SEARCH · Search NASA

Results for “Feature Extraction”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 973 records · Page 54

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Energy dependence of polarized 𝛾⁢𝛾 → 𝑒 + ⁢ 𝑒− in peripheral Au+Au collisions at $\sqrt{s_{NN}}$ = 54.4 and 200 GeV with the STAR experiment at RHIC

We report the differential yields at mid-rapidity of the Breit-Wheeler process (𝛾⁢𝛾 → 𝑒 + ⁢𝑒 − ) in peripheral Au+Aucollisions at $\sqrt{s_{NN}}$ = 54.4 and 200 GeV with the STAR experiment at the Relativistic Heavy Ion Collider (RHIC), as a function of energy $\sqrt{s_{NN}}$, 𝑒 + ⁢𝑒 − transverse momentum 𝑝 T , 𝑝$^{2}_{T}$, invariant mass 𝑀 𝑒⁢𝑒 , and azimuthal angle. In the invariant mass range of 0.4 < 𝑀 𝑒⁢𝑒 < 2.6GeV/𝑐 2 at low transverse momentum (𝑝 T < 0.15GeV/𝑐), the yields increase while the pair √⟨𝑝$^{2}_{T}$⟩ decreases with increasing $\sqrt{s_{NN}}$, a feature that is correctly predicted by the QED calculation. Here, the energy dependencies of the measured quantities are sensitive to the nuclear form factor, infrared divergence and photon polarization. The data are compiled and used to extract the charge radius of the Au nucleus.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2021

(U) A Theoretical Study of Asay Foil Trajectories (V.02)

We consider the trajectory of a generic Asay foil ejecta momentum diagnostic sensor, for scenarios where ejecta are produced at a planar surface and fly ballistically through a perfect vacuum to the sensor. To do so, we build upon a previously established mathematical formalism derived for the analytic study of stationary sensors (i.e., piezopins). First, we derive the momentum conservation equation for the problem, in a form amenable to accelerating sensors, in terms of a generic ejecta source areal mass function (“source model”). This defines an integro-differential equation (IDE) for the foil trajectory. When ejecta production is instantaneous - as is generally assumed in momentum diagnostic data analyses - the IDE leads to an implicit and easily calculable closed-form solution for the foil trajectory in a perfect system, as long as the ejecta particle velocity distribution is twice-integrable. General properties of the instant-production trajectory solution indicate the existence of a boundary condition the particle velocity distribution must satisfy in order for the analytically predicted foil trajectories to be compatible with certain features commonly observed in foil data. This boundary condition is identical to one derived previously from a consideration of piezopin data. Armed with the analytic solution for instant production, we also consider various techniques used to extract time-dependent cumulative ejecta masses from foil trajectories, and derive an expression for the error imposed by using an approximated equation of motion. This analytic trajectory solution furthermore makes it possible to examine the common practice of presenting inferred cumulative ejecta masses as a function of a normalized implied velocity; we derive conditions under which this methodology is and is not meaningful. We also propose a strategy for extending the instant-production trajectory solution to time-dependent source functions.

42 ENGINEERING↗

DOE OSTI2021

DEIMoS: An Open-Source Tool for Processing High-Dimensional Mass Spectrometry Data

We present DEIMoS: Data Extraction for Integrated Multidimensional Spectrometry, a Python application programming interface and command-line tool for high-dimensional mass spectrometry data analysis workflows, offering ease of development and access to efficient algorithmic implementations. Functionality includes feature detection, feature alignment, collision cross section calibration, isotope detection, and MS/MS spectral deconvolution, with the output comprising detected features aligned across study samples and characterized by mass, CCS, tandem mass spectra, and isotopic signature. Notably, DEIMoS operates on N-dimensional data, largely agnostic to acquisition instrumentation: algorithm implementations utilize all dimensions simultaneously to (i) offer greater separation between features, improving detection sensitivity, (ii) increase alignment/feature matching confidence among datasets, and (iii) mitigate convolution artifacts in tandem mass spectra. We demonstrate DEIMoS with LC-IMS-MS/MS data, demonstrating the advantages of a multidimensional approach in each data processing step.

Colby, Sean M↗

DOE OSTI2019

Deep learning as a tool for neural data analysis: Speech classification and cross-frequency coupling in human sensorimotor cortex

A fundamental challenge in neuroscience is to understand what structure in the world is represented in spatially distributed patterns of neural activity from multiple single-trial measurements. This is often accomplished by learning a simple, linear transformations between neural features and features of the sensory stimuli or motor task. While successful in some early sensory processing areas, linear mappings are unlikely to be ideal tools for elucidating nonlinear, hierarchical representations of higher-order brain areas during complex tasks, such as the production of speech by humans. Here, we apply deep networks to predict produced speech syllables from a dataset of high gamma cortical surface electric potentials recorded from human sensorimotor cortex. We find that deep networks had higher decoding prediction accuracy compared to baseline models. Having established that deep networks extract more task relevant information from neural data sets relative to linear models (i.e., higher predictive accuracy), we next sought to demonstrate their utility as a data analysis tool for neuroscience. We first show that deep network's confusions revealed hierarchical latent structure in the neural data, which recapitulated the underlying articulatory nature of speech motor control. We next broadened the frequency features beyond high-gamma and identified a novel high-gamma-to-beta coupling during speech production. Finally, we used deep networks to compare task-relevant information in different neural frequency bands, and found that the high-gamma band contains the vast majority of information relevant for the speech prediction task, with little-to-no additional contribution from lower-frequency amplitudes. Together, these results demonstrate the utility of deep networks as a data analysis tool for basic and applied neuroscience.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

Comprehensive Neural Posterior Estimation for Galaxy-Galaxy Strong Lensing

We present a deep learning model based on neural posterior estimation (NPE) for comprehensive extraction of astrophysical parameters from galaxy-scale strong gravitational lenses. The unprecedentedly large amount of galaxy-scale strong lenses expected in future cosmological surveys (${\cal O}(10^5)$) promises to enable valuable statistical constraints in various studies ranging from galaxy formation to the nature of dark matter, but it also poses a significant challenge for traditional modelling pipelines. To this end, our automated model includes several new, state-of-the-art features and approaches leveraging the framework of simulation-based inference (SBI). We infer a total of 20 parameters describing the mass and light profiles of both lens and source galaxies, using simulated raw multi-band data modelled under noise and observing conditions expected by the Legacy Survey of Space and Time (LSST), with its summary statistics generated by a residual network. We examine the efficacy of multi-band data in extracting nearly 20 model parameters simultaneous from strong lensing images including lens light. Finally, We perform a comprehensive set of diagnostics for SBI models, evaluating the model's prediction accuracy, stability, and uncertainty quantification.

Zhao, Roy J. [Chicago U., KICP]↗

DOE OSTI2018

Influence of Microbial Membranes on Biofilm Formation, Scaling Reactions, and Production Chemistry in Engineered Energy Systems (Final Report)

Humans are engineering the deep subsurface for the purposes of energy-capture, energy-storage, and environmental mitigation. Despite the potential for microbial communities to significantly alter the effectiveness of these technologies, we have historically lacked a clear understanding of how microbial cells interact in a fractured shale system. Existing knowledge has been heavily based on genomic information, lacking studies on in situ physiologies, kinetic rates, and attachment mechanisms governing biogeochemical reactions in the system. The overarching goal of this research was to advance understanding of engineered hydrocarbon extraction systems as they evolve from newly developed environments to mature use. Plasma membranes protect microbial cells, mediate many critical functions, and respond to intracellular cues and ecological perturbations through physicochemical modifications. The project objectives were to 1) characterize variables influencing growth and membrane features of shale taxa, 2) characterize the interactions between shale matrices and microorganisms and 3) elucidate engineered and environmental processes driving biogeochemical signatures at the field scale. Objectives were realized through laboratory experiments involving growth of pure cultures and enrichments in continuous culture and biofilms; access to samples from two DOE-NETL funded energy-capture sites: Marcellus Shale Energy and Environmental Laboratory (MSEEL and MSEEL II); and through collaborations with scientists and engineers at the Environmental Molecular Science Laboratory (EMSL) and National Energy Technology Laboratory (NETL). Key findings showed (a) Halanaerobium congolense WG10 makes its membrane more fluid and thicker through increasing its polyunsaturated fatty acid (PUFA) composition under non-optimal salinities (7% and 20%) compared to the optimum (13% NaCl); (b) anionic phospholipids within the plasma membrane are important for high salinity tolerance in H. congolense WG10 while zwitterionic species modulate biofilm physiologies in both isolate and mixed cultures, and (c) tyrosine accumulation with overexpression of Na+/H+ antiporters are essential for osmoprotection in H. congolense WG10 under high salinity while upregulation of aromatic amino acid synthesis and cell envelope maintenance underpin biofilm formation. It is clear from these results that microbial persistence in fractured shale reservoirs under environmental and engineered perturbations requires strategic plasma membrane lipid chemistry adjustments and metabolic restructuring, with implications for improved shale microbiology management and more efficient hydrocarbon extraction approaches targeting microorganisms.

03 NATURAL GAS↗

DOE OSTI2023

Convoluted filtering for process cycle modeling

Principles of materials science and engineering, physics, mathematics, and information science are used to extract knowledge and insights from the process-structure–property-performance relationships hidden in materials data. The process-structure modeling can be accelerated without loss of interpretability, with artificial intelligence tools that mimic the salient features of the process and process-structure relations. In this work, a novel convoluted model-filtering technique was exploited to build and successfully train the Convoluted Filter (CoFi) artifacts for Fe-based alloy heat treatment cycles. The artifacts were pre-trained to filter out deep models that change the surrogate microstructure state after the heat treatment at ambient conditions. Direct representation of the thermal cycle features within knowledge Graph facilitated development of meaningful data models for microstructure evolution, which reduce overfitting to limited datasets.

36 MATERIALS SCIENCE↗

DOE OSTI2020

Fractal breakage of porous carbonate sand particles: Microstructures and mechanisms

In situ, three-dimensional (3D) characterizations of particle breakage in porous carbonate sands are presented, for the first time, with synchrotron-based micro computed tomography. Evolution of grain-scale characteristics are identified and quantified via elaborate image processing and topology analyses. The sequential 3D images reveal distinctly different fracture mechanisms for carbonate sands from silica sands. The angular shape of carbonate sand particles facilitates bending fracture, and particles with a lower sphericity and a higher porosity are more prone to break. 3D crack networks extracted from fractured particles imply considerable cleavage along initial pores. The fractal dimension of crack networks increases with external loading due to crack branching via cleavage. The resultant fragment size distribution also appears fractal and the fractal feature is valid down to the breakage limit of calcium carbonate. Crack propagation along the initial pores reduces the energy barrier for particle breakage and thus fracture strength of particles

Carbonate sands↗

DOE OSTI2019

Methylation protocol for the retrospective detection of isopropyl-, pinacolyl- and cyclohexylmethylphosphonic acids, indicative markers for the nerve agents sarin, soman and cyclosarin, at low levels in soils using EI-GC–MS

A practical and efficient protocol for the derivatization and detection by GC-EI-MS of isopropyl-, pinacolyl- and cyclohexylmethylphosphonic acids, key diagnostic degradation products of the nerve agents sarin, soman and cyclosarin respectively, in six different types of soil matrices is presented here. The method involves the in situ conversion of the phosphonic acids to their respective methyl esters using trimethyloxonium tetrafluoroborate when present in the soils at low levels (10 μg g -1 ) without any prior extractions or soil preparation. The soils employed in our study were Nebraska EPA soil, Georgia soil, silt, Virginia type A soil, regular sand and Ottawa sand and were chosen for their vast differences in composition and physical features. Appealing attributes of the protocol include its rapidity (t < 30 min), mildness (ambient temperature), and practicality that includes the production of the phosphonic methyl esters that can be easily detected by GC-EI–MS and corroborated with the instrument's internal NIST spectral library or the Organisation for the Prohibition of Chemical Weapons (OPCW) central analytical database (OCAD v.21_2019). The overall efficacy of the protocol was then tested on a soil sample featured in the 44th OPCW PT that our laboratory participated in. After preparing the soil so as to give pinacolyl methylphosphonic acid at a 5 μg g -1 concentration, the acid was successfully methylated and detected by GC-EI-MS. The protocol's performance mirrors that of the universally employed diazomethane protocol but accomplishes this without any of the explosive hazards and time consuming reagent preparation commonly associated with it.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2025

Denoising Autoencoder for Reconstructing Sensor Observation Data and Predicting Evapotranspiration: Noisy and Missing Values Repair and Uncertainty Quantification

Abstract Machine learning (ML) methods applied in scientific research often deal with interrelated features in high‐dimensional data. Reducing data noise and redundancy is needed to increase prediction accuracy and efficiency especially when dealing with data from field sensors. We explored an unsupervised learning method, the denoising autoencoder (DAE), to extract the underlying data structure from noisy raw data in the context of predicting hydrologic quantities from multiple field sensors. These sensors have intrinsic instrumental noise and occasional malfunctions that cause missing values. Our DAE neural network reconstructed meteorological sensor data containing noise and missing values to predict evapotranspiration in a mountainous watershed. The DAE reconstructed the sensor variables with a mean coefficient of determination value of 0.77 across 15 dimensions representing individual sensors. It reduced variance and bias uncertainties compared to a classical autoencoder model. The reconstruction quality varied across dimensions depending on their cross‐correlation and alignment with the underlying data structure. Uncertainties arising from the model structure were overall higher than those resulting from data corruption. We attached the DAE structure to a downstream ET‐prediction neural network in three formats and achieved reasonably accurate ET predictions . The use of the DAE notably reduced variance uncertainty in ET prediction. However, excessive variance reduction may be accompanied by an increase in bias due to the intrinsic bias‐variance tradeoff. Our method of evaluating and reducing uncertainties in aggregated data from different sources can be used to improve predictive models, process understanding, and uncertainty quantification for better water resource management. Plain Language Summary We present a machine learning method, namely the denoising autoencoder, which reduces the effects of data noise and missing values typically present in scientific data sets collected through sensor measurements. This method selects the most relevant information from noisy raw data collected by the instruments and fills in missing values. To demonstrate the effectiveness of our method, we applied it to predict evapotranspiration, a hydrologic variable that represents the water moved from the land surface to the atmosphere through a combination of evaporation and plant water use (transpiration). We also used a random sampling technique (the Monte Carlo method) to compare the uncertainty in the predictions when using the raw and noisy data versus the reconstructed data. The denoising process produced more accurate predictions of evapotranspiration with less uncertainty. Improved predictions of evapotranspiration can lead to a better understanding and accounting of water budgets. This ML approach is broadly suitable for a wide variety of applications that involve noisy sensor data with missing values. Key Points We used a denoising autoencoder (DAE) neural network to reduce noise in meteorological and soil sensor observations by on average We used Monte Carlo sampling to estimate the bias and variance of all model outputs, including uncertainty sources from data and the model We attached the DAE component to a downstream neural network to predict ET with the variance reduced by , compared to that without the DAE

denoising autoencoder↗

DOE OSTI2019

Gut anatomical properties and microbial functional assembly promote lignocellulose deconstruction and colony subsistence of a wood-feeding beetle

Beneficial microbial associations enhance the fitness of most living organisms, and wood-feeding insects offer some of the most striking examples of this. Odontotaenius disjunctus is a wood-feeding beetle that possesses a digestive tract with four main compartments, each of which contains well-differentiated microbial populations, suggesting that anatomical properties and separation of these compartments may enhance energy extraction from woody biomass. In this work, using integrated chemical analyses, we demonstrate that lignocellulose deconstruction and fermentation occur sequentially across compartments, and that selection for microbial groups and their metabolic pathways is facilitated by gut anatomical features. Metaproteogenomics showed that higher oxygen concentration in the midgut drives lignocellulose depolymerization, while a thicker gut wall in the anterior hindgut reduces oxygen diffusion and favours hydrogen accumulation, facilitating fermentation, homoacetogenesis and nitrogen fixation. We demonstrate that depolymerization continues in the posterior hindgut, and that the beetle excretes an energy- and nutrient-rich product on which its offspring subsist and develop. Our results show that the establishment of beneficial microbial partners within a host requires both the acquisition of the microorganisms and the formation of specific habitats within the host to promote key microbial metabolic functions. Together, gut anatomical properties and microbial functional assembly enable lignocellulose deconstruction and colony subsistence on an extremely nutrient-poor diet.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2025

Anomalously slow hot carrier cooling and insulator-to-metal transition in a photo-doped Mott insulator

Photo-doped Mott insulators can exhibit novel photocarrier transport and relaxation dynamics and non-equilibrium phases. However, time-resolved real-space imaging of these processes are still lacking. Here, we use scanning ultrafast electron microscopy (SUEM) to directly visualize the spatial-temporal evolution of photoexcited species in a spin-orbit assisted Mott insulator α-RuCl 3 . At low optical fluences, we observe extremely long hot photocarrier transport time over one nanosecond, almost an order of magnitude longer than any known values in conventional semiconductors. At higher optical fluences, we observe nonlinear features suggesting a photo-induced insulator-to-metal transition, which is unusual in a large-gap Mott insulator. Our results demonstrate the rich physics in a photo-doped Mott insulator that can be extracted from spatial-temporal imaging and showcase the capability of SUEM to sensitively probe photoexcitations in strongly correlated electron systems.

42 ENGINEERING↗

DOE OSTI2023

Front-End Engineering and Design: Project Tundra Carbon Capture System (Final Report)

As the next phase of Project Tundra, Minnkota Power Cooperative (MPC) completed a front-end engineering and design (FEED) study to install a post-combustion CO 2 capture system (CCS) at the MPC operated (Square Butte Electric Cooperative-owned) Milton R. Young Station (MRYS) Unit 2 (MRY2), a 477-MW power plant fueled by North Dakota lignite. The project team completed the FEED with Fluor’s Econamine FG Plus℠ (EFG+) technology and is taking the next steps leading up to start of construction. Team members included FEED technical lead Fluor Enterprises (Fluor); owner’s engineer and balance of plant (BOP) engineer Burns & McDonnell (BMcD); leading carbon capture consultants (David Greeson Consulting, Hunt International Energy Services); environmental consultants (AECOM, Agora Environmental), other BOP engineering consultants (Golder Associates, Nels); cost-share funding agency, the North Dakota Industrial Commission (NDIC); and the Energy & Environmental Research Center (EERC). Project Tundra’s goal is to implement carbon capture, utilization, and storage (CCUS) in North Dakota, while preserving the use of lignite. Future options could lead to revitalizing legacy oil fields and creating a new CO 2 enhanced oil recovery (EOR) industry. The topic of this report is the FEED study for the carbon dioxide capture portion of Project Tundra. Building on the findings of a pre-FEED study for MRY2, the key deliverables contained in this FEED study are: a) design, costing, and performance data needed to commence project financing activities; b) engineering and material balances required to file for all project permits; and c) a final project schedule. Based on the results of the previous pre-FEED study, MPC and its team evaluated two options to provide the large amount of steam needed to operate the CCS: 1) installation of new natural gas package boilers and a new pipeline to supply natural gas to the facility and 2) extraction of steam from the existing MRYS steam turbine generator. The FEED study was performed using the package boilers option, as it was deemed to have the lowest technical risk, overall cost, and cost uncertainty. A key feature of the design, however, was the mixing the natural gas boilers’ flue gas with the MRY2 flue gas. This increased the size of the CCS system. This design also enabled the tie-in of Unit 1 flue gas for times when Unit 2 is offline. The CCS was designed to capture a combined 12,978 short tons per day (STPD) of CO 2 from the flue gases produced by the existing MRY Unit 2 along with flue gas produced by the new boiler package within the CCS. The 12,978 STPD CO 2 capture is achieved by normally processing 100% of the total available MRY Unit 2 flue gas (considering full load operation) and 100% of the boiler package flue gas, then removing 90% of the CO 2 available from the processed flue gas streams.

01 COAL, LIGNITE, AND PEAT↗

DOE OSTI2024

Double-Side Cooled 1.2kV, 300A SiC MOSFET Phase-leg Modules for 200 kW, > 100 kW/L Traction Inverters

The packaging of a double-side cooled 1.2 kV, 149 A SiC phase-leg modules has been reported in recent years for making 100 kW, 100 kW/L traction inverters. Each phase-leg module consists of two SiC MOSFETs, one per switch position. Six of the phase-leg modules are assembled into a segmented inverter configuration to meet the power and power density requirement. In this work, the layout of the phase-leg module was redesigned to include four of the SiC MOSFETs, two per switch position, with the aim of doubling the power to 200 kW and increasing the power density beyond 100 kW/L, but with only a 10.25% footprint increase. Key features of the packaging technology developed in the previous work were implemented in the current work, which include silver sintering for chip bonding and porous silver inter-posts for bonding device source pads to substrate. Parasitic extraction simulation showed that the four-chip module has a low parasitic inductance of 4.7 nH, like the two-chip module. Static characterization of the four-chip prototypes showed a low average on-resistance of 9 mΩ and a low average leakage current of 5 nA at Vds of 1.2 kV.

Zhang, Zichen↗

DOE OSTI2025

MARIE: A Python-Based Framework for Comprehensive Fuel Recycling Modeling

One of the most pressing challenges to the continued deployment of nuclear energy systems is in the ultimate management and disposition of discharged fuel assemblies. While reprocessing and recovery of valuable materials from UNF assemblies has been considered as part of an overall strategy for minimization of the volume of reactor-based wastes to be managed, the deployment of commercial-scale reprocessing facilities presents an enormous economic challenge. The MARIE software package has been developed as a means of confronting this challenge. Representing components of a generic fuel reprocessing operation as individual physical processes, MARIE is designed as a modular framework intended to allow for analysis and cost-optimization for a hypothetical reprocessing facility while realistically accounting for the physical characteristics of the used fuel source term, such as decay heat, activity, and radiation dose (informing corresponding shielding requirements). Capabilities supported by MARIE include head-end operations such as fuel shearing, voloxidation, and dissolution; generic solvent extraction operations informed by available open-literature data; a suite of unit operations intended to represent electrochemical processing of used fuel assemblies (i.e., oxide reduction, electrorefining, and electrowinning); and finally, accounting for both costs and physical features of discharged waste streams, which can be used to inform follow-on analyses such as the feasibility of deep-borehole disposal of HLW. This paper presents an overview of the MARIE software capabilities, including how individual unit operations are implemented to enable a larger-scale optimization of a hypothetical reprocessing operation on aspects such as cost and recovery of valuable materials.

Skutnik, Steve [ORNL] (ORCID:000000016441135X)↗

DOE OSTI2023

Model selection and signal extraction using Gaussian Process regression

We present a novel computational approach for extracting localized signals from smooth background distributions. We focus on datasets that can be naturally presented as binned integer counts, demonstrating our procedure on the CERN open dataset with the Higgs boson signature, from the ATLAS collaboration at the Large Hadron Collider. Our approach is based on Gaussian Process (GP) regression — a powerful and flexible machine learning technique which has allowed us to model the background without specifying its functional form explicitly and separately measure the background and signal contributions in a robust and reproducible manner. Unlike functional fits, our GP-regression-based approach does not need to be constantly updated as more data becomes available. We discuss how to select the GP kernel type, considering trade-offs between kernel complexity and its ability to capture the features of the background distribution. We show that our GP framework can be used to detect the Higgs boson resonance in the data with more statistical significance than a polynomial fit specifically tailored to the dataset. Finally, we use Markov Chain Monte Carlo (MCMC) sampling to confirm the statistical significance of the extracted Higgs signature.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2019

Factors that impact the outcomes in testicular germ cell tumors in low–middle-income countries

Germ cell tumors (GCTs) are one of the most common tumors in adolescents and young adults. There is paucity of data on GCT from low–middle-income countries (LMIC). The present study was conducted to assess the demographic features, clinical manifestations, pathology, and outcomes of GCT patients treated at our center. Patients with testicular GCT above the age of 18 years, treated at our center from 2001 to 2015 were included in the study. Data were extracted retrospectively from the case records. Event-free survival (EFS) and overall survival (OS) were calculated using the Kaplan–Meier method and the variables were compared using the log-rank test. The study included 421 patients among whom 128 (30%) had a histological diagnosis of seminoma and 293 (70%) had non-seminomatous germ cell tumor (NSGCT). Metastatic disease at presentation was observed in 83/128 (65%) with seminoma and 254/293 (87%) with NSGCT. According to the International Germ Cell Cancer Collaborative Group (IGCCCG) risk stratification for metastatic disease, good- and intermediate-risk seminoma were observed in 55/83 (66%) and 28/83 (34%) patients, respectively, and good-, intermediate-, and poor-risk NSGCT were observed in 82/254 (32%), 76/254 (30%), and 96/254 (38%) patients, respectively. The median follow-up was 32.3 months (range 0.03–200 months). The 3-year OS for the entire cohort was 80.3%. The 3-year OS for seminoma was 91.4%, and for NSGCT was 75.3%. Factors significantly associated with inferior EFS and OS on multivariate analysis included poor performance status, scrotal orchidectomy, carboplatin-based regimen, NSGCT histology, and treatment default. Patients with testicular GCT in India present in an advanced stage and higher IGCCCG risk compared to Western data. Factors unique to LMIC like treatment default, bulky disease, dose compromise, and scrotal orchidectomy have a negative impact on the outcome.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

DOE OSTI2024

Accelerating the Discovery of New, Single Phase High Entropy Ceramics via Active Learning

High-entropy ceramics have garnered interest due to their remarkable hardness, compressive strength, thermal stability, and fracture toughness; yet the discovery of new high-entropy ceramics (out of a tremendous number of possible elemental permutations) still largely requires costly, inefficient, trial-and-error experimental and computational approaches. The entropy forming ability (EFA) factor was recently proposed as a computational descriptor that positively correlates with the likelihood that a 5-metal high-entropy carbide (HECs) will form the desired single phase, homogeneous solid solution; however, discovery of new compositions is computationally expensive. If you consider 8 candidate metals, the HEC EFA approach uses 49 optimizations for each of the 56 unique 5-metal carbides, requiring a total of 2744 costly density functional theory calculations. Here, we describe an orders-of-magnitude more efficient active learning (AL) approach for identifying novel HECs. To begin, we compared numerous methods for generating composition-based feature vectors (e.g., magpie and mat2vec), deployed an ensemble of machine learning (ML) models to generate an average and distribution of predictions, and then utilized the distribution as an uncertainty. Here we then deployed an AL approach to extract new training data points where the ensemble of ML models predicted a high EFA value or was uncertain of the prediction. Our approach has the combined benefit of decreasing the amount of training data required to reach acceptable prediction qualities and biases the predictions toward identifying HECs with the desired high EFA values, which are tentatively correlated with the formation of single phase HECs. Using this approach, we increased the number of 5-metal carbides screened from 56 to 15,504, revealing 4 compositions with record-high EFA values that were previously unreported in the literature. Our AL framework is also generalizable and could be modified to rationally predict optimized candidate materials/combinations with a wide range of desired properties (e.g., mechanical stability, thermal conductivity).

36 MATERIALS SCIENCE↗