Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “deep learning method”

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 1,369 records · Page 76
All sourcesNASA NTRSDOE OSTI
DOE OSTI2023

Measuring photometric redshifts for high-redshift radio source surveys

With the advent of deep, all-sky radio surveys, the need for ancillary data to make the most of the new, high-quality radio data from surveys like the Evolutionary Map of the Universe (EMU), GaLactic and Extragalactic All-sky Murchison Widefield Array survey eXtended, Very Large Array Sky Survey, and LOFAR Two-metre Sky Survey is growing rapidly. Radio surveys produce significant numbers of Active Galactic Nuclei (AGNs) and have a significantly higher average redshift when compared with optical and infrared all-sky surveys. Thus, traditional methods of estimating redshift are challenged, with spectroscopic surveys not reaching the redshift depth of radio surveys, and AGNs making it difficult for template fitting methods to accurately model the source. Machine Learning (ML) methods have been used, but efforts have typically been directed towards optically selected samples, or samples at significantly lower redshift than expected from upcoming radio surveys. This work compiles and homogenises a radio-selected dataset from both the northern hemisphere (making use of Sloan Digital Sky Survey optical photometry) and southern hemisphere (making use of Dark Energy Survey optical photometry). We then test commonly used ML algorithms such as k-Nearest Neighbours (kNN), Random Forest, ANNz, and GPz on this monolithic radio-selected sample. We show that kNN has the lowest percentage of catastrophic outliers, providing the best match for the majority of science cases in the EMU survey. We note that the wider redshift range of the combined dataset used allows for estimation of sources up to z = 3 before random scatter begins to dominate. When binning the data into redshift bins and treating the problem as a classification problem, we are able to correctly identify ≈ 76% of the highest redshift sources—sources at redshift z > 2.51 —as being in either the highest bin (z > 2.51) or second highest (z = 2.25).

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

Probing for Artifacts: Detecting Imagenet Model Evasions

While deep learning models have made incredible progress across a variety of machine learning tasks, they remain vulnerable to adversarial examples crafted to fool otherwise trustworthy models. In this work we approach this problem through the lens of a detection framework. We propose a classification network that uses the hidden layer activations of a trained model as inputs to detect adversarial artifacts in an input. We train this classification network simultaneously against multiple adversarial algorithms to create a more robust detector and show higher detection rates than several alternatives. The novelty of our approach is in the scale and scope of probing Imagenet models for adversarial artifacts. In addition, we propose an improvement to feature squeezing, another common adversarial example detection method.

Rounds, Jeremiah↗
DOE OSTI2021

Improving scalability of parallel CNN training by adaptively adjusting parameter update frequency

Synchronous SGD with data parallelism, the most popular parallelization strategy for CNN training, suffers from the expensive communication cost of averaging gradients among all workers. The iterative parameter updates of SGD cause frequent communications and it becomes the performance bottleneck. In this paper, we propose a lazy parameter update algorithm that adaptively adjusts the parameter update frequency to address the expensive communication cost issue. Our algorithm accumulates the gradients if the difference of the accumulated gradients and the latest gradients is sufficiently small. Here, the less frequent parameter updates reduce the per-iteration communication cost while maintaining the model accuracy. Our experimental results demonstrate that the lazy update method remarkably improves the scalability while maintaining the model accuracy. For ResNet50 training on ImageNet, the proposed algorithm achieves a significantly higher speedup (739.6 on 2048 Cori KNL nodes) as compared to the vanilla synchronous SGD (276.6) while the model accuracy is almost not affected (<0.2% difference).

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Deep Learning Scene Classification Experiments in Automatic Detection of Slums on Planetscope Imagery

Population growth is increasingly happening in slum settlements of the large urban centers in the Global South. The term "slum" encompasses a wide range of communities, located mostly in underserved areas, and often exhibiting distinct structural and functional informalities with a relatively high concentration of marginalized populations. To address the issues confronting slums for effective planning and development, including the realistic estimation of the resident population, identifying them accurately is fundamental. Given the disagreements over a universal definition, diverse characteristic features, and socio-political limitations, global detection of slums is a veritable challenge. In this paper, we present experiments in slum detection using a scene classification algorithm and 3-meter spatial resolution satellite imagery. We train and evaluate the model for slum detection in Mumbai, India for the year 2023 and test the temporal generalization of the trained model on Mumbai in 2020 and 2018. In addition, we explore the pathways toward geographic generalization to Kolkata and Delhi (India). We discuss several limitations in the workflow and model, situate our findings in the existing literature, and suggest improvements and alternatives. With this, we establish baseline methods and experiments as a first step towards developing an image-based global slum detection framework and algorithm. This work adds to the community discussion on methods, data challenges, and open questions related to the detection of slums globally. With this research, we hope to improve our understanding of human settlements, especially in critical areas, improve population estimates, and help measure progress towards the sustainable development goals.

Arndt, Jacob↗
DOE OSTI2025

Accelerating template generation in resonant anomaly detection searches with optimal transport

We introduce Resonant Anomaly Detection with Optimal Transport (RAD-OT), a method for generating signal templates in resonant anomaly detection searches. RAD-OT leverages the fact that the samples from the conditional probability density of the target features vary approximately linearly along the optimal transport path connecting the resonant feature. This does not assume that the conditional density itself is linear with the resonant feature, allowing RAD-OT to efficiently capture multimodal relationships, changes in resolution, etc. By solving the optimal transport problem, RAD-OT can quickly build a template by interpolating between the background distributions in two sideband regions. We demonstrate the performance of RAD-OT using the LHC Olympics R&D dataset, where we find comparable sensitivity and improved stability with respect to deep learning-based approaches.

Automation↗
DOE OSTI2024

Improved detection of low-frequency within-host variants from deep sequencing: A case study with human papillomavirus

Abstract High-coverage sequencing allows the study of variants occurring at low frequencies within samples, but is susceptible to false-positives caused by sequencing error. Ion Torrent has a very low single nucleotide variant (SNV) error rate and has been employed for the majority of human papillomavirus (HPV) whole genome sequences. However, benchmarking of intrahost SNVs (iSNVs) has been challenging, partly due to limitations imposed by the HPV life cycle. We address this problem by deep sequencing three replicates for each of 31 samples of HPV type 18 (HPV18). Errors, defined as iSNVs observed in only one of three replicates, are dominated by C→T (G→A) changes, independently of trinucleotide context. True iSNVs, defined as those observed in all three replicates, instead show a more diverse SNV type distribution, with particularly elevated C→T rates in CCG context (CCG→CTG; CGG→CAG) and C→A rates in ACG context (ACG→AAG; CGT→CTT). Characterization of true iSNVs allowed us to develop two methods for detecting true variants: (1) VCFgenie, a dynamic binomial filtering tool which uses each variant’s allele count and coverage instead of fixed frequency cut-offs; and (2) a machine learning binary classifier which trains eXtreme Gradient Boosting models on variant features such as quality and trinucleotide context. Each approach outperforms fixed-cut-off filtering of iSNVs, and performance is enhanced when both are used together. Our results provide improved methods for identifying true iSNVs in within-host applications across sequencing platforms, specifically using HPV18 as a case study.

Virology↗
DOE OSTI2025

Motion Dynamics of Motile Microbes in Pore-Networks and its Implications for Reactive Transport Processes

This report outlines new methods to improve simulations of microbial transport and microbially mediated reactions in porous media. A range of experimental, modeling, and machine learning tools are introduced to make these simulations faster, more reliable, and useful for real-world applications. At the microscopic level, the study investigates how different types of bacteria move through confined spaces. A new artificial intelligence tool called DeepTrackStat, is introduced to track motions dynamics as observed in videos of particles migrating through pore networks. This tool is especially helpful for studying fast-moving microbes and requires less computing power than traditional tracking methods. At larger scales, the research looks at how microbes and chemicals interact in zones where surface water and groundwater meet. To connect the small- and large-scale findings, the study presents a neural network model called STAMNet. This tool helps scale up detailed small-scale microbial motion behaviors to predict large-scale environmental changes more efficiently. By combining lab experiments, computer models, and artificial intelligence, the research presented supports smarter environmental decision-making, especially in bioremediation of contaminated groundwater and protection of water quality.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Quantitative Imaging of Cobalt Phthalocyanine Distribution on Carbon Nanotubes: A Deep Learning Approach to Catalyst Characterization

Electrochemical reduction of carbon dioxide (CO 2 ) offers a pathway to valuable products, with catalysts playing a crucial role. This study investigates the distribution of cobalt tetraaminophthalocyanine (CoPc-NH 2 ) immobilized on carbon nanotubes (CNTs), utilizing high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) to characterize CoPc-NH 2 distribution. A challenge in the quantitative HAADF-STEM analysis is the introduction of bias from manual Co atom identification. To address this, we developed and trained a convolutional neural network (CNN) using a data set generated from images of CoPc-NH 2 /CNT samples with varying Co loadings. The CNN, implemented in TensorFlow and Keras, facilitated Co atom detections. Analysis of the CNN-generated data confirmed a correlation between Co loading and surface density, consistent with findings from UV–vis spectroscopy. Furthermore, the application of Ripley’s L(d) function highlighted the presence of slight Co atom clustering. Furthermore, this work demonstrates the utility of the combined HAADF-STEM and CNN approach for providing spatially resolved information about catalyst distribution on nonplanar supports, revealing structural details that are typically lost through other characterization methods.

HAADF-STEM↗
DOE OSTI2026

Microsecond-latency feedback at a particle accelerator by online reinforcement learning on hardware

The commissioning and operation of future large-scale scientific experiments will challenge current tuning and control methods. Reinforcement learning (RL) algorithms are a promising solution due to their ability to dynamically adapt to changing environments and consider delayed consequences. In many real-world applications, RL policies must produce actions in real time, often within microseconds to milliseconds, imposing significant constraints on system latency and computational overhead that conventional machine learning libraries are not designed to handle. To control phenomena in real time at these timescales, RL needs to be deployed on-the-edge, namely on dedicated hardware located near the system it controls, without relying on a host CPU or cloud-based inference. In this work we present the design and deployment of an experience accumulator system in a particle accelerator. In this system, deep-RL algorithms run using hardware acceleration and act within a few microseconds, enabling the use of RL for control of phenomena like beam instabilities. The training uses the collected data offline to reduce the number of operations carried out on the acceleration hardware. The proposed architecture was tested in real experimental conditions at the Karlsruhe research accelerator, a synchrotron light source, where the system was used to control artificially induced horizontal betatron oscillations in real-time, with a control loop period of just 2.7 μs. The results showed a performance comparable to the commercial feedback system available at the accelerator, demonstrating the viability and potential of this approach. Due to the self-learning and reconfiguration capability of this implementation, a seamless application to other control problems is possible. Applications range from particle accelerators to large-scale research and industrial facilities.

FPGA↗
DOE OSTI2020

Enforcing constraints for time series prediction in supervised, unsupervised and reinforcement learning

We assume that we are given a time series of data from a dynamical system and our task is to learn the flow map of the dynamical system. We present a collection of results on how to enforce constraints coming from the dynamical system in order to accelerate the training of deep neural networks to represent the flow map of the system as well as increase their predictive ability. In particular, we provide ways to enforce constraints during training for all three major modes of learning, namely supervised, unsupervised and reinforcement learning. In general, the dynamic constraints need to include terms which are analogous to memory terms in model reduction formalisms. Such memory terms act as a restoring force which corrects the errors committed by the learned flow map during prediction. For supervised learning, the constraints are added to the objective function. For the case of unsupervised learning, in particular generative adversarial networks, the constraints are introduced by augmenting the input of the discriminator. Finally, for the case of reinforcement learning and in particular actor-critic methods, the constraints are added to the reward function. In addition, for the reinforcement learning case, we present a novel approach based on homotopy of the action-value function in order to stabilize and accelerate training. We use numerical results for the Lorenz system to illustrate the various constructions.

Stinis, Panagiotis↗
DOE OSTI2023

Deep learning symmetries and their Lie groups, algebras, and subalgebras from first principles

Abstract We design a deep-learning algorithm for the discovery and identification of the continuous group of symmetries present in a labeled dataset. We use fully connected neural networks to model the symmetry transformations and the corresponding generators. The constructed loss functions ensure that the applied transformations are symmetries and the corresponding set of generators forms a closed (sub)algebra. Our procedure is validated with several examples illustrating different types of conserved quantities preserved by symmetry. In the process of deriving the full set of symmetries, we analyze the complete subgroup structure of the rotation groups SO (2), SO (3), and SO (4), and of the Lorentz group S O ( 1 , 3 ) . Other examples include squeeze mapping, piecewise discontinuous labels, and SO (10), demonstrating that our method is completely general, with many possible applications in physics and data science. Our study also opens the door for using a machine learning approach in the mathematical study of Lie groups and their properties.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Using GANs to predict milling stability from limited data

Milling is a key manufacturing process that requires the selection of operating parameters that provide efficient performance. However, the presence of chatter, a self-excited vibration causing poor surface finish and potential damage to the machine and cutting tool, makes it challenging to select the appropriate parameters. To predict chatter, stability maps are commonly used, but their generation requires expensive data, making it difficult to employ these maps in industry. Therefore, there is a pressing need for an approach that can accurately predict stability maps using limited experimental data. This study introduces the new Encoder GAN (EGAN) approach based on Generative Adversarial Networks (GANs) that predicts stability maps using limited experimental data. The approach consists of the encoder, generator, and discriminator subnetworks and uses the trained encoder and generator to predict the target stability map. This versatile method can be applied to various tool setups and can accurately predict stability maps with limited experimental data (five to 10 cutting tests) even when there is little information available for unknown parameters. In conclusion, the study evaluates the proposed approach using both numerical data and experiments and demonstrates its superior performance compared to state-of-the-art benchmarks.

42 ENGINEERING↗
DOE OSTI2021

Multiscale Data-Driven Seismic Full-Waveform Inversion With Field Data Study

Seismic full-waveform inversion (FWI), which uses iterative methods to estimate high-resolution subsurface models from seismograms, is a powerful imaging technique in exploration geophysics. In recent years, the computational cost of FWI has grown exponentially due to the increasing size and resolution of seismic data. Moreover, it is a nonconvex problem and can encounter local minima due to the limited accuracy of the initial velocity models or the absence of low frequencies in the measurements. To overcome these computational issues, we develop a multiscale data-driven FWI method based on fully convolutional networks (FCNs). In preparing the training data, we first develop a real-time style transform method to create a large set of synthetic subsurface velocity models from natural images. We then develop two convolutional neural networks with encoder-decoder structures to reconstruct the low- and high-frequency components of the subsurface velocity models, separately. To validate the performance of our data-driven inversion method and the effectiveness of the synthesized training set, we compare it with conventional physics-based waveform inversion approaches using both synthetic and field data. Finally, these numerical results demonstrate that, once our model is fully trained, it can significantly reduce the computation time and yield more accurate subsurface velocity models in comparison with conventional FWI.

58 GEOSCIENCES↗
DOE OSTI2026

Reconstruction of atmospheric neutrinos in DUNE’s horizontal-drift far-detector module

This paper reports on the capabilities in reconstructing and identifying atmospheric neutrino interactions in one of the Deep Underground Neutrino Experiment’s (DUNE) far detector modules, a liquid argon time projection chamber (LArTPC) with horizontal drift (FD-HD) of ionization electrons. The reconstruction is based upon the workflow developed for DUNE’s long-baseline oscillation analysis, with some necessary machine-learning models’ retraining and the addition of features relevant only to atmospheric neutrinos such as the neutrino direction reconstruction. Where relevant, the impact of the detection of the charged particles of the hadronic system is emphasized, and comparisons are carried out between the case when lepton-only information is considered in the reconstruction (as is the case for many neutrino oscillation experiments), versus when all particles identified in the LArTPC were included. Three neutrino direction reconstruction methods have been developed and studied for the atmospheric analyses: using lepton-only information, using all reconstructed particles, and using only correlations from reconstructed hits. The results indicate that incorporating more than just lepton information significantly improves the resolution of both neutrino direction and energy reconstruction. The angle reconstruction algorithms developed in this work result in no strong dependence on particle direction for reconstruction efficiencies or neutrino flavor identification. This comprehensive review of the reconstruction of atmospheric neutrinos in DUNE’s FD-HD LArTPC is the first step towards developing a first neutrino oscillation sensitivity analysis, which will ready DUNE for its first measurements.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Machine-Learned Force Field for Molecular Dynamics Simulations of Nonequilibrium Ammonia Synthesis on Iron Catalysts

Ammonia (NH 3 ) is one of the most important industrial chemicals. The conventional NH 3 synthesis method-the Haber–Bosch process-converts atmospheric nitrogen (N 2 ) into NH 3 using H 2 with an iron (Fe) catalyst. However, this process requires high pressures (100–200 atm) and temperatures (700–800 K) near thermal equilibrium. Recently, Fe-based nanocatalysts have been reported to produce promising NH 3 yields under atmospheric pressures and temperature-modulated nonequilibrium conditions. Understanding the mechanism of nonequilibrium catalysis with programmed temperature variation could help to optimize this fully electrified and less energy-intensive process. Although reactive molecular dynamics (RMD) simulations can be a useful tool to model nonequilibrium catalytic processes, they require the development of accurate force fields (i.e., interatomic potentials). Here, we present a machine-learned (ML) force field within the Deep Potential MD (DPMD) framework, trained using periodic density functional theory (DFT) calculations, to model NH 3 synthesis on Fe catalysts with various surface adsorbates such as *N, *H, *N 2 , *H 2 , *NH, *NH 2 , and *NH 3 . Here, we generated the DFT data from static models of elementary reactions on the most stable (110) surface of body-centered cubic Fe, which then were augmented by data from constant number of particles–volume–temperature (NVT) DFT-MD trajectories at various temperatures. Finally, we utilized the fully optimized ML force field to investigate reaction dynamics at an Fe(110) surface at linearly increasing temperatures using NVT-DPMD simulations. Our simulations indicate that pulsed temperature ramping could prove favorable for NH3 synthesis. For example, we conducted ramping under multiple sets of conditions: (i) from 900 to 1200 K over periods of 0.1–0.3 ns for Fe surfaces precovered with N or NH along with H; and (ii) from 300 to 600 K over 0.1–0.3 ns for Fe surfaces precovered with NH 3 . While our simulations so far are limited to short time scales (very rapid heating), these observations shed light on the mechanism of the high NH 3 synthesis rate achieved in a novel temperature-modulated nonequilibrium catalytic reactor using pulsed heating and cooling.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2022

Accurate and Rapid Forecasts for Geologic Carbon Storage via Learning-Based Inversion-Free Prediction

Carbon capture and storage (CCS) is one approach being studied by the U.S. Department of Energy to help mitigate global warming. The process involves capturing CO 2 emissions from industrial sources and permanently storing them in deep geologic formations (storage reservoirs). However, CCS projects generally target “green field sites,” where there is often little characterization data and therefore large uncertainty about the petrophysical properties and other geologic attributes of the storage reservoir. Consequently, ensemble-based approaches are often used to forecast multiple realizations prior to CO 2 injection to visualize a range of potential outcomes. In addition, monitoring data during injection operations are used to update the pre-injection forecasts and thereby improve agreement between forecasted and observed behavior. Thus, a system for generating accurate, timely forecasts of pressure buildup and CO 2 movement and distribution within the storage reservoir and for updating those forecasts via monitoring measurements becomes crucial. This study proposes a learning-based prediction method that can accurately and rapidly forecast spatial distribution of CO 2 concentration and pressure with uncertainty quantification without relying on traditional inverse modeling. The machine learning techniques include dimension reduction, multivariate data analysis, and Bayesian learning. The outcome is expected to provide CO 2 storage site operators with an effective tool for timely and informative decision making based on limited simulation and monitoring data.

58 GEOSCIENCES↗
DOE OSTI2022

Deep Hierarchical Super Resolution for Scientific Data

We present a novel technique for hierarchical super resolution (SR) with neural networks (NNs), which upscales volumetric data represented with an octree data structure to a high-resolution uniform gridwith minimal seam artifacts on octree node boundaries. Our method uses existing state-of-the-art SR models and adds flexibility to upscale input data with varying levels of detail across the domain, instead of only uniform grid data that are supported in previous approaches.The key is to use a hierarchy of SR NNs, each trained to perform 2x SR between two levels of detail, with a hierarchical SR algorithm that minimizes seam artifacts by starting from the coarsest level of detail and working up.We show that our hierarchical approach outperforms baseline interpolation and hierarchical upscaling methods, and demonstrate the usefulness of our proposed approach across three use cases including data reduction using hierarchical downsampling+SR instead of uniform downsampling+SR, computation savings for hierarchical finite-time Lyapunov exponent field calculation, and super-resolving low-resolution simulation results for a high-resolution approximation visualization.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

Physics-informed neural networks for heterogeneous poroelastic media

This study presents a novel physics-informed neural network (PINN) framework for modeling poroelasticity in heterogeneous media with material interfaces. The approach introduces a composite neural network (CoNN) where separate neural networks predict displacement and pressure variables for each material. While sharing identical activation functions, these networks are independently trained for all other parameters. To address challenges posed by heterogeneous material interfaces, the CoNN is integrated with the Interface-PINNs (I-PINNs) framework (Sarma et al., Comput. Methods Appl. Mech. Eng. 429: 117135, 2024), allowing different activation functions across material interfaces. Further, this ensures accurate approximation of discontinuous solution fields and gradients. Performance and accuracy of this combined architecture were evaluated against the conventional PINNs approach, a single neural network (SNN) architecture, and the eXtended PINNs (XPINNs) framework through two one-dimensional benchmark examples with discontinuous material properties. The results show that the proposed CoNN with I-PINNs architecture achieves an RMSE that is two orders of magnitude better than the conventional PINNs approach and is at least 40 times faster than the SNN framework. Compared to XPINNs, the proposed method achieves an RMSE at least one order of magnitude better and is 40% faster.

42 ENGINEERING↗