Search NASA⌕ Search

SEARCH · Search NASA

Results for “Tensor train”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2024

Chapter 4: Physically informed deep learning networks for simulating microstructure evolution of 3D polycrystals

As discussed in the previous chapter, high energy diffraction microscopy (HEDM) is used to study the micromechanical evolution of a material during in situ loading. HEDM experiments have been used to verify crystal plasticity (CP) simulations [119, 91, 90, 120], for experimental planning, material design, and to further analyze experimental results. However, Fast Fourier transform-based CP (CP-FFT) or finite element-based CP (CP-FE) methods are often too slow to be used in real-time during an experiment. CP-FFT is faster than CP-FE simulations due to the absence of meshing, but can still take hours to simulate the response of a single volume depending on the size and number of strain steps [127]. Reducing computation time would create a larger exploration space in planning and design, and enable faster analysis of experimental results and real-time feedback during an experiment. This research expands upon previous works to develop a workflow for predicting the full-field evolution of a 3D polycrystal. The workflow is simplified from previous works to predict only orientation and elastic strain tensors (from which stress tensors are calculated). The network is physically informed through loss functions and network architecture for a more robust model. The orientation predictions are informed about the cubic crystal symmetry of the material by incorporating disorientation and misorientation information into the network architecture and loss. The Von Mises stress is used to enforce the correct stress-strain trends in the strain tensor predictions. Additional total strain steps from the elastic and elastoplastic region are included to better capture the stress-strain evolution at smaller total strain steps. Material and hardening parameters are additional inputs into the networks to further inform the network and to study the network’s ability to predict different materials other than those used for training.

36 MATERIALS SCIENCE↗
DOE OSTI2014

TU-C-12A-02: Development of a Multiparametric Statistical Response Map for Quantitative Imaging

Purpose: Quantitative imaging biomarkers (QIB) are becoming increasingly utilized in early phase clinical trials as a means of non-invasively assessing treatment response and associated response heterogeneity. The aim of this study was to develop a flexible multiparametric statistical framework to predict voxel-by-voxel response of several potential MRI QIBs. Methods: Patients with histologically proven glioblastomas (n=11) were treated with chemoradiation (with/without bevacizumab) and underwent one baseline and two mid-treatment (3–4wks) MRIs. Dynamic contrast-enhanced (3D FSPGR, 6.3sec/phase, 0.1 mmol/kg Gd-DTPA), dynamic susceptibility contrast (2D GRE-EPI, 1.5sec/phase, 0.2mmol/kg Gd-DTPA), and diffusion tensor (2D DW-EPI, b=0, 1200 s/mm{sup 2}, 27 directions) imaging acquisitions were obtained during each study. Mid-treatment and pre-treatment images were rigidly aligned, and regions of partial response (PR), stable disease (SD), and progressive disease (PD) were contoured in consensus by two experienced radiation oncologists. Voxels in these categories were used to train ordinal (PR<SD<PD) and logistic (PR/SD<PD) models incorporating the apparent diffusion coefficient, fractional anisotropy, Ktrans, ve, vp, and relative cerebral blood volume imaging biomarkers, as well as treatment type. Leave-one-out cross-validation was performed at the patient level to assess model prediction accuracy. Results: Ordinal regression resulted in model prediction accuracies of 60% (PR), 0% (SD), 81% (PD), and 69% (overall), with coefficients of variation (COV) of 9.4%, 9.6%, and 23.6%, respectively. Logistic regression resulted in accuracies of 82.0% (PR/SD), 46.2% (PD), and 76.2% (overall) with COVs of 22.4%, 45.7%, and 23.8%, respectively. Conclusion: Despite limited patient numbers, this feasibility pilot study demonstrates that ordinal and logistic regression models potentially provide a flexible statistical framework for incorporating longitudinal multiparametric quantitative imaging data and, for these particular data, can potentially distinguish PR from PD. Additional training data exhibiting a full range of treatment responses is expected to allow prediction of all three treatment outcomes. Moreover, these models are extensible to multi-modality, multiparametric analyses.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2023

AdversarialTensors

This library builds a framework for defending ML models against adversarial attacks. The library will be developed at various stages leading to publication and software release at each stage. We employ tensor decomposition strategies as preprocessing stages for the first stage to provide robustness against the prominent adversarial noise. In the second stage, we develop a latent noise generator capable of generating novel adversarial noise that threatens the existing state-of-the-art defense strategy. In the third stage, we develop a UNSUP-GAN model, where the generator is trained to denoise against latent noise and most adversarial noises. This generator can provide a robust adversarial attack against any unseen attack.

Bhattarai, Manish↗
DOE OSTI2023

Reduced Order Model to Predict Dispersion of Flammable Refrigerant into a Space

As the HVAC&R industry mobilizes to deploy more low-GWP refrigerants, relevant standards are being continually reviewed and updated. Those include the general safety standards ISO 5149 and ASHRAE 15, and the equipment standards IEC and UL. The standards systematically set the allowable maximum amount of refrigerant that should be used in different equipment types and different applications. To do so, they rely on predictions of how a leaked refrigerant mass will disperse into a space. Dispersion characteristics, such as total flammable volume and its residence time, determine the risk associated with the presence of the flammable refrigerant. The standards have included provisions for the use of flammable refrigerants for approximately two decades. They relied on limited analytical analyses and test cases in their development. Dispersion of a refrigerant into a space is complex. Computational fluid dynamics (CFD) are the most accurate in predicting a given problem. However, CFD is computationally expensive and requires specialized expertise and resources and is not suitable for use by standards development working group as prediction tool. This paper presents the development of a reduced order model (ROM) that predicts the key dispersion characteristics relevant to the dispersion of a leaked refrigerant into a space for any combination of input variables. The inputs are the refrigerant release height, the total released refrigerant mass and its release flow rate, the refrigerant molecular weight, the ventilation flow rate, the floor area and height of the space, recirculation air flow rate, and the tightness of the space. The outputs are histograms of volume fraction of the room in prescribed concentration bins and the total mass of the refrigerant in each bin normalized by the total refrigerant charge at 13 prescribed simulation time stamps between 1 and 900 seconds. The ROM is constructed from a set of CFD simulations with carefully chosen combinations of input parameters. The selection if done using a multidimensional sparse grid which is a generalization of the classical tensor approach but offers additional flexibility and thus can be more carefully tuned towards a specific model. The tuning is done to improve the accuracy, measured in the difference between the output values of the ROM and the CFD model, while minimizing the computational cost, measured in number of CFD simulations which is orders of magnitude more expensive than the processing the training data.

Edwards, Dean↗
DOE OSTI2014

SU-D-BRF-06: A Brachytherapy Simulator with Realistic Haptic Force Feedback and Real-Time Ultrasounds Image Simulation for Training and Teaching

Purpose: Surgical procedures require dexterity, expertise and repetition to reach optimal patient outcomes. However, efficient training opportunities are usually limited. This work presents a simulator system with realistic haptic force-feedback and full, real-time ultrasounds image simulation. Methods: The simulator is composed of a custom-made Linear-DELTA force-feedback robotic platform. The needle tip is mounted on a force gauge at the end effector of the robot, which responds to needle insertion by providing reaction forces. 3D geometry of the tissue is using a tetrahedral finite element mesh (FEM) mimicking tissue properties. As the needle is inserted/retracted, tissue deformation is computed using a mass-tensor nonlinear visco-elastic FEM. The real-time deformation is fed to the L-DELTA to take into account the force imparted to the needle, providing feedback to the end-user when crossing tissue boundaries or needle bending. Real-time 2D US image is also generated synchronously showing anatomy, needle insertion and tissue deformation. The simulator is running on an Intel I7 6- core CPU at 3.26 MHz. 3D tissue rendering and ultrasound display are performed on a Windows 7 computer; the FEM computation and L-DELTA control are executed on a similar PC using the Neutrino real-time OS. Both machines communicate through an Ethernet link. Results: The system runs at 500 Hz for a 8333-tetrahedron tissue mesh and a 100-node angular spring needle model. This frame rate ensures a relatively smooth displacement of the needle when pushed or retracted (±20 N in all directions at speeds of up to 2 m/s). Unlike commercially-available haptic platforms, the oblong workspace of the L-DELTA robot complies with that required for brachytherapy needle displacements of 0.1m by 0.1m by 0.25m. Conclusion: We have demonstrated a real-life, realistic brachytherapy simulator developed for prostate implants (LDR/HDR). The platform could be adapted to other sites or training for other types of needle-based procedures.

62 RADIOLOGY AND NUCLEAR MEDICINE↗
DOE OSTI2014

Automated Algorithms for Quantum-Level Accuracy in Atomistic Simulations: LDRD Final Report.

This report summarizes the result of LDRD project 12-0395, titled "Automated Algorithms for Quantum-level Accuracy in Atomistic Simulations." During the course of this LDRD, we have developed an interatomic potential for solids and liquids called Spectral Neighbor Analysis Poten- tial (SNAP). The SNAP potential has a very general form and uses machine-learning techniques to reproduce the energies, forces, and stress tensors of a large set of small configurations of atoms, which are obtained using high-accuracy quantum electronic structure (QM) calculations. The local environment of each atom is characterized by a set of bispectrum components of the local neighbor density projected on to a basis of hyperspherical harmonics in four dimensions. The SNAP coef- ficients are determined using weighted least-squares linear regression against the full QM training set. This allows the SNAP potential to be fit in a robust, automated manner to large QM data sets using many bispectrum components. The calculation of the bispectrum components and the SNAP potential are implemented in the LAMMPS parallel molecular dynamics code. Global optimization methods in the DAKOTA software package are used to seek out good choices of hyperparameters that define the overall structure of the SNAP potential. FitSnap.py, a Python-based software pack- age interfacing to both LAMMPS and DAKOTA is used to formulate the linear regression problem, solve it, and analyze the accuracy of the resultant SNAP potential. We describe a SNAP potential for tantalum that accurately reproduces a variety of solid and liquid properties. Most significantly, in contrast to existing tantalum potentials, SNAP correctly predicts the Peierls barrier for screw dislocation motion. We also present results from SNAP potentials generated for indium phosphide (InP) and silica (SiO 2 ). We describe efficient algorithms for calculating SNAP forces and energies in molecular dynamics simulations using massively parallel computers and advanced processor ar- chitectures. Finally, we briefly describe the MSM method for efficient calculation of electrostatic interactions on massively parallel computers.

Thompson, Aidan Patrick↗
DOE OSTI2025

Multiplexing Focusing Analyzer for Efficient Stress-Strain Measurements

Statement of the problem or situation that is being addressed. Although thermal and cold neutron scattering is widely used and is critical for success in many areas of materials science and engineering, relatively low neutron fluxes severely limit applications of not only laboratory neutrons generators, but also large national neutron facilities. State-of-the-art thermal and cold neutron sources are large expensive national facilities, which serve diverse community of scientific and industrial users. The constant need to improve the instruments performance, stems from the fact that neutron methods are gaining in popularity, and becoming more and more powerful, while new neutron sources are not being constructed to keep pace with the developments and needs of the scientific community. Small research reactors at universities and National Labs, and laboratory-based neutron generators, are necessary not only for education and training, but also when samples cannot be transported to other facilities. However, the standard neutron techniques, which were developed for high-flux facilities, require much higher efficiencies to be used effectively with the low fluxes of small sources. Thus, the efficient use of neutron sources, such as with our proposed analyzer, is important for the progress and broader use of these neutron techniques. General statement of how this problem is being addressed. We propose to design and demonstrate novel diffractive optical device, which will enable very efficient residual stress neutron diffractometers. The proposed device will be a multi-foil analyzer, where each foil is constructed of focusing bent single crystals of Si. Such device will enable polychromatic residual stress neutron diffraction. At large national facilities, such as at Oak Ridge National Laboratory, these analyzers would enable very fast measurements for determining residual stress tensors, raster large samples or screen multiple samples. Commercial Applications and Other Benefits The outcome of this project would be the demonstration of commercial devices, novel neutron optical components, which could be utilized to improve the performance of existing instruments or build novel neutron scattering instruments at DOE neutron facilities and commercial laboratory neutron sources. These new devices will widen the scope of research conducted using neutrons and enable measurements not feasible at present. Summary for Members of Congress Thermal and cold neutron beams are a powerful materials science probe, which provide unique information about the structure of matter. The proposed innovations expand the reach of neutron-based investigations to new materials and industries by enabling new instrumentation capabilities, thereby greatly enhancing and expanding the role of small, laboratory-based neutron instrumentation, and improving education and training of neutron users.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2023

Machine Learning Full NMR Chemical Shift Tensors of Silicon Oxides with Equivariant Graph Neural Networks

The nuclear magnetic resonance (NMR) chemical shift tensor is a highly sensitive probe of the electronic structure of an atom and furthermore its local structure. Recently, machine learning has been applied to NMR in the prediction of isotropic chemical shifts from a structure. Current machine learning models, however, often ignore the full chemical shift tensor for the easier-to-predict isotropic chemical shift, effectively ignoring a multitude of structural information available in the NMR chemical shift tensor. Here we use an equivariant graph neural network (GNN) to predict full 29 Si chemical shift tensors in silicate materials. The equivariant GNN model predicts full tensors to a mean absolute error of 1.05 ppm and is able to accurately determine the magnitude, anisotropy, and tensor orientation in a diverse set of silicon oxide local structures. When compared with other models, the equivariant GNN model outperforms the state-of-the-art machine learning models by 53%. The equivariant GNN model also outperforms historic analytical models by 57% for isotropic chemical shift and 91% for anisotropy. The software is available as a simple-to-use open-source repository, allowing similar models to be created and trained with ease.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Bridging the Gap Between LLMs and LNS with Dynamic Data Format and Architecture Codesign

Deep neural networks (DNNs) have achieved tremendous success in the past few years. However, their training and inference demand exceptional computational and memory resources. Quantization has been shown as an effective approach to mitigate the cost, with the mainstream data types reduced from FP32 to FP16/BF16 and recently FP8 in the latest NVIDIA H100 GPUs. With increasingly aggressive quantization, however, the conventional floating-point formats suffer from limited precision in representing numbers around zero. Recently, NVIDIA demonstrated the potential of using a Logarithmic Number System (LNS) for the next generation of tensor cores. While LNS mitigates the hurdles in representing small numbers, in this work we observed a mismatch between LNS and the emerging Large Language Models (LLM), where LLM exhibits significant outliers when directly adopting the LNS format. In this paper, we present a data-format/architecture codesign to bright this gap. On the format side, we propose a dynamic LNS format to flexibly represent outliers at a higher precision, by exploiting asymmetry in the LNS representation and identifying outliers through a per-vector basis. On the architecture side, for demonstration, we realize the dynamic LNS format in a systolic array, which can handle the irregularity of the outliers at runtime. We implement our approach on an Alveo U280 FPGA as a prototype. Experimental results show that our design can effectively handle the outliers and resolve the mismatch between LNS and LLM, contributing to an accuracy improvement of 15.4% and 16% over the floating-point and the original LNS baselines, using four state-of-the-art LLM models. Our observation and design lay a solid foundation for the large-scale adoption of the LNS format in the next-generation deep learning hardware.

Haghi, Pouya↗
DOE OSTI2019

Axisymmetric and Focusing Analyzers to Enable Efficient Powder and Residual Stress Neutron Diffractometers

Project Summary: Company: Adelphi Technology, Inc., Redwood City, CA 94063-4121 DOE Phase 1 STTR PI: Dr. Jay Theodore Cremer, Jr. Chief Scientist at Adelphi Technology, Inc. DOE Phase 1 STTR with: Dr. Boris Khaykovich of Massachusetts Institute of Technology (MIT) MIT Nuclear Reactor Laboratory DOE Phase 1 STTR Title: Axisymmetric and Focusing Analyzers to Enable Efficient Powder and Residual Stress Neutron Diffractometers Statement of the problem or situation that is being addressed: Although thermal and cold neutron scattering is widely used, and is critical for success in many areas of materials science and engineering, relatively low neutron fluxes severely limit applications of not only laboratory neutrons generators, but also large national neutron facilities. State-of-the-art thermal and cold neutron sources are large expensive national facilities, which serve diverse community of scientific and industrial users. The constant need to improve the instruments performance, stems from the fact that neutron methods are gaining in popularity, and becoming more and more powerful, while new neutron sources are not being constructed, so as to keep pace with the developments and needs of the scientific community. Small research reactors at universities and National Labs, and laboratory-based neutron generators, are necessary not only for education and training, but also when samples cannot be transported to other facilities. However, the standard neutron techniques, which were developed for high-flux facilities, require much higher efficiencies to be used effectively with the low fluxes of small sources. Thus, the efficient use of neutron sources, such as with our proposed analyzer, is important for the progress and broader use of these neutron techniques. General statement of how this problem is being addressed: We strive to achieve the goal of efficient use of available thermal and cold neutrons, by introducing here, novel analyzers for powder diffraction and stress-strain measurements. These analyzers will enable polychromatic cold and thermal neutron diffractometers, which will have much higher throughput than existing instruments. At large national facilities, such as at Oak Ridge National Laboratory, these analyzers would enable very fast instruments for screening of multiple samples, for measuring kinetic processes, or for measuring residual stress tensors. Commercial Applications and Other Benefits: By combining the novel analyzers with the Adelphi thermal neutron generator, we will open a new market, which includes the testing of samples that cannot be transported, such as nuclear fuels, or the testing of parts during production. Given the reasonable cost and size of our proposed analyzer systems, along with the continuing development of more powerful thermal neutron generators, the proposed analyzers will be suitable for weak neutrons sources, which are installed at small to large university, industrial, and government laboratories. Key Words: Neutron powder diffraction, residual stress measurements, thermal neutron generators, TRIGA reactors, low flux thermal and cold neutron sources, neutron diffraction. Summary for Members of Congress: Thermal and cold neutron beams are a powerful materials science probe, which provide unique information about the structure of matter. The proposed innovations expand the reach of neutron-based investigations to new materials and industries by enabling new instrumentation capabilities, thereby greatly enhancing and expanding the role of small, laboratory-based neutron instrumentation, and improving education and training of neutron users.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2026

A User-Friendly GUI Tool for Automated Microstructural Analysis of Fiber-Reinforced Composites and Porous Structures

Understanding and quantifying microstructural features such as fiber orientation and porosity is critical for predicting the mechanical behavior and performance of fiber-reinforced polymer composites. Traditional manual analysis is time-consuming, subjective, and unsuitable for high-throughput datasets. We present a graphical user interface (GUI) application that automates the analysis of microscopy images to extract key microstructural metrics, including fiber orientation tensors, fiber orientation distribution, porosity and pore size distribution. The app integrates multiple image segmentation techniques including global and local thresholding, clustering, and region-based approaches, offering flexibility for different types of image qualities and features. Users can load microstructural images, select regions of interest and segmentation techniques tailored to their image dataset. It also addresses a critical challenge in fiber orientation analysis: the ambiguities caused by touching, overlapping, or partially cut fibers. It supports autorun examples for standardized workflows, enabling reproducible analysis and facilitating training and benchmarking. This tool significantly reduces manual intervention, enhances consistency, and accelerates data generation for structure–property modeling, process optimization, and digital materials research. The tool is intended for use by materials scientists, engineers, and researchers engaged in composite characterization, quality control, and machine learning-based microstructural studies.

Chawla, Komal [ORNL] (ORCID:0000000190327565)↗
DOE OSTI2019

DeepCMB: Lensing reconstruction of the cosmic microwave background with deep neural networks

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitational lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2019

DeepCMB: Lensing Reconstruction of the CMB with Deep Neural Networks

Next-generation cosmic microwave background (CMB) experiments will have lower noise and therefore increased sensitivity, enabling improved constraints on fundamental physics parameters such as the sum of neutrino masses and the tensor-to-scalar ratio r. Achieving competitive constraints on these parameters requires high signal-to-noise extraction of the projected gravitational potential from the CMB maps. Standard methods for reconstructing the lensing potential employ the quadratic estimator (QE). However, the QE performs suboptimally at the low noise levels expected in upcoming experiments. Other methods, like maximum likelihood estimators (MLE), are under active development. In this work, we demonstrate reconstruction of the CMB lensing potential with deep convolutional neural networks (CNN) - ie, a ResUNet. The network is trained and tested on simulated data, and otherwise has no physical parametrization related to the physical processes of the CMB and gravitati onal lensing. We show that, over a wide range of angular scales, ResUNets recover the input gravitational potential with a higher signal-to-noise ratio than the QE method, reaching levels comparable to analytic approximations of MLE methods. We demonstrate that the network outputs quantifiably different lensing maps when given input CMB maps generated with different cosmologies. We also show we can use the reconstructed lensing map for cosmological parameter estimation. This application of CNN provides a few innovations at the intersection of cosmology and machine learning. First, while training and regressing on images, we predict a continuous-variable field rather than discrete classes. Second, we are able to establish uncertainty measures for the network output that are analogous to standard methods. We expect this approach to excel in capturing hard-to-model non-Gaussian astrophysical foreground and noise contributions.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

A review of the application of 2D isotropic-anisotropic correlation NMR spectroscopy in structural studies of chalcogenide glasses

The application of solid-state high-resolution NMR spectroscopy in the structural investigation of chalcogenide glasses in Ge/As/P/Si-X (X = S, Se,Te) systems has remained challenging even for the spin-1/2 nuclides ( 29 Si, 31 P, 77 Se, 125 Te), owing to their low natural abundance (except for 31 P), slow spin-lattice relaxation rate and large CSA and chemical shift distribution induced line broadening effects. However, most of these deleterious effects can be successfully overcome in two-dimensional (2D) isotropic-anisotropic correlation NMR experiments, especially when performed at high magnetic field and in conjunction with the Car-Purcell-Meiboom-Gill (CPMG) echo train acquisition. Here, we present a short introduction to the basic principles of such experiments and review their applications over the last decade in deciphering various short- and intermediate- range structural characteristics of chalcogenide glasses in S-Se, Se-Te, Ge-Se, As-Se and Si-Se systems as well as in investigating the molecular dynamics in a P-Se supercooled liquid. We anticipate possible future applications of these 2D isotropic-anisotropic correlation NMR experiments, particularly in conjunction with density functional theory-based calculations of NMR chemical shift tensor parameters and additional signal enhancement schemes, in addressing complex structural correlations and distributions in chalcogenide glasses.

2D NMR↗
DOE OSTI2024

Evaluating Physics-Informed Neural Network Performance for Seismic Discrimination between Earthquakes and Explosions

In this article, we evaluate adding a weak physics constraint, that is, a physics‐based empirical relationship, to the loss function with a physics‐informed manner in local distance explosion discrimination in the hope of improving the generalization capability of the machine learning (ML) model. We compare the proposed model with the two‐branch model we previously developed, as well as with a pure data‐driven model. Unexpectedly, the proposed model did not consistently outperform the pure data‐driven model. By varying the level of inconsistency in the training data, we find this approach is modulated by the strength of the physics relationship. In conclusion, this result has important implications for how to best incorporate physical constraints in ML models.

58 GEOSCIENCES↗
DOE OSTI2023

BuildingsBench: A Benchmark for Universal Building Load Forecasting [SWR-23-51]

The residential and commercial building stock in the United States is responsible for a significant percentage of energy consumption and greenhouse gas emissions. Electrification of end-uses, as well as decarbonizing the electrical grid through renewable energy sources such as solar and wind, constitutes the pathway to zero-emission buildings. Forecasting day-ahead building energy consumption is an integral part of this solution. Currently, specialized forecasting models are hand-made for each individual building, which is time-consuming, expensive, and leads to duplicated efforts. BuildingsBench is a Python software framework for training and comparing generalized machine learning models for universal building load forecasting. This challenge tasks a single foundational model to generalize its forecasts for a wide variety of buildings, across geographic regions, building types, weather patterns, and more. This software provide code for pre-training such models and subsequently evaluating their performance on a suite of hundreds of diverse real and synthetic buildings. BuildingsBench is a platform for: - Large-scale pretraining with the synthetic Buildings-900K dataset for short-term load forecasting (STLF). Buildings-900K is statistically representative of the entire U.S. building stock and is extracted from the NREL End-Use Load Profiles database. - Benchmarking on two tasks evaluating generalization: zero-shot STLF and transfer learning for STLF. We provide an index-based PyTorch Dataset for large-scale pretraining, easy data loading for multiple real building energy consumption datasets as PyTorch Tensors or Pandas DataFrames, simple (persistence) to advanced (transformer) baselines, metrics management, and more.

Emami, Patrick↗
DOE OSTI2024

The use of digital thread for reconstruction of local fiber orientation in a compression molded pin bracket via deep learning

A deep convolutional neural network (DCNN) was used for microstructure reconstruction using artificial intelligence (MR-AI) by predicting local average fiber orientation distributions (FOD) in a 3D prepreg platelet molded composite (PPMC) pin bracket. To train the MR-AI model, surface strain fields from residual stresses simulated in PPMC plates were used as the input to the DCNN. A training dataset included PPMC plates with various degrees of global fiber alignment, based on the information obtained from high-fidelity flow simulation of a pin bracket. Further, the MR-AI model was then deployed to analyze FOD in the 3D pin bracket by conducting thermo-elastic residual stress analysis. Initially, the MR-AI model was established entirely on the synthetic simulation data. Then, a μCT scan of a physically molded pin bracket was used to create a finite element model that provided data for additional validation of the DCNN model. For the μCT scan finite element pin bracket the MR-AI model predicted the distribution of fiber orientation tensor components with MAE of 0.10 indicating a global prediction error of 10%. For the flow simulated pin bracket, the MR-AI model predicted the distribution of fiber orientation tensor components with a global prediction error of 11%. The MR-AI model showed the ability to predict regions of varying alignment in the base and flange of the pin bracket. The proposed MR-AI methodology allows for rapid prediction of FOD in geometrically complex parts and offers a promising path to detecting unique fiber orientation states in molded components.

42 ENGINEERING↗
DOE OSTI2024

Distributed-Memory Sparse Deep Neural Network Inference Using Global Arrays

Partitioned Global Address Space (PGAS) models exhibit tremendous promise in developing efficient and productive distributed-memory parallel applications. They have been used extensively in scientific computations due to conveniently offering a ``shared-memory''-like model and convenient interfaces that separate communication with synchronization. Traditionally, PGAS communication models have been applied to dense/contiguously distributed data, but most modern applications depict varied levels of sparsity. Existing PGAS models require certain adaptations to support distributed sparse computations, since associated computations often require matrix arithmetic, in addition to data movement. The Global Arrays toolkit from Pacific Northwest National Laboratory (PNNL) is one of the earliest PGAS models to combine one-sided data communication and distributed matrix operations and is still used in the popular NWChem quantum chemistry suite. Recently, we have expanded the Global Arrays toolkit to support common sparse operations, like sparse matrix-dense matrix multiplies (SpMM), sparse matrix-sparse matrix multiplication (SpGEMM) and Sampled Dense-Dense Matrix Multiplication (SDDMM). As it turns out, these operations are the bedrock of sparse Deep Learning (DL); sparse deep neural networks and Graph Neural Networks (GNNs) have gained increasing attention recently in achieving speedups on training and inference with reduced memory footprints. Unlike scientific applications in High Performance Computing (HPC), modern (distributed-memory capable) DL toolkits often rely on non-standardized and closed-source vendor software optimizations, creating challenges in software-hardware co-design at scale. Our goal is to support a variety of distributed-memory sparse matrix operations and helper functions in the newly created Sparse Global Arrays (SGA), such that it is possible to build portable and productive Machine Learning scenarios for algorithm/software and hardware codesign purposes. Contemporary data-parallel schemes for training/inference are undergoing a major overhaul since model replication limits scalability and causes resource inefficiencies. As such, we have adopted tensor parallelism in decomposing the model and inputs, to mitigate memory issues. Current implementation is built on top of MPI and uses CPUs to maximize the portability across the platforms.

Distributed computing, machine learning↗