Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “evaluation datasets”

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 1,747 records · Page 97
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Quasi-stationary Planetary Scale Waves in Modern Climate Models

We evaluate climate model simulations of geopotential height at the largest space-time scales, considering the first Fourier components (in longitude) of long-term climatological means. Quasi-stationary planetary scale waves emerge from this analysis. Variations of these waves are important components of climate change, for example associated with regional drought. Systematic examination of their long-term climatology provides context for understanding their evolution in time. We compare four reanalysis datasets with “historical runs” from the latest version of the Coupled Model Intercomparison Project (CMIP) and find general agreement, but we also identify model errors outside the range of both observational uncertainty and the uncertainty implied by interannual variability. Errors in wave phase have possible implications for climate model projections of regional climate and of shifting weather patterns. One model in the set we examine, originally created nearly 20 years before the others, exhibits noticeably greater errors.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

PPI DataHub Project Data Package: S. elongatus PCC 7942 Circadian Control Bioproduction Transcriptomics (PB-DP3)

The purpose of this experiment was to evaluate how circadian clock regulation impacts carbon partitioning between storage, growth, and product synthesis in Synechococcus elongatus PCC 7942 in providing insights to strategies for enhanced bioproduction. Sample data was acquired using a Illumina HiSeq sequencer system and processed for RNA sequencing (RNA-Seq) expression analysis. Transcriptomic differential expression analysis revealed coordinated circadian clock-driven adjustment of the cell cycle and rewiring of energy and carbon metabolism. Processed RNA-Seq datasets are openly accessible from the PNNL DataHub project dataset download page and contain secondary processed RNA-seq results files and supporting metadata materials linked to relevant source code information supporting data transparency and reuse.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Automated Quantification of Wind Turbine Blade Leading Edge Erosion from Field Images

Wind turbine blade leading edge erosion is a major source of power production loss and early detection benefits optimization of repair strategies. Two machine learning (ML) models are developed and evaluated for automated quantification of the areal extent, morphology and nature (deep, shallow) of damage from field images. The supervised ML model employs convolutional neural networks (CNN) and learns features (specific types of damage) present in an annotated set of training images. The unsupervised approach aggregates pixel intensity thresholding with calculation of pixel-by-pixel shadow ratio (PTS) to independently identify features within images. The models are developed and tested using a dataset of 140 field images. The images sample across a range of blade orientation, aspect ratio, lighting and resolution. Each model (CNN v PTS) is applied to quantify the percent area of the visible blade that is damaged and classifies the damage into deep or shallow using only the images as input. Both models successfully identify approximately 65% of total damage area in the independent images, and both perform better at quantifying deep damage. The CNN is more successful at identifying shallow damage and exhibits better performance when applied to the images after they are preprocessed to a common blade orientation.

Aird, Jeanie A.↗
DOE OSTI2025

Application of PRIM for understanding patterns in carbon dioxide model-observation differences

Reducing uncertainties in regional carbon balances requires a better understanding of CO 2 transport in synoptic weather systems. Here, we apply the Patient Rule Induction Method (PRIM), a data-mining method to identify high-density regions for a target-class within an input parameter space, to airborne observations of potential temperature, wind speed, water vapor mixing ratio, and CO 2 dry mol fraction gathered during the Atmospheric Carbon and Transport (ACT)-America Summer 2016 and Winter 2017 campaigns. ACT observations were targeted at expert-designated cases of fair weather and near-frontal warm and cold sector air at atmospheric boundary-layer, lower-, and higher free tropospheric levels (ABL, LFT, and HFT, respectively). We investigate atmospheric characteristics of these pre-defined cases and associated CO 2 model-observation-differences in the mesoscale WRF-Chem model. PRIM results separate winter- and summertime observations as well as observations from ABL, LFT, and HFT with enrichment factors of 4.0–20.5 inside the PRIM box compared to the entire dataset but cannot distinguish between near-frontal warm and cold sector observations in the higher free troposphere. Analyzing of the parameter space constrained by PRIM, we find that large magnitude model observation differences preferentially associated with times when atmospheric conditions are less typical. This association suggests that PRIM could provide a useful tool for isolating atmospheric conditions with large-magnitude and non-Gaussian CO 2 -residuals for targeted transport model evaluation and to potentially improve inversion results during synoptically active periods.

Gerken, Tobias [James Madison Univ., Harrisonburg,↗
NASA NTRS2019

Assimilation of Remotely Sensed Leaf Area Index into the Noah-MP Land Surface Model: Impacts on Water and Carbon Fluxes and States over the Continental U.S.

Accurate representation of vegetation states is required for the modeling of terrestrial water-energy-carbon exchanges and the characterization of the impacts of natural and anthropogenic vegetation changes on the land surface. This study presents a comprehensive evaluation of the impact of assimilating remote sensing-based Leaf Area Index (LAI) retrievals over the Continental U.S. in the Noah-MP land surface model, during a time period of 2000 to 2017. The results demonstrate that the assimilation has a beneficial impact on the simulation of key water budget terms such as soil moisture, evapotranspiration, snow depth, terrestrial water storage and streamflow, when compared with a large suite of reference datasets. In addition, the assimilation of LAI is also found to improve the carbon fluxes of Gross Primary Production (GPP) and Net Ecosystem Exchange (NEE). Most prominent improvements in the water and carbon variables are observed over the agricultural areas of the U.S., where assimilation improves the representation of vegetation seasonality impacted by cropping schedules. The systematic, added improvements from assimilation in a configuration that employs high quality boundary conditions highlight the significant utility of LAI data assimilation in capturing the impacts of vegetation changes.

anthropogenic vegetation↗
DOE OSTI2025

The Profiled Feldman-Cousins Method for Confidence Interval Construction for the Nova 3-Flavor Oscillation Analysis

The small interaction cross-section of neutrinos makes experimental neutrino physics particularly responsive to technological advancements. A significant development leveraged by the NOvA experiment is large-scale parallel processing, enabling novel computational approaches to longstanding experimental challenges. Central to managing the resulting high-throughput data is NOvA’s implementation of the Freight Train model, designed for efficient data production and handling.This dissertation details the methodology and execution of the NOvA 2024 3-Flavor Oscillation Analysis, supported by a comprehensive dataset spanning ten years. It emphasizes frequentist results refined through the Feldman-Cousins (FC) technique, specifically addressing confidence interval corrections in parameter estimation. The computational intensity associated with Feldman-Cousins arises from extensive Monte Carlo simulations, which were substantially mitigated through parallel computing on the Perlmutter supercomputer at the National Energy Research Scientific Computing Center (NERSC), employing the MPI framework.To further enhance computational efficiency, an Importance Sampling method is introduced and evaluated, demonstrating significant potential to reduce complexity, particularly in exploring extreme parameter space regions. This thesis presents both the successful application of advanced computational resources and the development of sophisticated statistical techniques, aiming to enhance the precision and scope of neutrino oscillation analyses.

Dye ajdye11190@gmail.com, Andrew Joseph [Mississip↗
DOE OSTI2024

Evaluating the contribution of plant metabolic pathways in the light to the ATP:NADPH demand using a meta-analysis of isotopically non-stationary metabolic flux analyses

Abstract Balancing the ATP: NADPH demand from plant metabolism with supply from photosynthesis is essential for preventing photodamage and operating efficiently, so understanding its drivers is important for integrating metabolism with the light reactions of photosynthesis and for bioengineering efforts that may radically change this demand. It is often assumed that the C3 cycle and photorespiration consume the largest amount of ATP and reductant in illuminated leaves and as a result mostly determine the ATP: NADPH demand. However, the quantitative extent to which other energy consuming metabolic processes contribute in large ways to overall ATP: NADPH demand remains unknown. Here, we used the metabolic flux networks of numerous recently published isotopically non-stationary metabolic flux analyses (INST-MFA) to evaluate flux through the C3 cycle, photorespiration, the oxidative pentose phosphate pathway, the tricarboxylic acid cycle, and starch/sucrose synthesis and characterize broad trends in the demand of energy across different pathways and compartments as well as in the overall ATP:NADPH demand. These data sets include a variety of species including Arabidopsis thaliana , Nicotiana tabacum , and Camelina sativa as well as varying environmental factors including high/low light, day length, and photorespiratory levels. Examining these datasets in aggregate reveals that ultimately the bulk of the energy flux occurred in the C3 cycle and photorespiration, however, the energy demand from these pathways did not determine the ATP: NADPH demand alone. Instead, a notable contribution was revealed from starch and sucrose synthesis which might counterbalance photorespiratory demand and result in fewer adjustments in mechanisms which balance the ATP deficit.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Impacts of urbanization, antecedent rainfall event, and cyclone tracks on extreme floods at Houston reservoirs during Hurricane Harvey

The objective of this study is to evaluate the effects of urbanization, antecedent rainfall event (ARE), and varying cyclone tracks on the streamflow—and thus the subsequent reservoir status—during the floods caused by Hurricane Harvey in August-September 2017. Through a hydrological modeling approach, we examined how these factors influenced the inflows, peak pool elevations and outflows of the two most important detention reservoirs in the Houston region, the Addicks and Barker Reservoirs. A high-resolution rainfall reanalysis dataset for extreme storm events, along with a suite of synthetic rainfall values from a variety of storm tracks, were adopted to represent both the truth and the maximum possible rainfall during the Hurricane Harvey period. Results showed the following: Urbanization only led to slight increases in peak inflows, not necessarily leading to an increase in peak pool elevations. ARE contributed to the peak inflow and pool elevation slightly. In contrast, if the cyclone had followed the most adverse track consistent with earlier forecasts (all else being equal), the total volumetric flow into the two reservoirs could have been significantly larger (37.00% and 48.70% respectively), thus increasing the peak pool elevations by 1.06 and 1.37 meters respectively. These results suggest that large uncertainties exist for flood management at a watershed scale during hurricanes, because of the uncertainties with regard to the cyclone track. This is true even if storm-relative precipitation rates could be predicted perfectly.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2021

Omics-Lethal Human Viruses, Ebola Experiment EH001

The purpose of this experiment was to evaluate the human patient peripheral blood mononuclear cells (PBMC) response to Zaire Ebola Makona virus infection during the 2013-2016 epidemic in West Africa. Samples were obtained from whole blood collections from human patients in 2015 who were naturally infected with Ebola virus during the West African Ebola virus epidemic and healthy individuals (6 month collections) where resulting blood serum was processed for proteome, metabolome, and lipidome expression analysis. Secondary host-associated viral dataset downloads contain one or more statistically processed (normalization data transformation) quantitative dataset collections resulting in qualitative expression analyses of primary host-pathogen experimental study designs. Leveraging unique high-resolution Omics capabilities for proteomics, metabolomics, and lipidomics dataset downloads each have a direct relationship to a primary sample submission corresponding to a specific Ebola virus infection.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Simple Diesel Train Fuel Consumption Model for Real-Time Train Applications

This paper introduces a simple diesel train energy consumption model that calculates the instantaneous energy consumption using vehicle operational input variables, including the instantaneous speed, acceleration, and roadway grade, which can be easily obtained from global positioning system (GPS) loggers. The model was tested against real-world data and produced an error of -1.33% for all data and errors ranging from -12.4% to +8.0% for energy consumption of four train datasets amounting to a total of 5854 km trips. The study also validated the proposed model with separate data that were collected between Valencia and Cuenca, Spain, which had a total length of 198 km and found that the model was accurate, yielding a relative error of -1.55% for the total energy consumption. These results show that the proposed model can be used by train operators, transportation planners, policy makers, and environmental engineers to evaluate the energy consumption effects of train operational projects and train simulation within intermodal transportation planning tools.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
NASA NTRSDate not supplied

Root Cause Analysis of the Data Refinement Process – Medical Conditions Capability Resource Tables

The medical system for spaceflight thus far has been designed to support missions in low earth orbit (LEO). Crew capabilities are limited and heavily dependent on the team of medical support staff at Mission Control Center (MCC) to guide diagnosis and management. However, missions to the Moon and Mars will suffer from several constraints that will make this ground support focused approach to care ineffective. In order to update and modify medical system design, NASA has relied on Probabilistic Risk Assessment (PRA) modeling to mitigate medical risk through trade space analysis. Specifically, capability resource tables (CRT’s) were developed to create a dataset of resources required to manage a list of accepted medical conditions significant in exploration spaceflight. With 120 conditions, this dataset contained hundreds of capabilities and thousands of resources with tens of thousands of cells of data. Initially these tables were built in excel for high throughput during development, but ultimately had to be transferred, managed, and modified into the Evidence Library database for modeling purposes. The process of collating and reviewing the Evidence Library revealed numerous errors in the dataset that had to be corrected through iterative changes. Several error types emerged during this process and can be broken into specific classifications defined as “input”, “transcription”, “structural”, “branching”, and “information”. In reviewing these error types through the root cause analysis (RCA) approach, we were able to identify the contributors to these errors which included single data review points, changing product end goals, limited software selection, time constraints and several others. By reviewing and evaluating the underlying causes we can provide possible system improvements that can be implemented for current and future data management in PRA model inputs.

A. Anderson↗
NASA NTRS2024

Asheville Urban Development II: Mapping Urban Heat to Support Cooling Initiatives and Climate Resilience Planning in the Greater Asheville Area

Asheville, North Carolina experiences the urban heat island effect, where temperatures in the city are higher than in surrounding rural areas. This effect intensifies with increased urbanization and less vegetative cover. Asheville’s urban heat island was exacerbated by population increases and tree cover decline, escalating the need for heat mitigation. We partnered with the City of Asheville’s Sustainability Department and Asheville GreenWorks whose actions prioritize sustainable city planning and equitable climate resilience. Using NASA Earth observations and ancillary datasets we spatially mapped urban heat, heat vulnerability, and cooling and adaptive capacity from 2019-2023. To map urban heat, we used Landsat 8 and 9 Operational Land Imager and Thermal Infrared Sensor for land surface temperature and albedo data and the ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station for evapotranspiration data. We assessed heat vulnerability using the urban heat data andthe Centers for Disease Control and Prevention’s Social Vulnerability Index. To evaluate cooling and adaptive capacity we used the InVEST Urban Cooling Model, integrating our heat vulnerability analysis with land use and cover data from Sentinel-1 Synthetic Aperture Rada rand Sentinel-2 Multispectral Instrument. Our results revealed distinct spatial patterns of urban heat, heat vulnerability, and cooling and adaptive capacity in Asheville with downtown as the focal hotspot and an outward decreasing radial pattern. These findings highlight targeted need for interventions to reduce heat impacts, address environmental injustices, and enhance climate resilience. Our project provided research to local organizations that can be used for heat mitigation in the greater Asheville area.

Authors not in NED but are confirmed contractors. ↗
NASA NTRSDate not supplied

Quantum Leap: Evaluating the Feasibility of Quantum Machine Learning Using NASA Earth Observational Data

This study explores the feasibility of leveraging quantum machine learning (QML) to analyze NASA Earth Observational (EO) data for climate change research, with a particular focus on the phenomenon of ”crop frosting” which has become more prevalent due to climate change. We implemented and evaluated two QML models, the Variational Quantum Classifier (VQC) and Quantum Support Vector Classifier (QSVC), in both simulated and real quantum computing environments using a 127 qubit IBM quantum processor. Our study emphasizes the scientific rigor in comparing these quantum models with a classical Support Vector Machine (SVM) classifier, highlighting their performance in processing climate data. The results offer valuable insights into the potential scientific advantages, limitations, and scalability of QML for analyzing EO datasets, thus paving the way for more advanced climate modeling and predictive analytics using quantum computing. We showcased how Environmental Interaction Knowledge Graphs (EIKGs) and Digital Twins (DTs) can be integrated into this study. This research underscores the transformative potential of Classical and QML leveraging KGs and DT to address the multifaceted challenges posed by climate change.

Quantum Computing↗
NASA NTRS2016

Combining Livestock Production Information in a Process-Based Vegetation Model to Reconstruct the History of Grassland Management

Grassland management type (grazed or mown) and intensity (intensive or extensive) play a crucial role in the greenhouse gas balance and surface energy budget of this biome, both at field scale and at large spatial scale. However, global gridded historical information on grassland management intensity is not available. Combining modelled grass-biomass productivity with statistics of the grass-biomass demand by livestock, we reconstruct gridded maps of grassland management intensity from 1901 to 2012. These maps include the minimum area of managed vs. maximum area of unmanaged grasslands and the fraction of mown vs. grazed area at a resolution of 0.5deg by 0.5deg. The grass-biomass demand is derived from a livestock dataset for 2000, extended to cover the period 19012012. The grass-biomass supply (i.e. forage grass from mown grassland and biomass grazed) is simulated by the process-based model ORCHIDEE-GM driven by historical climate change, risingCO2 concentration, and changes in nitrogen fertilization. The global area of managed grassland obtained in this study increases from 6.1 x 10(exp 6) km(exp 2) in 1901 to 12.3 x 10(exp 6) kmI(exp 2) in 2000, although the expansion pathway varies between different regions. ORCHIDEE-GM also simulated augmentation in global mean productivity and herbage-use efficiency over managed grassland during the 20th century, indicating a general intensification of grassland management at global scale but with regional differences. The gridded grassland management intensity maps are model dependent because they depend on modelled productivity. Thus specific attention was given to the evaluation of modelled productivity against a series of observations from site-level net primary productivity (NPP) measurements to two global satellite products of gross primary productivity (GPP) (MODIS-GPP and SIF data). Generally, ORCHIDEE-GM captures the spatial pattern, seasonal cycle, and inter-annual variability of grassland productivity at global scale well and thus is appropriate for global applications presented here.

Livestock dataset↗
DOE OSTI2020

Technology Summary L-21044 Tunnel Sensor Hydrogeology Evaluations

Illicit tunneling along the U.S. Mexican border is a persistent national security problem. Several types of sensors could be to used to detect tunnels, but the hydrogeology of the subsurface makes some technologies ineffective. This effort used available borehole and surface geology data to develop 3D hydrogeologic models, which could then provide the input for optimizing subsurface sensor technologies to detect tunnels. Depending on the specific type of sensor, the following physical properties of the subsurface are important: longitudinal speed (velocity-p), shear speed (velocity-s), acoustic attenuation, electrical permittivity, electrical conductivity, density and elastic properties. LLNL developed a procedure which uses hydrogeological datasets to determine 3D hydrogeology cross sections. The physical properties of the rocks are generally not known precisely, thus ranges of properties would need to be considered. This cross-section data can then be utilized to enhance modeling studies to optimize sensor types/systems.

58 GEOSCIENCES↗
DOE OSTI2026

Advancing Asset Management in Water Infrastructure Systems

Aging water system infrastructure, including drinking water, wastewater, and stormwater, poses a growing challenge for utilities and municipalities. These water systems have well documented challenges with respect to their age, condition, and level of service. ASCE annual report cards consistently rate these infrastructure systems in the United States as underfunded, overcapacity, or past service life (ASCE 2025 Report Card). For example, Chini and Stillwell (2017) estimated that the mean water loss in drinking water systems, i.e., non-revenue water, is approximately 16% across the United States. These concerns are not just relegated to the United States, with Courtenay, British Columbia, identifying 17% of their water main pipes as in a ‘poor’ condition state, defined as a category condition 5 out of 5 (City of Courtenay, 2024). These cases illustrate the challenges utilities are facing to manage extensive networks of infrastructure to deliver a consistent and high level of service. For buried infrastructure such as water systems, studies suggest that preventative interventions can lead to lower maintenance costs and fewer service disruptions (Mazumder et al, 2018; Li et al, 2014). The demonstrated need and benefit of appropriately applied asset management is juxtaposed against the relatively sparse literature that evaluates water systems within an asset management construct. Since 2020, just 37 papers specifically reference asset management in the Journal of Water Resources Planning and Management. Of those, only a few specifically look to develop strategies for improved asset management. Therefore, we highlight four key research areas that represent opportunities for advancement of asset management research for water systems. First, advances in condition assessment and forecasting are needed to better estimate asset deterioration using diverse datasets. Second, machine learning (ML) and artificial intelligence (AI) hold promise for predictive maintenance and investment prioritization, though questions of generalizability and model transparency remain. Third, applying a value of information framework can guide utilities in making cost-effective sensor deployment and data collection decisions, to direct monitoring strategies towards data-informed asset management decisions. Finally, integrated infrastructure management is critical, requiring coordinated planning with other infrastructure systems and stakeholder engagement to reduce costs and enhance service delivery.

Chini, Christopher M.↗
DOE OSTI2026

Flavor, transverse momentum, and azimuthal dependence of charged pion multiplicities in semi-inclusive deep-inelastic scattering with 10.6 GeV electrons

Measurements of semi-inclusive deep-inelastic scattering multiplicities for 𝜋 + and 𝜋 − from proton and deuteron targets are reported on a grid of hadron kinematic variables 𝑧, 𝑃 𝑇 , and 𝜙* for leptonic kinematic variables in the range 0.3<𝑥<0.6 and 3<𝑄 2 <5GeV 2 . Data were acquired in 2018 and 2019 at Jefferson Lab Hall C with a 10.6 GeV electron beam impinging on 10-cm-long liquid hydrogen and deuterium targets. Scattered electrons and charged pions were detected in the High Momentum Spectrometer and Super High Momentum Spectrometer, respectively. The multiplicities were fitted for each bin in (𝑥,𝑄 2 ,𝑧,𝑃𝑡) to extract the 𝜙*—independent 𝑀 0 and the azimuthal modulations ⟨cos⁡(𝜙*)⟩ and ⟨cos⁡(2⁢𝜙*)⟩. The 𝑃𝑡 dependence of the 𝑀 0 results was found to be remarkably consistent for the four cases studied: 𝑒⁢𝑝→𝑒⁢𝜋 + ⁢𝑋, 𝑒⁢𝑝→𝑒⁢𝜋 − ⁢𝑋, 𝑒⁢𝑑→𝑒⁢𝜋 + ⁢𝑋, 𝑒⁢𝑑→𝑒⁢𝜋 − ⁢𝑋 over the range 0GeV<𝑃 𝑡 <0.4GeV, as were the multiplicities evaluated near 𝜙*=180∘ over the extended range 0GeV<𝑃 𝑡 <0.7GeV. The Gaussian widths of the 𝑃 𝑡 dependence exhibit a quadratic increase with 𝑧. The cos⁡(𝜙*) modulations were found to be consistent with zero for 𝜋 + , in agreement with previous world data, while the 𝜋 − moments were, in many cases, significantly greater than zero. The cos⁡(2⁢𝜙*) modulations were found to be consistent with zero. The higher statistical precision of this dataset of about 20 000 individual multiplicity values, compared with previously published data, should allow improved determinations of quark transverse momentum distributions and higher twist contributions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2026

When does global attention help: a unified empirical study on atomistic graph learning

Graph neural networks (GNNs) are widely used as surrogates for costly experiments and first-principles simulations to study the behavior of compounds at atomistic scale, and their architectural complexity is constantly increasing to enable the modeling of complex physics. While most recent GNNs combine more traditional message passing neural networks (MPNNs) layers to model short-range interactions with more advanced graph transformers (GTs) with global attention mechanisms to model long-range interactions, it is still unclear when global attention mechanisms provide real benefits over well-tuned MPNN layers due to inconsistent implementations, features, or hyperparameter tuning. We introduce the first unified, reproducible benchmarking framework–built on HydraGNN–that enables seamless switching among four controlled model classes: MPNN, MPNN with chemistry/topology encoders, GPS-style hybrids of MPNN with global attention, and fully fused localglobal models with encoders. Using seven diverse open-source datasets for benchmarking across regression and classification tasks, we systematically isolate the contributions of message passing, global attention, and encoder-based feature augmentation. Our study shows that encoder-augmented MPNNs form a robust baseline, while fused localglobal models yield the clearest benefits for properties governed by long-range interaction effects. We further quantify the accuracycompute trade-offs of attention, reporting its overhead in memory. Together, these results establish the first controlled evaluation of global attention in atomistic graph learning and provide a reproducible testbed for future model development.

Equivariant graph neural networks↗