Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “generative machine learning”

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 397 records · Page 22

Data-driven material models for atomistic simulation

The central approximation made in classical molecular dynamics simulation of materials is the interatomic potential used to calculate the forces on the atoms. Great effort and ingenuity is required to construct viable functional forms and find accurate parametrizations for potentials using traditional approaches. Machine learning has emerged as an effective alternative approach to develop accurate and robust interatomic potentials. Beginning with a very general model form, the potential is learned directly from a database of electronic structure calculations and therefore can be viewed as a multiscale link between quantum and classical atomistic simulations. Risk of inaccurate extrapolation exists outside the narrow range of time and length scales where the two methods can be directly compared. Here, we use the spectral neighbor analysis potential (SNAP) and show how a fit can be produced with minimal interpolation errors which is also robust in extrapolating beyond training. To demonstrate the approach, we have developed a tungsten-beryllium potential suitable for the full range of binary compositions. Subsequently, large-scale molecular dynamics simulations were performed of high energy Be atom implantation onto the (001) surface of solid tungsten. The machine learned W-Be potential generates a population of implantation structures consistent with quantum calculations of defect formation energies. A very shallow ( < 2 nm ) average Be implantation depth is predicted which may explain ITER diverter degradation in the presence of beryllium.

36 MATERIALS SCIENCE↗

Detecting Important Drivers of Gridded Population Modeling With Machine Learning

High-resolution population datasets have been lever-aged across a broad swath of domains, such as climate change, public policy, humanitarian aid, and rescue operations, among others. Machine learning methods were adopted to generate high-resolution or gridded population estimates by using various geospatial input features such as buildings, roads, and nighttime lights. In this study, we evaluate the importance of population features using Random Forest models across three levels of analysis, utilizing permutation measures. Our research aims to address key questions to enhance our understanding of high-resolution population modeling, such as: Are certain features globally (10 countries collectively) more important than others? Do optimal features vary by country? Within each country, do feature importance differ across administrative units? What similarities exist in feature importance at the global, country, and administrative unit levels? To answer these questions, we leverage the Kneedle algorithm to automate the selection of optimum features. We find that there are patterns displayed by features across spatial boundaries, evidenced by the same feature being the most important indicator of population across 7 of the 10 countries modeled. Our findings indicate that while important features may vary across geographies, certain features consistently hold greater importance than others agnostic of geography.

Lebakula, Viswadeep [ORNL] (ORCID:0000000152935914↗

MICCO: An Enhanced Multi-GPU Scheduling Framework for Many-Body Correlation Functions

Calculation of many-body correlation functions is one of the critical kernels utilized in many scientific computing areas, especially in Lattice Quantum Chromodynamics (Lattice QCD). It is formalized as a sum of a large number of contraction terms each of which can be represented by a graph consisting of vertices describing quarks inside a hadron node and edges designating quark propagations at specific time intervals. Due to its computation- and memory-intensive nature, real-world physics systems (e.g., multi-meson or multi-baryon systems) explored by Lattice QCD prefer to leverage multi-GPUs. Different from general graph processing, many-body correlation function calculations show two specific features: a large number of computation-/data-intensive kernels and frequently repeated appearances of original and intermediate data. The former results in expensive memory operations such as tensor movements and evictions. The latter offers data reuse opportunities to mitigate the data-intensive nature of many-body correlation function calculations. However, existing graph-based multi-GPU schedulers cannot capture these data-centric features, thus resulting in a sub-optimal performance for many-body correlation function calculations. To address this issue, this paper presents a multi-GPU scheduling framework, MICCO, to accelerate contractions for correlation functions particularly by taking the data dimension (e.g., data reuse and data eviction) into account. This work first performs a comprehensive study on the interplay of data reuse and load balance, and designs two new concepts: local reuse pattern and reuse bound to study the opportunity of achieving the optimal trade-off between them. Based on this study, MICCO proposes a heuristic scheduling algorithm and a machine-learning-based regression model to generate the optimal setting of reuse bounds. Specifically, MICCO is integrated into a real-world Lattice QCD system, Redstar, for the first time running on multiple GPUs. The evaluation demonstrates MICCO outperforms other state-of-art works, achieving up to 2.25× speedup in synthesized datasets, and 1.49× speedup in real-world correlation functions.

Wang, Qihan↗

High-precision inversion of dynamic radiography using hydrodynamic features

While radiography is routinely used to probe complex, evolving density fields in research areas ranging from materials science to shock physics to inertial confinement fusion and other national security applications, complications resulting from noise, scatter, complex beam dynamics, etc. prevent current methods of reconstructing density from being accurate enough to identify the underlying physics with sufficient confidence. In this work, we show that using only features that are robustly identifiable in radiographs and combining them with the underlying hydrodynamic equations of motion using a machine learning approach of a conditional generative adversarial network (cGAN) provides a new and effective approach to determine density fields from a dynamic sequence of radiographs. In particular, we demonstrate the ability of this method to outperform a traditional, direct radiograph to density reconstruction in the presence of scatter, even when relatively small amounts of scatter are present. Our experiments on synthetic data show that the approach can produce high quality, robust reconstructions. We also show that the distance (in feature space) between a testing radiograph and the training set can serve as a diagnostic of the accuracy of the reconstruction.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

NNL.Fe.qSNAP-ZBL.2024.1: A Fe Spectral Neighbor Analysis Potential for Radiation Damage Simulations

The NNL.Fe.qSNAP-ZBL.2024.1 machine-learned potential (MLP) has been generated to support the development of an elemental body-centered cubic (BCC) Fe athermal recombination corrected neutron damage model and simulations of primary recoil atom (PRA) cascades in BCC Fe. This MLP is a quadratic spectral neighbor analysis potential (qSNAP) hybridized with the universal Ziegler-Beirsack-Littmark (ZBL) potential at short-range and is named according to Naval Nuclear Laboratory MLP naming conventions (NNL.material-system.MLP-type.year.version). Training set calculations for Fe are presented along with the subsequent MLP fitting procedure. A key criterion of the fitting procedure is that ZBL describes the short-range interaction with minimal impact on the MLP. The MLP is compared to density functional theory (DFT) predicted properties relevant to radiation damage simulation, including threshold displacement energies, for validation. The NNL.Fe.qSNAP-ZBL.2024.1 potential is considered suitable for molecular dynamics (MD) simulations of radiation defects up to 800 K and PRA cascades in BCC Fe up to around 10 keV. The potential can additionally be used on a limited basis for recoils of 10–20 keV, within which range the emergence of structures outside the training set in cascade simulations may cause system instabilities.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

MSD CoP Webinar: Advancing MSD Research with Artificial Intelligence

Context: This webinar was hosted by the MultiSector Dynamics Community of Practice (MSD CoP; https://multisectordynamics.org). Abstract: Recent advances in Artificial Intelligence (AI) are quickly changing the landscape of tools available to conceptualize, execute, and disseminate research. We posit that research efforts in Multi-Sector Dynamics can benefit from these advances; the new AI in MSD Working Group thus aims to identify and quantify opportunities and risks associated with their implementation. In this webinar, we will first introduce the new AI Working Group, which was initially conceived during the first MSD workshop in October 2023. Next, our panelists will explore how generative AI, explainable AI, and machine learning can help us improve modeling efforts in multiple domains, including climate science, hydrology, and energy systems. Finally, we will discuss the aims of the working group, gather inputs from the community, and suggest directions for the next steps. Presenters : Andrea Castelletti (Politecnico di Milano; Invited Speaker), Chaopeng Shen (Pennsylvania State University; Invited Speaker), Nicole Jackson (Sandia National Laboratory; Invited Speaker), Stefano Galelli (Cornell University; Co-Chair), David Gold (Utrecht University; Co-Chair), Jillian Sturtevant (Baylor University; Communications Officer) Moderator: Pat M. Reed (MSD CoP Facilitation Team) This webinar was held on: June 14th, 2024 from 12-1:30 PM EST

Artificial Intelligence↗

Development and experimental qualification of novel disruption prevention techniques on DIII-D

Novel disruption prevention solutions spanning a range of control regimes are being developed and tested on DIII-D to enable ITER success. First, a new real-time control algorithm has been developed and tested for regulating nearness to stability limits and maintaining safety-margins. Its first application has been for reliable prevention of vertical displacement events (VDEs) by adjusting plasma elongation (κ) and the inner-gap between the plasma and inner-wall in response to real- time open-loop VDE growth rate (γ) estimators. VDEs were robustly prevented up to average open-loop growth rates of 800 rad/s with initial tunings, with only applying shape modification when near safety limits. Second, the disruption risk during fast, emergency shutdown after large tearing and locked modes can be significantly improved by transitioning to a limited topology during shutdown. More than 50% of emergency limited shutdowns after locked modes reach a final normalized current I N < 0.3 before terminating, scaling to the 3 MA ITER requirement. This is in contrast to diverted shutdowns, the majority of which disrupt at I N > 0.8. Despite improvements, these results highlight the critical importance of early prevention. Third, a novel emergency shut down method has been developed which excites instabilities to form a warm, helical core post-thermal quench. The current quench extends to ~100ms and avoids VDEs and runaway electron generation. Novel real-time machine learning disruption prediction has been integrated with the DIII-D proximity controller, and a real- time compatible multi-mode MHD spectroscopy technique has been developed. Results presented here were enabled by a focused effort, the Disruption Free Protocol, in DIII-D’s 2019-20 campaign to complement disruption prevention experiments with a large piggy-back program. In addition to testing novel techniques, it is estimated to have helped avoid 32 potential disruptions in piggyback operations with rapid, early shutdowns after large rotating n=1 or locked modes.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Modeling Weather Impact on Ground Delay Programs

Scheduled arriving aircraft demand may exceed airport arrival capacity when there is abnormal weather at an airport. In such situations, Federal Aviation Administration (FAA) institutes ground-delay programs (GDP) to delay flights before they depart from their originating airports. Efficient GDP planning depends on the accuracy of prediction of airport capacity and demand in the presence of uncertainties in weather forecast. This paper presents a study of the impact of dynamic airport surface weather on GDPs. Using the National Traffic Management Log, effect of weather conditions on the characteristics of GDP events at selected busy airports is investigated. Two machine learning methods are used to generate models that map the airport operational conditions and weather information to issued GDP parameters and results of validation tests are described.

Wang, Yao↗

Analysis and Prediction of Weather Impacted Ground Stop Operations

When the air traffic demand is expected to exceed the available airport's capacity for a short period of time, Ground Stop (GS) operations are implemented by Federal Aviation Administration (FAA) Traffic Flow Management (TFM). The GS requires departing aircraft meeting specific criteria to remain on the ground to achieve reduced demands at the constrained destination airport until the end of the GS. This paper provides a high-level overview of the statistical distributions as well as causal factors for the GSs at the major airports in the United States. The GS's character, the weather impact on GSs, GS variations with delays, and the interaction between GSs and Ground Delay Programs (GDPs) at Newark Liberty International Airport (EWR) are investigated. The machine learning methods are used to generate classification models that map the historical airport weather forecast, schedule traffic, and other airport conditions to implemented GS/GDP operations and the models are evaluated using the cross-validations. This modeling approach produced promising results as it yielded an 85% overall classification accuracy to distinguish the implemented GS days from the normal days without GS and GDP operations and a 71% accuracy to differentiate the GS and GDP implemented days from the GDP only days.

Analysis↗

Artificial Intelligence - The Future of Space Communications

Presentation focus on the Cognitive Communications Project Overview. The project aims to develop cognitive communications technologies to increase mission science return and improve resource efficiencies. Another key aspect of the project is understanding the fundamental aspects of cognitive technology and developing artificial intelligence to advance the future of space communications. This involves the use of machine learning algorithms in the next generation architecture for space communications in the effort to increase efficiency, autonomy and increased performance of the space communication and navigations next generation architecture.

Briones, Janette C.↗

A Regional Map of Mangrove Extent for Myanmar, Thailand, and Cambodia Shows Losses of 44% By 1996

Southeast Asia is home to some of the planet’s most carbon-dense and biodiverse mangrove ecosystems. There is still much uncertainty with regards to the timing and magnitude of changes in mangrove cover over the past 50 years. While there are several regional to global maps of mangrove extent in Southeast Asia over the past two decades, data prior to the mid-1990s is limited due to the scarcity of Earth Observation data of sufficient quality and the historical limitations to publicly available EO. We present a classification of mangrove extent using Landsat 1-2 MSS Tier 2 data from 1972-1977 for three Southeast Asian countries: Myanmar, Thailand, and Cambodia. Mangrove extent land cover maps were generated using a Random Forest machine learning algorithm that effectively mapped a total of 15,420.51 km2. Accuracy assessments indicated that the classification for the mangrove and non-mangrove class had a producer’s accuracy of 79.63 % and 98.24 %a user’s accuracy of 89.58 % and 96.21 %, and an overall accuracy of 95.26 % . We found a decline of 6,830 km2 between the 1970’s and 2020, showing that44.29% of the mangrove area in these countries has been lost in the past 48 years. Most of this loss occurred between the 1970’s and 1996; rates of deforestation declined dramatically after 1996. This study also urges that these results be considered within the context of the local need for these coastal resources that helped them endure the political and financial challenges of their coastal homes.

Priscilla Baltezar↗

Active learning for robust, high-complexity reactive atomistic simulations

Machine learned reactive force fields based on polynomial expansions have been shown to be highly effective for describing simulations involving reactive materials. Nevertheless, the highly flexible nature of these models can give rise to a large number of candidate parameters for complicated systems. In these cases, reliable parameterization requires a well-formed training set, which can be difficult to achieve through standard iterative fitting methods. In this paper, we present an active learning approach based on cluster analysis and inspired by Shannon information theory to enable semi-automated generation of informative training sets and robust machine learned force fields. The use of this tool is demonstrated for development of a model based on linear combinations of Chebyshev polynomials explicitly describing up to four-body interactions, for a chemically and structurally diverse system of C/O under extreme conditions. We show that this flexible training database management approach enables development of models exhibiting excellent agreement with Kohn–Sham density functional theory in terms of structure, dynamics, and speciation.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

GOOML: Geothermal Operational Optimization with Machine Learning

Geothermal Operational Optimization with Machine Learning (GOOML) is a project focused on maximizing increased availability and capacity from existing industrial-scale geothermal generation assets. The GOOML project will develop a suite of machine learning-based algorithms that analyze historical production datasets and provide predictive setpoints for geothermal field operations. Historical datasets from New Zealand and the US will provide the input to develop digital geothermal system twins which allow prediction of market conditions, maintenance operations and steamfield optimization. The algorithms will identify key parameters within fields and suggest setpoints for components of the system to maintain optimal generation. Set-points can be instructed to follow mass flow restrictions, generation maximization and optimal field/reservoir balance and give field operators a guide by which generation can be optimized. The datasets that will be used to develop GOOML are sourced from operating geothermal fields in New Zealand and the United States with varying degrees of complexity. This will ensure that most geothermal systems can utilize the GOOML tool to assist in optimizing operations. GOOML aims to achieve a step-change in geothermal operations by developing state-of-the-art machine learning algorithms, comprehensive data analytics, and a first-of-its-kind automated, intelligent geothermal system model.

algorithms↗

MaLTESE: Large-Scale Simulation-Driven Machine Learning for Transient Driving Cycles

Optimal engine operation during a transient driving cycle is the key to achieving greater fuel economy, engine efficiency, and reduced emissions. In order to achieve continuously optimal engine operation, engine calibration methods use a combination of static correlations obtained from dynamometer tests for steady-state operating points and road and/or track performance data. As the parameter space of control variables, design variable constraints, and objective functions increases, the cost and duration for optimal calibration become prohibitively large. In order to reduce the number of dynamometer tests required for calibrating modern engines, a large-scale simulation-driven machine learning approach is presented in this work. A parallel, fast, robust, physics-based reduced-order engine simulator is used to obtain performance and emission characteristics of engines over a wide range of control parameters under various transient driving conditions (drive cycles). We scale the simulation up to 3,906 nodes of the Theta supercomputer at the Argonne Leadership Computing Facility to generate data required to train a machine learning model. The trained model is then used to predict various engine parameters of interest, and the results are compared with those predicted by the engine simulator. Our results show that a deepneural-network-based surrogate model achieves high accuracy: Pearson product-moment correlation values larger than 0.99 and mean absolute percentage error within 1.07% for various engine parameters such as exhaust temperature, exhaust pressure, nitric oxide, and engine torque. Once trained, the deep-neural-network-based surrogate model is fast for inference: it requires about 16 mu s for predicting the engine performance and emissions for a single design configuration compared with about 0.5 s per configuration with the engine simulator. Moreover, we demonstrate that transfer learning and retraining can be leveraged to incrementally retrain the surrogate model to cope with new configurations that fall outside the training data space.

deep learning↗

Invited: Software defined accelerators from learning tools environment

Next generation systems, such as edge devices, will need to provide efficient processing of machine learning (ML) algorithms along several metrics, including energy, performance, area, and latency. However, the quickly evolving field of ML makes it extremely difficult to generate accelerators able to support a wide variety of algorithms. At the same time, designing accelerators in hardware description languages (HDLs) by hand is hard and time consuming, and does not allow quick exploration of the design space. In this paper we present the Software Defined Accelerators From Learning Tools Environment (SODALITE), an automated open source high-level ML framework-to-verilog compiler targeting ML Application-Specific Integrated Circuits (ASICs) chiplets. The SODALITE approach will implement optimal designs by seamlessly combining custom components generated through high-level synthesis (HLS) with templated and fully tunable Intellectural Properties (IPs) and macros, integrated in an extendable resource library. Through a closed loop design space exploration engine, developers will be able to quickly explore their hardware designs along different dimensions.

High-Level Synthesis, Accelerators, Hardware-Softw↗

QSO photometric redshifts using machine learning and neural networks

ABSTRACT The scientific value of the next generation of large continuum surveys would be greatly increased if the redshifts of the newly detected sources could be rapidly and reliably estimated. Given the observational expense of obtaining spectroscopic redshifts for the large number of new detections expected, there has been substantial recent work on using machine learning techniques to obtain photometric redshifts. Here, we compare the accuracy of the predicted photometric redshifts obtained from deep learning (DL) with the k-nearest neighbour (kNN) and the decision tree regression (DTR) algorithms. We find using a combination of near-infrared, visible, and ultraviolet magnitudes, trained upon a sample of Sloan Digital Sky Survey quasi-stellar objects, that the kNN and DL algorithms produce the best self-validation result with a standard deviation of σΔz = 0.24 (σΔz(norm) = 0.11). Testing on various subsamples, we find that the DL algorithm generally has lower values of σΔz, in addition to exhibiting a better performance in other measures. Our DL method, which uses an easy to implement off-the-shelf algorithm with neither filtering nor removal of outliers, performs similarly to other, more complex, algorithms, resulting in an accuracy of Δz < 0.1 up to z ∼ 2.5. Applying the DL algorithm trained on our 70 000 strong sample to other independent (radio-selected) data sets, we find σΔz ≤ 0.36 (σΔz(norm) ≤ 0.17) over a wide range of radio flux densities. This indicates much potential in using this method to determine photometric redshifts of quasars detected with the Square Kilometre Array.

Curran, S. J.↗

SplitML

SplitML (Signal Processing Library for Interference rejecTion by Machine Learning) is a code repository for a set of tools for interference rejection in complex time-domain signals. The goal of the tools is to provide machine learning modeling capabilities for rejecting interference. Recent related machine learning algorithms for signal processing have focused mostly on speech enhancement or multi-speaker speech separation; the tools in SplitML will extend these innovations to generic time-domain signals of interest. SplitML includes tools for generating synthetic noisy signals, code for customized machine learning models applicable to interference rejection, and tools for evaluating such algorithms against standard signal processing techniques. These components are written in Python, a high-level programming language that takes advantage of the Python ecosystem of high-quality open-source packages for machine learning and signal processing.

Klein, Natalie↗