Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Machine Learning Algorithms”

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 1,675 records · Page 93
All sourcesNASA NTRSDOE OSTI
DOE OSTI2016

SU-D-206-02: Evaluation of Partial Storage of the System Matrix for Cone Beam Computed Tomography Using a GPU Platform

Purpose: Iterative reconstruction algorithms in computed tomography (CT) require a fast method for computing the intersections between the photons’ trajectories and the object, also called ray-tracing or system matrix computation. This work evaluates different ways to store the system matrix, aiming to reconstruct dense image grids in reasonable time. Methods: We propose an optimized implementation of the Siddon’s algorithm using graphics processing units (GPUs) with a novel data storage scheme. The algorithm computes a part of the system matrix on demand, typically, for one projection angle. The proposed method was enhanced with accelerating options: storage of larger subsets of the system matrix, systematic reuse of data via geometric symmetries, an arithmetic-rich parallel code and code configuration via machine learning. It was tested on geometries mimicking a cone beam CT acquisition of a human head. To realistically assess the execution time, the ray-tracing routines were integrated into a regularized Poisson-based reconstruction algorithm. The proposed scheme was also compared to a different approach, where the system matrix is fully pre-computed and loaded at reconstruction time. Results: Fast ray-tracing of realistic acquisition geometries, which often lack spatial symmetry properties, was enabled via the proposed method. Ray-tracing interleaved with projection and backprojection operations required significant additional time. In most cases, ray-tracing was shown to use about 66 % of the total reconstruction time. In absolute terms, tracing times varied from 3.6 s to 7.5 min, depending on the problem size. The presence of geometrical symmetries allowed for non-negligible ray-tracing and reconstruction time reduction. Arithmetic-rich parallel code and machine learning permitted a modest reconstruction time reduction, in the order of 1 %. Conclusion: Partial system matrix storage permitted the reconstruction of higher 3D image grid sizes and larger projection datasets at the cost of additional time, when compared to the fully pre-computed approach. This work was supported in part by the Fonds de recherche du Quebec - Nature et technologies (FRQ-NT). The authors acknowledge partial support by the CREATE Medical Physics Research Training Network grant of the Natural Sciences and Engineering Research Council of Canada (Grant No. 432290).

60 APPLIED LIFE SCIENCES↗
DOE OSTI2025

Best of both worlds: Enforcing detailed balance in machine learning models of transition rates

The slow microstructural evolution of materials often plays a key role in determining material properties. When the unit steps of the evolution process are slow, direct simulation approaches such as molecular dynamics become prohibitive and Kinetic Monte-Carlo (kMC) algorithms, where the state-to-state evolution of the system is represented in terms of a continuous-time Markov chain, are instead frequently relied upon to efficiently predict long-time evolution. The accuracy of kMC simulations however relies on the complete and accurate knowledge of reaction pathways and corresponding kinetics. This requirement becomes extremely stringent in complex systems such as concentrated alloys where the astronomical number of local atomic configurations makes the a priori tabulation of all possible transitions impractical. Machine learning models of transition kinetics have been used to mitigate this problem by enabling the efficient on-the-fly prediction of kinetic parameters. While conventional KMC methods based on transition state theory naturally yield reversible dynamics that exactly obey the detailed balance criterion, providing strong guarantees on the properties of the stationary distribution, many recently-proposed ML-based approaches to barrier predictions provide no such guarantees. In this study, we derive conditions under which physics-informed ML architectures exactly enforce the detailed balance condition by construction, even when relying on non-extensive descriptions of states in terms of local environments around mobile defects. In conclusion, using the diffusion of a vacancy in a concentrated alloy as an example, we show that such ML architectures also exhibit superior performance in terms of prediction accuracy, demonstrating that the imposition of physical constraints can facilitate the accurate learning of barriers at no increase in computational cost.

36 MATERIALS SCIENCE↗
DOE OSTI2019

Searching for Light Sterile Neutrinos with NOvA Through Neutral-Current Disappearance

Most appearance and disappearance neutrino oscillation experiments have observed evidence for neutrino oscillations that agree with the three-flavor neutrino paradigm. Explaining unexpected signal from short-baseline appearance experiments, such as LSND and MiniBooNE, in terms of neutrino oscillations requires the existence of light sterile neutrino(s). The study of light sterile neutrino induced flavor mixing in NOvA uses a long-baseline of 810 km between the Near Detector (ND) at Fermilab and the Far Detector (FD) in Minnesota. The signal for light sterile neutrino-driven oscillations is a deficit of Neutral-Current (NC) neutrino interactions at the FD concerning the ND prediction. In our analysis, we are employing current innovations in machine learning technologies for NC sterile neutrino searches with NOvA. I will detail the state-of-art algorithm which we employed and the improved results obtained in the NOvA NC Disappearance analysis.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2018

Opportunities in Machine Learning for Particle Accelerators

Machine learning (ML) is a subfield of artificial intelligence. The term applies broadly to a collection of computational algorithms and techniques that train systems from raw data rather than a priori models. ML techniques are now technologically mature enough to be applied to particle accelerators, and we expect that ML will become an increasingly valuable tool to meet new demands for beam energy, brightness, and stability. The intent of this white paper is to provide a high-level introduction to problems in accelerator science and operation where incorporating ML-based approaches may provide significant benefit. We review ML techniques currently being investigated at particle accelerator facilities, and we place specific emphasis on active research efforts and promising exploratory results. We also identify new applications and discuss their feasibility, along with the required data and infrastructure strategies. We conclude with a set of guidelines and recommendations for laboratory managers and administrators, emphasizing the logistical and technological requirements for successfully adopting this technology. This white paper also serves as a summary of the discussion from a recent workshop held at SLAC on ML for particle accelerators.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

Machine learning for fundamental spectroscopic and thermodynamic data of actinides and lanthanides

Accurately modeling optical spectra with absolute radiometric intensities is vital for nuclear forensics applications that depend on characterizing optical emissions from energetic nuclear phenomena. This requires precise knowledge of the individual atomic transition probabilities, known as Einstein A-coefficients, for each emission line. Obtaining these values theoretically or experimentally is often impractical due to the complex electronic structures and the number of transitions involved in atoms relevant to nuclear applications. In this study, we explore the use of machine learning to predict the Einstein A coefficients for atomic transitions. Seven models were evaluated that ranged from deep learning to decision tree algorithms, and found that gradient boosting performed best, specifically the Extreme Gradient Boosting (XGB) architecture, achieving a precision of 86% across transitions of 36 elements. Furthermore, the model was cross-validated using published transition probabilities reported in the literature and applied to estimate Pu plasma temperatures from a previous experiment conducted at Savannah River National Laboratory.

Atomic spectroscopy↗
DOE OSTI2016

sNebula, a network-based algorithm to predict binding between human leukocyte antigens and peptides

Understanding the binding between human leukocyte antigens (HLAs) and peptides is important to understand the functioning of the immune system. Since it is time-consuming and costly to measure the binding between large numbers of HLAs and peptides, computational methods including machine learning models and network approaches have been developed to predict HLA-peptide binding. However, there are several limitations for the existing methods. We developed a network-based algorithm called sNebula to address these limitations. We curated qualitative Class I HLA-peptide binding data and demonstrated the prediction performance of sNebula on this dataset using leave-one-out cross-validation and five-fold cross-validations. Furthermore, this algorithm can predict not only peptides of different lengths and different types of HLAs, but also the peptides or HLAs that have no existing binding data. We believe sNebula is an effective method to predict HLA-peptide binding and thus improve our understanding of the immune system.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2016

Fullrmc, a rigid body reverse monte carlo modeling package enabled with machine learning and artificial intelligence

Here, a new Reverse Monte Carlo (RMC) package ‘fullrmc’ for atomic or rigid body and molecular, amorphous or crystalline materials is presented. fullrmc main purpose is to provide a fully modular, fast and flexible software, thoroughly documented, complex molecules enabled, written in a modern programming language (python, cython ,C and C++ when performance is needed) and complying to modern programming practices. fullrmc approach in solving an atomic or molecular structure is different from existing RMC algorithms and software. In a nutshell, traditional RMC methods and software randomly adjust atom positions until the whole system has the greatest consistency with a set of experimental data. In contrast, fullrmc applies smart moves endorsed with reinforcement machine learning to groups of atoms. While fullrmc allows running traditional RMC modelling, the uniqueness of this approach resides in its ability to customize grouping atoms in any convenient way with no additional programming efforts and to apply smart and more physically meaningful moves to the defined groups of atoms. Also fullrmc provides a unique way with almost no additional computational cost to recur a group’s selection, allowing the system to go out of local minimas by refining a group’s position or exploring through and beyond not allowed positions and energy barriers the unrestricted three dimensional space around a group.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Machine Learning Enabled Sensor Fusion for In-Situ Defect Detection in Laser Powder Bed Fusion

Laser Powder Bed Fusion (L-PBF) Additive Manufacturing (AM) is among the metal 3D printing technologies most broadly adopted by the manufacturing industry. The current industry qualification paradigm for critical-application L-PBF parts relies heavily on expensive non-destructive inspection techniques such as X-Ray Computed Tomography (XCT), which significantly limits the use-cases of L-PBF. In situ monitoring of the process promises a less expensive alternative to ex situ testing, but existing sensor technologies and data analysis techniques struggle to detect sub-surface flaws (e.g., porosity and cracking) on production-scale L-PBF printers. RTX Technologies Research Center (RTRC) has licensed ORNL’s Peregrine software package – a printer- and camera-agnostic data analytics tool designed specifically for detecting process anomalies using in situ data collected during powder bed printing. The goal of this project was to feed temporally rich, multi-modal sensor data, including visible light, integrated near infrared (NIR), and spatially mapped co-axial melt pool thermal emission data into Peregrine to enable detection of subsurface flaws. XCT data was used as ground truth training data to allow Peregrine’s deep learning algorithms to recognize anomalies in these complex data streams in both test artifacts and industrially relevant geometries. Completion of this program has seen the successful implementation of multi-modal, multi-layer sensor data footprints for training of machine learning models in Peregrine. Flaws detected in XCT data have been successfully detected directly from this in situ data footprint, and initial analyses of the in situ probability-of-detection has been conducted, showing performance levels commensurate with traditional non-destructive evaluation (NDE) methods. The in situ monitoring methodology was then applied to an industrially relevant component that was using post-build NDE, highlighting the utility of the proposed method for hard-to-inspect AM components. As a direct result of this program, two journal manuscripts [1], [2] have been published in Additive Manufacturing, with additional manuscripts planned following program completion.

36 MATERIALS SCIENCE↗
DOE OSTI2018

A learning-based data-driven forecast approach for predicting future reservoir performance

Quantification of the predictive uncertainty of subsurface models has long been investigated. The traditional workflow is to calibrate prior models to match observed data, and then use the posterior models to simulate future system performance. Not only are these procedures computationally expensive, but they also have issues in maintaining geological model constraints during the calibration step. Data space inversion (DSI) was introduced recently to predict future system performance without the iterative history matching or model calibration step. In general, DSI approaches seek to establish a statistical relationship between the observed and forecast variables, as well as to quantify the predictive uncertainty of the forecast variables, by using an ensemble of uncalibrated prior models. Existing DSI approaches all require a number of complex transformation and mapping operations, which may deter their widespread use. Here in this study, we introduce a new and simpler DSI approach, the learning-based, data-driven forecast approach (LDFA), by combining dimension reduction and machine learning techniques to quickly provide accurate forecast results and reliably quantify corresponding uncertainty in the results. Our LDFA framework is demonstrated using two supervised learning algorithms, artificial neural network (ANN) and support vector regression (SVR), on two representative examples from reservoir engineering and geological carbon storage. Results suggest that our approach provides accurate forecast results (e.g., future oil production rate or cumulative injected CO 2 ) and reasonable predictive uncertainty intervals. Our framework is generic and may be applied to other surface and subsurface problems.

42 ENGINEERING↗
DOE OSTI2019

“Teamwork Makes the Dream Work”: Tribal Competition Evolutionary Search as a Surrogate for Free-Energy-Based Structural Predictions

Crystal structure prediction has been a grand challenge in material science owing to the large configurational space that one must explore. Evolutionary (genetic) algorithms coupled with first principles calculations are commonly used in crystal structure prediction to sample the ground and metastable states of materials based on configurational energies. However, crystal structure predictions at finite temperature (T), pressure (P), and composition (X) require a free-energy-based search that is often computationally expensive and tedious. Here in this paper, we introduce a new machine-learning workflow for structure prediction that is based on a concept inspired by the evolution of human tribes in primitive society. Our tribal genetic algorithm (GA) combines configurational sampling with evolutionary optimization to accurately predict entropically stabilized phases at finite (T, P, X), at a computational cost that is an order of magnitude smaller than that required for a free-energy-based search. In a departure from standard GA techniques, the populations of individuals are divided into multiple tribes based on a bond-order fingerprint, and genetic operations are modified to ensure that cluster configurations are sampled adequately to capture entropic contributions. Team competition introduced into the evolutionary process allows winning teams (representing a better set of individuals) to expand their sizes; this translates into a more expanded search of the phase space allowing us to explore solutions near possible global minimum. Each team explores a specific section of the structural phase space and avoids bias on solutions arising from the use of individual populations in a purely energy-based search. We demonstrate the efficacy of our approach by performing the structural prediction of a representative two-dimensional two-body system as well as Lennard-Jones clusters over a range of temperatures up to its melting point. Our approach outperforms the standard GA approaches and enables structural search under “real nonambient conditions” on both bulk systems and finite-sized clusters.

36 MATERIALS SCIENCE↗
DOE OSTI2026

Speedup of UEDGE Parameter Scans Using Machine-Learning Optimized OpenMP Parallelization and a Continuation Solver

This article presents the OpenMP parallelization of the preconditioning Jacobian assembly and right‐hand side residual evaluation in UEDGE. A continuation algorithm, utilizing the internal NKSOL implicit Jacobian‐Free Newton‐Krylov solver to efficiently scan physical parameters, is also presented. The implemented parallelization reduces the computational time for a benchmark scan run on 32 threads by compared to the serial version when using trained random forest regression models to identify the optimal decomposition of the system of equations. Random forest regression models applied to the UEDGE time‐dependent and continuation solver algorithms did not yield meaningful improvement in computational performance. A benchmark DIII‐D gas injection rate scan in the 0.35–0.75 kA interval, performed on a test cluster using the parallelized code and continuation solver, produced 1066 steady‐state solutions with a 22 s average wall‐clock computational time per steady‐state solution.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2019

Benchmarking Parametric and Machine Learning Models for Genomic Prediction of Complex Traits

The usefulness of genomic prediction in crop and livestock breeding programs has prompted efforts to develop new and improved genomic prediction algorithms, such as artificial neural networks and gradient tree boosting. However, the performance of these algorithms has not been compared in a systematic manner using a wide range of datasets and models. Using data of 18 traits across six plant species with different marker densities and training population sizes, we compared the performance of six linear and six non-linear algorithms. First, we found that hyperparameter selection was necessary for all non-linear algorithms and that feature selection prior to model training was critical for artificial neural networks when the markers greatly outnumbered the number of training lines. Across all species and trait combinations, no one algorithm performed best, however predictions based on a combination of results from multiple algorithms (i.e., ensemble predictions) performed consistently well. While linear and non-linear algorithms performed best for a similar number of traits, the performance of non-linear algorithms vary more between traits. Although artificial neural networks did not perform best for any trait, we identified strategies (i.e., feature selection, seeded starting weights) that boosted their performance to near the level of other algorithms. Our results highlight the importance of algorithm selection for the prediction of trait values.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2019

Benchmarking parametric and machine learning models for genomic prediction of complex traits

The usefulness of genomic prediction in crop and livestock breeding programs has prompted efforts to develop new and improved genomic prediction algorithms, such as artificial neural networks and gradient tree boosting. However, the performance of these algorithms has not been compared in a systematic manner using a wide range of datasets and models. Using data of 18 traits across six plant species with different marker densities and training population sizes, we compared the performance of six linear and six non-linear algorithms. First, we found that hyperparameter selection was necessary for all non-linear algorithms and that feature selection prior to model training was critical for artificial neural networks when the markers greatly outnumbered the number of training lines. Across all species and trait combinations, no one algorithm performed best, however predictions based on a combination of results from multiple algorithms (i.e. ensemble predictions) performed consistently well. While linear and non-linear algorithms performed best for a similar number of traits, the performance of non-linear algorithms vary more between traits. Although artificial neural networks did not perform best for any trait, we identified strategies (i.e. feature selection, seeded starting weights) that boosted their performance to near the level of other algorithms. Our results highlight the importance of algorithm selection for the prediction of trait values.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

A semi-supervised machine learning detector for physics events in tokamak discharges

Databases of physics events have been used in various fusion research applications, including the development of scaling laws and disruption avoidance algorithms, yet they can be time-consuming and tedious to construct. This paper presents a novel application of the label spreading semi-supervised learning algorithm to accelerate this process by detecting distinct events in a large dataset of discharges, given few manually labeled examples. A high detection accuracy (>85%) for H-L back transitions and initially rotating locked modes is demonstrated on a dataset of hundreds of discharges from DIII-D with manually identified events for which only 3 discharges are initially labeled by the user. Lower yet reasonable performance (~75%) is also demonstrated for the core radiative collapse, an event with a much lower prevalence in the dataset. Additionally, analysis of the performance sensitivity indicates that the same set of algorithmic parameters is optimal for each event. This suggests that the method can be applied to detect a variety of other events not included in this paper, given that the event is well described by a set of 0D signals robustly available on many discharges. Procedures for analysis of new events are demonstrated, showing automatic event detection with increasing fidelity as the user strategically adds manually labeled examples. Detections on Alcator C-Mod and EAST are also shown, demonstrating the potential for this to be used on a multi-tokamak dataset.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2021

Versatile feature learning with graph convolutions and graph structures

Graphs represent real world relationships, and graph embedding projects nodes in a graph to a latent space that can help simplify downstream tasks. Recent development of graph convolutions in deep learning significantly improves the performance of many learning tasks on graphs. Unfortunately, prior embedding methods either do not embed graphs with node features, or fail to produce high-quality embeddings for downstream learning tasks that result in large performance gap in comparison to direct learning on graphs.We present a versatile and effective embedding method, Conv2Vec, for embedding graphs with or without node features. It is based on graph convolutions with objective functions motivated by concepts and structures from classical graph algorithms. Conv2Vec produce high-quality embedding for both plain graphs and graphs with node features for downstream tasks.We evaluate the embeddings generated by Conv2Vec with a transductive node classification task. With the generated embeddings and very simple machine learning approaches, we are able to achieve accuracies similar to those achieved by direct learning with graph convolutions. Interestingly, if we strip the node features from the graph and thus learning an embedding has to rely entirely on the graph topology, node classification with our embedding significantly outperforms direct learning with various graph convolutions. This suggests that structures from classical graph algorithms may play an important role in learning on graphs.

Cong, Guojing↗
DOE OSTI2020

Accelerating Scientific Computing in the Post-Moore’s Era

Novel uses of graphical processing units for accelerated computation revolutionized the field of high-performance scientific computing by providing specialized workflows tailored to algorithmic requirements. As the era of Moore’s law draws to a close, many new non–von Neumann processors are emerging as potential computational accelerators, including those based on the principles of neuromorphic computing, tensor algebra, and quantum information. While development of these new processors is continuing to mature, the potential impact on accelerated computing is anticipated to be profound. We discuss how different processing models can advance computing in key scientific paradigms: machine learning and constraint satisfaction. Significantly, each of these new processor types utilizes a fundamentally different model of computation, and this raises questions about how to best use such processors in the design and implementation of applications. While many processors are being developed with a specific domain target, the ubiquity of spin-glass models and neural networks provides an avenue for multi-functional applications. Furthermore, this also hints at the infrastructure needed to integrate next-generation processing units into future high-performance computing systems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Accelerating phase-field-based microstructure evolution predictions via surrogate models trained by machine learning methods

Abstract The phase-field method is a powerful and versatile computational approach for modeling the evolution of microstructures and associated properties for a wide variety of physical, chemical, and biological systems. However, existing high-fidelity phase-field models are inherently computationally expensive, requiring high-performance computing resources and sophisticated numerical integration schemes to achieve a useful degree of accuracy. In this paper, we present a computationally inexpensive, accurate, data-driven surrogate model that directly learns the microstructural evolution of targeted systems by combining phase-field and history-dependent machine-learning techniques. We integrate a statistically representative, low-dimensional description of the microstructure, obtained directly from phase-field simulations, with either a time-series multivariate adaptive regression splines autoregressive algorithm or a long short-term memory neural network. The neural-network-trained surrogate model shows the best performance and accurately predicts the nonlinear microstructure evolution of a two-phase mixture during spinodal decomposition in seconds, without the need for “on-the-fly” solutions of the phase-field equations of motion. We also show that the predictions from our machine-learned surrogate model can be fed directly as an input into a classical high-fidelity phase-field model in order to accelerate the high-fidelity phase-field simulations by leaping in time. Such machine-learned phase-field framework opens a promising path forward to use accelerated phase-field simulations for discovering, understanding, and predicting processing–microstructure–performance relationships.

36 MATERIALS SCIENCE↗
DOE OSTI2024

Unsupervised Learning for Improved Gamma-Ray Spectrometry in Pixelated Cadmium Zinc Telluride (CZT) Detectors

Machine learning has been found to be ubiquitously useful across many industries, presenting an opportunity to improve radiation detection performance using data-driven algorithms. Improved detector resolution can aid in the detection, identification, and quantification of radionuclides. Here, in this work, a novel, data-driven, unsupervised learning approach is developed to improve detector spectral characteristics by learning, and subsequently rejecting, poorly performing regions of the pixelated detector. Feature engineering is used to fit individual characteristic photo peaks to a Doniach lineshape with a linear background model. Then, principal component analysis is used to learn a lower-dimension latent space representation of each photo peak where the pixels are clustered, and subsequently ranked, based on the cluster mean distance to an optimal point. Pixels within the worst cluster(s) are rejected to improve the full-width at half-maximum (FWHM) by 10% to 15% (relative to the bulk detector) at 50% net efficiency when applied to training data obtained from measurements of a 100 μCi 154 Eu source using a H3D M400i pixelated cadmium zinc telluride detector. These results compare well with, but do not outperform, a greedy algorithm that accumulates pixels in order of FWHM from lowest to highest used as a benchmark. In the future, this approach can be extended to include the detector energy and angular response. Finally, the model is applied to newly seen natural and enriched uranium spectra relevant for nuclear safeguards applications.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗