and other anomalies. We will address two fundamental challenges with TV-optimization: (i) the typical slow convergence of existing TV-optimization methods, and (ii) the selection of spatially varying TV parameters to promote interface detection. Additionally, we will apply such techniques to image data collected in additive manufacturing. In said context, stochasticity in build events induces flaws in the manufactured piece, compromising the integrity of said part. There is a critical need for in-situ monitoring to spot anomalies once they form, and in this setting we apply our total variation and hyperparameter solvers. We will develop a customized algorithm based on for extreme-scale TV-optimization that achieves super-linear or quadratic-convergence, a critical property for real-time, image-by-image analysis. A worst-case outcome is a preprocessing step that enhances image quality in-situ, specifically for out-of-focus and noisy images.

36 MATERIALS SCIENCE↗

DOE OSTI2023

Optimization of Distribution Feeder Topology: A Differential Programming Learning Approach

This paper presents a gradient based method for optimizing distribution feeder network topology under load un- certainty. We recast the optimal network reconfiguration problem as a learning problem where edge weights of a graph are learned to produce an optimized spanning tree for a distribution network. Using recent methods published on differentiable programming, we provide a data driven method for learning these weights. We test our method on 100 variations of an IEEE 15-bus test system. Our results show that our method outperforms more traditional mathematical programming-based approaches.

differentiable programming↗