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 1,783 records · Page 99

CryoSegNet: accurate cryo-EM protein particle picking by integrating the foundational AI image segmentation model and attention-gated U-Net

Picking protein particles in cryo-electron microscopy (cryo-EM) micrographs is a crucial step in the cryo-EM-based structure determination. However, existing methods trained on a limited amount of cryo-EM data still cannot accurately pick protein particles from noisy cryo-EM images. The general foundational artificial intelligence–based image segmentation model such as Meta’s Segment Anything Model (SAM) cannot segment protein particles well because their training data do not include cryo-EM images. Here, we present a novel approach (CryoSegNet) of integrating an attention-gated U-shape network (U-Net) specially designed and trained for cryo-EM particle picking and the SAM. The U-Net is first trained on a large cryo-EM image dataset and then used to generate input from original cryo-EM images for SAM to make particle pickings. CryoSegNet shows both high precision and recall in segmenting protein particles from cryo-EM micrographs, irrespective of protein type, shape and size. On several independent datasets of various protein types, CryoSegNet outperforms two top machine learning particle pickers crYOLO and Topaz as well as SAM itself. The average resolution of density maps reconstructed from the particles picked by CryoSegNet is 3.33 Å, 7% better than 3.58 Å of Topaz and 14% better than 3.87 Å of crYOLO. It is publicly available at https://github.com/jianlin-cheng/CryoSegNet

59 BASIC BIOLOGICAL SCIENCES↗

Sci-Fri AM: Quality, Safety, and Professional Issues 04: Predicting waiting times in Radiation Oncology using machine learning

We describe a method for predicting waiting times in radiation oncology. Machine learning is a powerful predictive modelling tool that benefits from large, potentially complex, datasets. The essence of machine learning is to predict future outcomes by learning from previous experience. The patient waiting experience remains one of the most vexing challenges facing healthcare. Waiting time uncertainty can cause patients, who are already sick and in pain, to worry about when they will receive the care they need. In radiation oncology, patients typically experience three types of waiting: Waiting at home for their treatment plan to be prepared Waiting in the waiting room for daily radiotherapy Waiting in the waiting room to see a physician in consultation or follow-up These waiting periods are difficult for staff to predict and only rough estimates are typically provided, based on personal experience. In the present era of electronic health records, waiting times need not be so uncertain. At our centre, we have incorporated the electronic treatment records of all previously-treated patients into our machine learning model. We found that the Random Forest Regression model provides the best predictions for daily radiotherapy treatment waiting times (type 2). Using this model, we achieved a median residual (actual minus predicted value) of 0.25 minutes and a standard deviation residual of 6.5 minutes. The main features that generated the best fit model (from most to least significant) are: Allocated time, median past duration, fraction number and the number of treatment fields.

60 APPLIED LIFE SCIENCES↗

Alert Triage v 0.1 beta

In the cyber security operations of a typical organization, data from multiple sources are monitored, and when certain conditions in the data are met, an alert is generated in an alert management system. Analysts inspect these alerts to decide if any deserve promotion to an event requiring further scrutiny. This triage process is manual, time-consuming, and detracts from the in-depth investigation of events. We have created a software system that uses supervised machine learning to automatically prioritize these alerts. In particular we utilize active learning to make efficient use of the pool of unlabeled alerts, thereby improving the performance of our ranking models over passive learning. We have demonstrated the effectiveness of our system on a large, real-world dataset of cyber security alerts.

Doak, JustinE.↗

Reconfiguring the metabolism of photosynthetic microbes for their development as biotechnological platforms [Slides]

LANL is at the forefront of genetic engineering of microalgae. We have developed genetic engineering toolboxes for many strains, including Picochlorum soloecismus, Nannochloropsis salina, and Chlorella sorokiniana. We have implemented such toolboxes for generating mutants with favorable phenotypes. We have the opportunity to integrate metabolic features from different species (cyanobacteria>microalgae>plants) into a synthetic biology discovery and developmental platform. We can leverage the cyanobacterial metabolic “simplicity” to engineer complex organisms, i.e. for the production of renewable polymers and unrivaled ‘omics and machine learning scientific collaboration.

59 BASIC BIOLOGICAL SCIENCES↗

Secondary organic aerosols derived from intermediate-volatility n-alkanes adopt low-viscous phase state

Abstract. Secondary organic aerosol (SOA) derived from n-alkanes, as emitted from vehicles and volatile chemical products, is a major component of anthropogenic particulate matter, yet the chemical composition and phase state are poorly understood and thus poorly constrained in aerosol models. Here we provide a comprehensive analysis of n-alkane SOA by explicit gas-phase chemistry modeling, machine learning, and laboratory experiments to show that n-alkane SOA adopts low-viscous semi-solid or liquid states. Our study underlines the complex interplay of molecular composition and SOA viscosity: n-alkane SOA with a higher carbon number mostly consists of less functionalized first-generation products with lower viscosity, while the SOA with a lower carbon number contains more functionalized multigenerational products with higher viscosity. This study opens up a new avenue for analysis of SOA processes, and the results indicate few kinetic limitations of mass accommodation in SOA formation, supporting the application of equilibrium partitioning for simulating n-alkane SOA formation in large-scale atmospheric models.

54 ENVIRONMENTAL SCIENCES↗

Alert Triage v 0.1 beta

In the cyber security operations of a typical organization, data from multiple sources are monitored, and when certain conditions in the data are met, an alert is generated in an alert management system. Analysts inspect these alerts to decide if any deserve promotion to an event requiring further scrutiny. This triage process is manual, time-consuming, and detracts from the in-depth investigation of events. We have created a software system that uses supervised machine learning to automatically prioritize these alerts. In particular we utilize active learning to make efficient use of the pool of unlabeled alerts, thereby improving the performance of our ranking models over passive learning. We have demonstrated the effectiveness of our system on a large, real-world dataset of cyber security alerts.

Doak, Justin E.↗

AI‐Driven Defect Engineering for Advanced Thermoelectric Materials

Thermoelectric materials offer a promising pathway to directly convert waste heat to electricity. However, achieving high performance remains challenging due to intrinsic trade-offs between electrical conductivity, the Seebeck coefficient, and thermal conductivity, which are further complicated by the presence of defects. This review explores how artificial intelligence (AI) and machine learning (ML) are transforming thermoelectric materials design. Advanced ML approaches including deep neural networks, graph-based models, and transformer architectures, integrated with high-throughput simulations and growing databases, effectively capture structure-property relationships in a complex multiscale defect space and overcome the “curse of dimensionality”. This review discusses AI-enhanced defect engineering strategies such as composition optimization, entropy and dislocation engineering, and grain boundary design, along with emerging inverse design techniques for generating materials with targeted properties. Finally, it outlines future opportunities in novel physics mechanisms and sustainability, highlighting the critical role of AI in accelerating the discovery of thermoelectric materials.

36 MATERIALS SCIENCE↗

G2Aero Database of Airfoils - Curated Airfoils

This dataset contains a curated set of 19,164 airfoil shapes from various applications and the data-driven design space of separable shape tensors (PGA space), which can be used as a parameter space for machine-learning applications focused on airfoil shapes. We constructed the airfoil dataset in two main stages. First, we identified 13 baseline airfoils from the NREL 5MW and IEA 15MW reference wind turbines. We reparameterized these shapes using least-squares fits of 8-order CST parametrizations, which involve 18 coefficients. By uniformly perturbing all 18 CST coefficients by +/-20% around each baseline airfoil, we generated 1,000 unique airfoils. Each airfoil was sampled with 1,001 shape landmarks whose x-coordinates followed a cosine distribution along the chord. This process resulted in a total of 13,000 airfoil shapes, each with 1,001 landmarks. In the second phase, we gathered additional airfoils from the extensive BigFoil database, which consolidates data from sources such as the University of Illinois Urbana-Champaign (UIUC) airfoil database, the JavaFoil database, the NACA-TR-824 database, and others. We undertook a thorough pre-processing step to filter out shapes with sparse, noisy, or incomplete data. We also removed airfoils with sharp leading edge and those exceeding our threshold for trailing edge thickness. Additionally, we thinned out the collection of NACA airfoils-- parametric sweeps of NACA airfoils with increasing thickness and camber present in BigFoil database-- by selecting every fourth step in the parameter sweeps. Finally, we regularized the airfoils by reparametrizing them with an 8-order CST parametrization (with 1,001 shape landmarks with x coordinated following cosine distribution along the chord) and removing airfoils with high reconstruction errors. This data pre-processing resulted in a set of 6,164 airfoils. In total, our curated airfoil dataset comprises 19,164 airfoils, each with 1,001 landmarks, and is stored in the curated_airfoils.npz file. Using this curated airfoil dataset, we utilized the separable shape tensors framework to develop a data-driven parameterization of airfoils based on principal geodesic analysis (PGA) of separable shape tensors. This PGA space is provided in PGAspace.npz file.

airfoils↗

Adversarial autoencoder ensemble for fast and probabilistic reconstructions of few-shot photon correlation functions for solid-state quantum emitters

Second-order photon correlation measurements [g (2) (τ) functions] are widely used to classify single-photon emission purity in quantum emitters or to measure the multiexciton quantum yield of emitters that can simultaneously host multiple excitations – such as quantum dots – by evaluating the value of g (2) (τ = 0). Accumulating enough photons to accurately calculate this value is time consuming and could be accelerated by fitting of few-shot photon correlations. Here, we develop an uncertainty-aware, deep adversarial autoencoder ensemble (AAE) that reconstructs noise-free g (2) (τ) functions from noise-dominated, few-shot inputs. The model is trained with simulated g (2) (τ) functions that are facilely generated by Poisson sampling time bins. The AAE reconstructions are performed orders-of-magnitude faster, with reconstruction errors and estimates of g (2) (τ = 0) that are lower in variance and similar in accuracy compared to Maximum likelihood estimation and Levenberg-Marquardt least-squares fitting approaches, for simulated and experimentally measured few-shot g (2) (τ) functions (~100 two-photon events) of InP/ZnS/ZnSe and CdS/CdSe/CdS quantum dots. The deep-ensemble model comprises eight individual autoencoders, allowing for probabilistic reconstructions of noise-free g (2) (τ) functions, and we show that the predicted variance scales inversely with number of shots, with comparable uncertainties to computationally intensive Markov chain Monte Carlo sampling. Furthermore, this work demonstrates the advantage of machine learning models to perform uncertainty-aware, fast, and accurate reconstructions of simple Poisson-distributed photon correlation functions, allowing for on-the-fly reconstructions and accelerated materials characterization of solid-state quantum emitters.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Accurate Prediction of Algal Biomass Lipid, Protein, and Carbohydrate Composition with Machine Learning Regression Modelling of Near-IR Spectra

During large scale algal biomass cultivation, it is difficult to reliably control relative composition to target levels. Rapid determination of chemical composition is feasible by using near infrared (NIR) spectral data. We sought to build and improve on reliable high-throughput screening prediction method based on partial least squares regression (PLSR) by the application of artificial neural networks (ANN) and associated optimization strategies. The algal biomass sample set was designed and created in an iterative process of culturing in physiologically diverse conditions at the GAI field site, followed by compositional analyses at NREL. The workflow allowed us to identify gaps in compositional space for informing the subsequent cultivation and sampling efforts and generated a high quality set of 210 unique samples with chemical analysis results, spectral scanning data, and cultivation metadata. We observed a significant improvement in the performance of carbohydrate content predictions using an optimized ANN model compared to PLSR, with > 16% reduction in mean absolute percent error (MAPE) when tested on the same set of reserved data. The optimized ANN models for FAME and protein prediction performed exceptionally well with 5.99% and 5.09% MAPE, respectively. Application of these methods to detection and quantification of minor biomass constituents that are relevant to certain product streams has shown positive preliminary results, opening the possibility for extensions to the outputs of this powerful data type. All models are accompanied by prediction uncertainties and unsupervised spectral outlier detection to alert an operator to unreliable spectral data. These tools can be deployed for rapid determination of algal culture status, and cultivation and biomass quality improvement.

algal biofuels↗

Computationally Efficient Training of Deep Neural Networks via Transfer Learning

Transfer learning is a highly successful approach to train deep neural networks for customized image recognition tasks. Transfer learning leverages deep learning models that are previously trained on massive datasets and re-trains the network for a novel image recognition dataset. Typically, the advantage of transfer learning has been measured in sample efficiency, but instead, we investigate the computational efficiency of transfer learning. A good pre-trained model provides features that can be used as input to a new classifier (usually the top layers of a neural network). We show that if a good pre-trained model is selected, that training a new classifier can be much more computationally efficient than training a deep neural network without transfer learning. The first step in transfer learning is to select a pre-trained model to use as a feed-forward network for generating features. Here this selection process either uses human intuition/convenience or a methodical but computationally intensive validation loop. Instead, we would prefer a method to select the pre-trained model that will produce the best transfer results for the least amount of computation. To this end, we provide a computationally efficient metric for the fit between a pre-trained model and a novel image recognition task. The better the fit, the less computation will be needed to re-train the pre-trained model for the novel task. As machine learning becomes ubiquitous, highly-accurate trained models will proliferate and computationally efficient transfer learning methods will enable rapid development of new image recognition models.

97 MATHEMATICS AND COMPUTING↗

Automated defect identification in electroluminescence images of solar modules

Solar photovoltaic (PV) modules are susceptible to manufacturing defects, mishandling problems or extreme weather events that can limit energy production or cause early device failure. Trained professionals use electroluminescence (EL) images to identify defects in modules, however, field surveys or inline image acquisition can generate millions of EL images, which are infeasible to analyze by rote inspection. Here, we develop a rapid automatic computer vision pipeline (~0.5 seconds/module) to analyze EL images and identify defects including cracks, intra-cell defects, oxygen-induced defects, and solder disconnections. Defect identification is achieved with a machine learning model (Random Forest, ResNet models and YOLO) trained on 762 manually-labeled EL images of PV modules. We compare model performance on an imbalanced real-world validation set containing 134 EL images and determine that ResNet18 and YOLO are the optimal models; we next evaluated these models on a dedicated testing set (129 module images) with resulting macro F1 scores of 0.83 (ResNet18) and 0.78 (YOLO). Using a field EL survey of a PV power plant damaged in a vegetation fire, we analyze 18,954 EL images (2.4 million cells) and inspect the spatial distribution of defects on the solar modules. The results find increased frequency of ‘crack’, ‘solder’ and ‘intra-cell’ defects on the edges of the solar module closest to the ground after fire. We also find an abnormal increase of striation rings on cells which were assumed to be caused mainly in fabrication process. Our methods are published as open-source software. It can also be used to identify other kinds of defects or process different types of solar cells with minor modification on models by transfer learning.

14 SOLAR ENERGY↗

Microscopic Mechanisms of Superionic Na-ion Conductivity in Crystalline and Amorphous NaMOCl 4 (M = Nb, Ta) Solid Electrolytes

Sodium-ion solid electrolytes offer a sustainable route toward next-generation batteries, but few match the performance of their lithium counterparts. Halide-based NaMOCl 4 (M = Nb, Ta) has recently emerged as a promising analogue to LiMOCl 4 , yet its structure–transport relationships remain unclear due to poor crystallinity in experiments. Here, we combine density functional theory and machine-learned molecular dynamics to reveal that crystalline NaMOCl 4 exhibits negligible room-temperature conductivity with high activation barriers arising from vacancy-mediated diffusion below an order–disorder transition. Above this transition, rotational and translational motion of the [MO 2/2 Cl 4 – ] ∞ chains create new Na sites and enhances transport. In contrast, the amorphous phase inherently supports facile, three-dimensional Na diffusion through dynamic framework flexibility. These results show that ordered crystalline phases hinder ionic transport, while disorder – either thermally induced or structural – facilitates it, revising prior assumptions from the Li system and providing design principles for high-conductivity Na halide electrolytes.

Wei, Grace [University of California, Berkeley, CA↗

The Illumination of Thunderclouds by Lightning: 4. Volumetric Thunderstorm Imagery

Optical instruments such as the Geostationary Lightning Mapper (GLM) detect lightning based on transient changes in cloud illumination. The horizontal location of lightning is determined from the coordinates of the pixels on the imaging array illuminated during the flash. However, the vertical position of the lightning pulses (approximated by GLM “groups”) below the cloud top cannot be routinely measured from a single space-based instrument. In our prior work, we have developed a machine learning algorithm that can infer optical source altitude for a given pulse based on how the optical energy is distributed across the group footprint and the local Advanced Baseline Imager Cloud-Top Height (CTH). In this fourth part of our thundercloud illumination study, we leverage these source altitudes to generate volumetric GLM imagery of a Colombia thunderstorm. We find that 3D versions of the current GLM meteorological imagery products (that describe thunderstorm kinematics) and thundercloud imagery products (that depict how the flashes appear from space) provide additional insights into lightning activity in the thunderstorm that are lost in the vertical integration used to generate the current 2D GLM gridded products. This new volumetric imaging capability provides a more comprehensive picture of where lightning occurs in the storm, how its physical characteristics vary across three-dimensional space, and how its optical emissions interact with surrounding the cloud medium.

54 ENVIRONMENTAL SCIENCES↗

Accurate prediction of X-ray pulse properties from a free-electron laser using machine learning

Free-electron lasers providing ultra-short high-brightness pulses of X-ray radiation have great potential for a wide impact on science, and are a critical element for unravelling the structural dynamics of matter. To fully harness this potential, we must accurately know the X-ray properties: intensity, spectrum and temporal profile. Owing to the inherent fluctuations in free-electron lasers, this mandates a full characterization of the properties for each and every pulse. While diagnostics of these properties exist, they are often invasive and many cannot operate at a high-repetition rate. Here, we present a technique for circumventing this limitation. Employing a machine learning strategy, we can accurately predict X-ray properties for every shot using only parameters that are easily recorded at high-repetition rate, by training a model on a small set of fully diagnosed pulses. Lastly, this opens the door to fully realizing the promise of next-generation high-repetition rate X-ray lasers.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DIII-D High Field Side Lower Hybrid Current Drive Experiment Overview

In preparation for high field side lower hybrid current drive (HFS LHCD) experiments in DIII-D, the HFS LHCD was to be commissioned and physics experiments commence once the system operated up to 300 kW for 0.5 s. The initial physics experiments sought to characterize coupling, wave propagation, and driven current measurements. In HFS LHCD first campaign, the maximum power was limited to <200 kW due to waveguide pressure leaks limiting the number of available modules and power per module. Here we summarize commissioning progress and initial physics observations. The HFS LHCD coupler was optimized for high qmin, DIII-D discharges where efficient off-axis current at r/a~0.6-0.8 is desired. The coupler n|| spectrum is peaked at 2.7 and is predicted to generate ~0.14 MA/MW coupled for 1.6 MW injected. In preparation, the HFS scrape-off layer density profile was characterized and found to have steeper profiles and lower fluctuation levels than the low field side. Furthermore, the HFS SOL density profile can be accurately predicted using global plasma quantities using machine learning. Thus far, one module has injected ~100 kW for 0.5 s with <5% reflected power. Nonthermal electrons have been observed on lower frequency channels of the electron cyclotron emission radiometer correlated with the LH power indicating core wave absorption. To avoid 30R neutral beam heat flux, a split launcher is proposed to avoid the high heat flux region while maintaining power spectrum and directivity.

Wukitch, Stephen J. [Massachusetts Inst. of Techno↗

Machine learning surrogate for charged particle beam dynamics with space charge based on a recurrent neural network with aleatoric uncertainty

In this work, we develop a machine learning (ML) model with aleatoric uncertainty for the low energy beam transport (LEBT) region of the LANSCE linear accelerator in which we model the transport of a space-charge-dominated 750 keV proton beam through a lattice of 22 quadrupole magnets. Our ML model is developed based on data generated by a Kapchinsky–Vladimirsky (KV) envelope model of beam transport. We show that a recurrent neural network can be used as a dynamical surrogate model for fast prediction of the LEBT beam envelope. Furthermore, we endow the model with the prediction of aleatoric uncertainty and compare three different approaches. We demonstrate that the ML-based uncertainty quantification models are well calibrated and produce good estimates of the regions where the model is less certain about its predictions. This ML framework is a necessary step in the development of a real-time virtual diagnostic tool with uncertainty quantification that can be integrated into more complex downstream tasks (e.g., adaptive control or learning flexible control policies via reinforcement learning) for improved efficiency in beam operations. In future work, we plan to expand on this preliminary study by considering more realistic envelope models that include longitudinal momentum spread and dispersive effects in bending magnets, as well as particle tracking codes with 3D space charge (such as and ). Published by the American Physical Society 2024

43 PARTICLE ACCELERATORS↗