SEARCH · Engineering Papers
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.
NeuralMie was developed to improve the representation of aerosol optical properties in E3SM, the DOE's Earth system model. NeuralMie is able to closely emulate numerical solutions to the problem at a fraction of the computational cost. TAMie was written to generate training data for NeuralMie.
In this work, we detail two uncertainty quantification (UQ) methods that provide complementary information. Readout ensembling, by finetuning only the readout layers of an ensemble of foundation models, provides information about model uncertainty. Amending the final readout layer to predict upper and lower quantiles replaces point predictions with distributional predictions, which provide information about uncertainty within the underlying training data. We demonstrate our approach with the MACE-MP-0 model, applying UQ to both the foundation model and a series of finetuned models. The uncertainties produced by the ensemble and quantile methods are demonstrated to be distinct measures by which the quality of the NNP output can be judged.
GenomeFace is meta-genome binning software. Metagenomic binning, the process of grouping DNA sequences into taxonomic units, is critical for understanding the functions, interactions, and evolutionary dynamics of microbial communities. We propose a deep learning approach to binning using two neural networks, one based on composition and another on environmental abundance, dynamically weighting the contribution of each based on characteristics of the input data. Trained on over 43,000 prokaryotic genomes, our network for composition-based binning is inspired by metric learning techniques used for facial recognition. Using a task-specific, multi-GPU accelerated algorithm to cluster the embeddings produced by our network, our binner leverages marker genes observed to be universally present in nearly all taxa to grade and select optimal clusters of sequences from a hierarchy of candidates. We evaluate our approach on four simulated datasets with known ground truth. Our linear time integration of marker genes recovers more near complete genomes than state of the art but computationally infeasible solutions using them, while being over an order of magnitude faster. Finally, we demonstrate the scalability and acuity of our approach by testing it on three of the largest metagenome assemblies ever performed. Compared to other binners, we produced 47%-183% more near complete genomes. From these datasets, we find over the genomes of over 3000 new candidate species which have never been previously cataloged, representing a potential 4% expansion of the known bacterial tree of life.
SF-26-078 This software provides a high-level workflow for satellite-based flood and surface-water mapping using Sentinel-1 SAR and Sentinel-2 multispectral imagery. It supports preparation of remote-sensing data, training and evaluation of deep-learning models, and generation of high-resolution water extent predictions, including methods to improve robustness to SAR speckle noise and cross-sensor image misalignment.
This repository solves the problem of driver identification through vehicular and biometric data. Through an embedding-based approach and a novel loss function, we're able to distinguish between different drivers' behaviors. This also provides preprocessing for reproducibility of results.The code preprocesses vehicular data, trains neural networks, and outputs predictions.This code introduces a novel embedding-based neural network with a 91% rank-1 accuracy, as well as all code to reproduce training and results.
This datastet contains kernels extracted from OpenMP microbenchmarks collected in DataRaceBench. The purpose of this dataset is to serve as training data for machine learning with programming language processing (PLP).
Full-waveform inversion (FWI) is an accurate imaging approach for modeling the velocity structure by minimizing the misfit between recorded and predicted seismic waveforms. However, the strong nonlinearity of FWI resulting from fitting oscillatory waveforms can trap the optimization in local minima. We have adopted a neural-network-based full-waveform inversion (NNFWI) method that integrates deep neural networks with FWI by representing the velocity model with a generative neural network. Neural networks can naturally introduce spatial correlations as regularization to the generated velocity model, which suppresses noise in the gradients and mitigates local minima. Furthermore, the velocity model generated by neural networks is input to the same partial differential equation (PDE) solvers used in conventional FWI. The gradients of the neural networks and PDEs are calculated using automatic differentiation, which back propagates gradients through the acoustic PDEs and neural network layers to update the weights of the generative neural network. Experiments on 1D velocity models, the Marmousi model, and the 2004 BP model determine that NNFWI can mitigate local minima, especially for imaging high-contrast features such as salt bodies, and it significantly improves the inversion in the presence of noise. Adding dropout layers to the neural network model also allows analyzing the uncertainty of the inversion results through Monte Carlo dropout. NNFWI opens a new pathway to combine deep learning and FWI for exploiting the characteristics of deep neural networks and the high accuracy of PDE solvers. Because NNFWI does not require extra training data and optimization loops, it provides an attractive and straightforward alternative to conventional FWI.
Structural health monitoring (SHM) systems are used to inform operation of structural systems subject to loads and environments that may affect their integrity. SHM systems rely on continuous monitoring of the structure to determine its health state. These systems are often coupled with a model of the deployed structure to determine the consequences of changes in the system by forecasting the response to future states. These models, which may be thought of as digital twins, need to be updated to reflect the latest state of the structural system. This work makes use of an uncertainty-aware machine learning model that enforces distance preservation of the original input space to determine deviations from the training data input space distributions. This workflow enables domain shift detection to determine whether damage is present in the structure. The uncertainty metrics generated by this network are then used in a Bayes risk framework to design an optimal damage detector given cost and risk considerations. The approach is demonstrated on a computational example with simulated damage.
We propose a new modeling approach for scatter estimation and descattering in polyenergetic X-ray computed tomography (CT) based on fitting models to local neighborhoods of a training set. X-ray CT is widely used in medical and industrial applications. X-ray scatter, if not accounted for during reconstruction, creates a loss of contrast in CT reconstructions and introduces severe artifacts including cupping, shading, and streaks. Even when these qualitative artifacts are not apparent, scatter can pose a major obstacle in obtaining quantitatively accurate reconstructions. Our approach to estimating scatter is, first, to generate a training set of 2D radiographs with and without scatter using particle transport simulation software. To estimate scatter for a new radiograph, we adaptively fit a scatter model to a small subset of the training data containing the radiographs most similar to it. We compared local and global (fit on full data sets) versions of several X-ray scatter models, including two from the recent literature, as well as a recent deep learning-based scatter model, in the context of descattering and quantitative density reconstruction of simulated, spherically symmetrical, single-material objects comprising shells of various densities. Our results show that, when applied locally, even simple models provide state-of-the-art descattering, reducing the error in density reconstruction due to scatter by more than half.
Fluorescence microscopy imaging speed is fundamentally limited by the measurement signal-to-noise ratio (SNR). To improve image SNR for a given image acquisition rate, computational denoising techniques can be used to suppress noise. However, common techniques to estimate a denoised image from a single frame either are computationally expensive or rely on simple noise statistical models. These models assume Poisson or Gaussian noise statistics, which are not appropriate for many fluorescence microscopy applications that contain quantum shot noise and electronic Johnson–Nyquist noise, therefore a mixture of Poisson and Gaussian noise. In this paper, we show convolutional neural networks (CNNs) trained on mixed Poisson and Gaussian noise images to overcome the limitations of existing image denoising methods. The trained CNN is presented as an open-source ImageJ plugin that performs real-time image denoising (within tens of milliseconds) with superior performance (SNR improvement) compared to conventional fluorescence microscopy denoising methods. The method is validated on external datasets with out-of-distribution noise, contrast, structure, and imaging modalities from the training data and consistently achieves high-performance ( > <#comment/> 8 d B ) denoising in less time than other fluorescence microscopy denoising methods.
Unsupervised deep learning techniques are widely used to identify anomalous behaviour. The performance of such methods is a product of the amount of training data and the model size. However, the size is often a limiting factor for the deployment on resource-constrained devices. Here, we present a novel procedure based on knowledge distillation for compressing an unsupervised anomaly detection model into a supervised deployable one and we suggest a set of techniques to improve the detection sensitivity. Compressed models perform comparably to their larger counterparts while significantly reducing the size and memory footprint.
Convolutional neural network (CNN), a deep learning algorithm, has gained popularity in technological applications that rely on interpreting images (typically, an image is a 2D field of pixels). Transport phenomena is the science of studying different fields representing mass, momentum, or heat transfer. Some of the common fields are species concentration, fluid velocity, pressure, and temperature. Each of these fields can be expressed as an image(s). Consequently, CNNs can be leveraged to solve specific scientific problems in transport phenomena. Herein, we show that such problems can be grouped into three basic categories: (a) mapping a field to a descriptor (b) mapping a field to another field, and (c) mapping a descriptor to a field. After reviewing the representative transport phenomena literature for each of these categories, we illustrate the necessary steps for constructing appropriate CNN solutions using sessile liquid drops as an exemplar problem. If sufficient training data is available, CNNs can considerably speed up the solution of the corresponding problems. Finally, the present discussion is meant to be minimalistic such that readers can easily identify the transport phenomena problems where CNNs can be useful as well as construct and/or assess such solutions.
In this study, we develop Gaussian process regression (GPR) models of isotropic hyperelastic material behavior. First, we consider the direct approach of modeling the components of the Cauchy stress tensor as a function of the components of the Finger stretch tensor in a Gaussian process. We then consider an improvement on this approach that embeds rotational invariance of the stress-stretch constitutive relation in the GPR representation. This approach requires fewer training examples and achieves higher accuracy while maintaining invariance to rotations exactly. Finally, we consider an approach that recovers the strain-energy density function and derives the stress tensor from this potential. Although the error of this model for predicting the stress tensor is higher, the strain-energy density is recovered with high accuracy from limited training data. The approaches presented here are examples of physics-informed machine learning. They go beyond purely data-driven approaches by embedding the physical system constraints directly into the Gaussian process representation of materials models.
Incorporating detailed chemical kinetic models is critical for accurate simulations of reacting flows. However, detailed models involve a large number of thermochemical (TC) state variables. Solving the governing equations to evolve these TC variables becomes impractical for real-world applications. In this work, we propose an autoencoder (AE) neural network (NN)-based reduced model to accelerate such simulations. The AE NN is first trained to find a low-dimensional latent representation of the TC states. Then, the evolving state of a chemical system can be tracked by solving the equations of the latent variables instead of the original TC equations. We demonstrate the reduced model in a syngas CO/H 2 combustion system, using training data collected from canonical perfectly stirred reactors (PSRs). It is found that the AE model can reduce the dimension of the combustion system from 12 to 2 while maintaining low reconstruction error and excellent elemental mass conservation for the test dataset. In the a posteriori test, the combustion states obtained from solving the two latent equations are compared to those from solving the 12 equations of the full model. The AE reduced method is found to be able to capture the diverse combustion states on the top two branches of the S-curve well including the extinction turning point, but with higher prediction errors for states near the ignition turning point.
Abstract We propose a deep spatiotemporal recurrent neural network, DeepShake, to project future shaking intensity directly from current ground-motion observations. DeepShake is a network-based forecasting model, able to predict future shaking intensity at all stations within a network given previously measured ground shaking. The model is not given any a priori knowledge of station locations; instead, it learns wave propagation amplitudes and delays solely from training data. We developed DeepShake with the 35,679 earthquakes from the 2019 Ridgecrest sequence. Tasked with alerting for modified Mercalli intensity (MMI) IV+ shaking on 3568 validation earthquakes at least 5 s in advance, DeepShake achieves an equal error rate of 11.4%. For the Mw 7.1 earthquake that hit Ridgecrest on 5 July 2019, DeepShake was able to provide targeted alerts to all stations inside the network 5 s prior to the arrival of MMI IV+ waveforms. DeepShake demonstrates that deep spatiotemporal neural networks can effectively provide one-step earthquake early warning with reasonable accuracy and latency.
Low emittance electron beams are of high importance at facilities like the LCLS-II at SLAC. Emittance dilution effects due to off-axis beam transport for a TESLA-type cryomodule (CM) have been shown at the Fermilab Accelerator Science and Technology facility. The results showed the correlation between the electron beam-induced cavity high-order modes (HOMs) and submacropulse centroid slewing and oscillation downstream of the CM. Mitigation of emittance dilution can be achieved by reducing the HOM signals and the variances in the submacropulse beam positions downstream of the CM. Here we present a Machine Learning based optimization and model construction for HOM signal level reduction using Neural Networks and Gaussian Processes. To gather training data we performed experiments using single bunch and 50 bunch electron beams with charges up to 125 pC/b. We measured HOM signals of all cavities and beam position with a set of BPMs downstream of the CM. The beam trajectory was changed using V/H125 corrector set located upstream of the CM. The results presented here will inform the LCLS-II injector commissioning and will serve as a prototype for HOM reduction and emittance preservation.
Given the growing emphasis on system resilience, it is important to develop software-level error detectors that help trap hardware-level faults with reasonable accuracy while minimizing false alarms as well as the performance overhead introduced. We present a technique that approaches this idea by taking stencil computations as our target, and synthesizing detectors based on machine learning. In particular, we employ linear regression to generate computationally inexpensive models which form the basis for error detection. Our technique has been incorporated into a new open-source library called SORREL. In addition to reporting encouraging experimental results, we demonstrate techniques that help reduce the size of training data. We also discuss the efficacy of various detectors synthesized, as well as our future plans.
This project was original proposed as a fundamental advance in the treatment of probabilistic modeling as it applies to nuclear fission calculations of nuclear database inputs and produced particle distributions in heavy-ion collisions. However, funding decisions made within LLNL and the DOE Office of Science led to the project proceeding through three distinct stages. The first two, probabilistic calculations of the fission and jet energy loss relate follow directly from the proposed research. The fission task led to the development of a proof-of-principle Gaussian Process Emulator for time-dependent generator coordinate method (TDGCM) calculation for the fission of Pu-240. To study energy loss in heavy-ion collisions with a probabilistic model, we generated a training data set using the Pythia event generator superimposed on a heavy-ion background with a single quench-factor parameter. We then used kernel density estimation (KDE) to recover the distribution of quench-factors for a calibration data set. In the third stage, we proceeded to work through collaborations, collecting new data within the ATLAS experiment at the Large Hadron Collider (LHC), and further developing bayesian methodology within the Jet Energy-loss Tomography with a Statistically and Computationally Advanced Program Envelope (JETSCAPE) Collaboration.