SEARCH · Search NASA

Results for “association prediction”

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 181 records · Page 10

All sources NASA NTRS DOE OSTI

DOE OSTI2020

Phase Switching as the Origin of Large Piezoelectric Response in Organic-Inorganic Perovskites: A First-Principles Study

Piezoelectrics are critical functional components of many practical applications such as sensors, ultrasonic transducers, actuators, medical imaging, telecommunications. So far, the best performing piezoelectrics are ferroelectric ceramics, many of which are toxic, heavy, hard, and cost-ineffective. Recently, a groundbreaking discovery of extraordinarily large piezoelectric coefficients in the family of organic-inorganic perovskites gave a hope for a cheaper, environmentally friendly, inexpensive, light-weight and flexible alternative. However, the origin of such response in organic-inorganic ferroelectrics whose spontaneous polarization is an order of magnitude smaller than for inorganic counterparts remains unclear. In this study we employ first-principles simulations to predict that the mechanism associated with large piezoelectric constants is of extrinsic origin and associated with switching between the stable phase and previously overlooked energetically competitive metastable phase that can be stabilized by the external stress. Here, the phase switching changes the polarization direction and, therefore, gives origin to large piezoelectric response, similar to PbZr$_{1-x}$Ti$_x$O$_3$ near morphotropic phase boundary. Existence of such metastable phases is likely to manifest as the dynamical molecular disorder above the Curie temperature and, therefore, could be intrinsic to the entire family of organic-inorganic ferroelectrics with such disorder.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

DOE OSTI2021

Coupling of regional geophysics and local soil-structure models in the EQSIM fault-to-structure earthquake simulation framework

Accurate understanding and quantification of the risk to critical infrastructure posed by future large earthquakes continues to be a very challenging problem. Earthquake phenomena are quite complex and traditional approaches to predicting ground motions for future earthquake events have historically been empirically based whereby measured ground motion data from historical earthquakes are homogenized into a common data set and the ground motions for future postulated earthquakes are probabilistically derived based on the historical observations. This procedure has recognized significant limitations, principally due to the fact that earthquake ground motions tend to be dictated by the particular earthquake fault rupture and geologic conditions at a given site and are thus very site-specific. Historical earthquakes recorded at different locations are often only marginally representative. There has been strong and increasing interest in utilizing large-scale, physics-based regional simulations to advance the ability to accurately predict ground motions and associated infrastructure response. However, the computational requirements for simulations at frequencies of engineering interest have proven a major barrier to employing regional scale simulations. In a U.S. Department of Energy Exascale Computing Initiative project, the EQSIM application development is underway to create a framework for fault-to-structure simulations. This framework is being prepared to exploit emerging exascale platforms in order to overcome computational limitations. This article presents the essential methodology and computational workflow employed in EQSIM to couple regional-scale geophysics models with local soil-structure models to achieve a fully integrated, complete fault-to-structure simulation framework. Here, the computational workflow, accuracy and performance of the coupling methodology are illustrated through example fault-to-structure simulations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Preliminary Enhanced LAW Glass Formulation Algorithm

This report summarizes the Preliminary Enhanced LAW Glass Formulation Algorithm, its background information, and the calculations it performs. The Preliminary Enhanced LAW Glass Formulation Algorithm is a tool that has been developed in MATLAB to formulate glass at a given a waste composition, while attempting to maximize waste loading. It is intended for use at the Hanford Waste Treatment and Immobilization Plant, where nuclear waste will be vitrified into glass for safe, long-term storage. The formulated glass is required to meet several processing and product quality constraints. In addition, calculations must account for associated uncertainties in constraint prediction and measurement. The algorithm also allows for the ability to hold glass compositions closer to tested regions in order to increase confidence in the output. The algorithm must also follow nuclear quality assurance procedures and submit to rigorous validation and verification testing.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2021

TAFI/Kebab End of Project Report

This report focuses on the two primary goals set forth in Sandia’s TAFI effort, referred to here under the name Kebab. The first goal is to overlay a trajectory onto a large database of historical trajectories, all with very different sampling rates than the original track. We demonstrate a fast method to accomplish this, even for databases that hold over a million tracks. The second goal is to then demonstrate that these matched historical trajectories can be used to make predictions about unknown qualities associated with the original trajectory. As part of this work, we also examine the problem of defining the qualities of a trajectory in a reproducible way.

42 ENGINEERING↗

DOE OSTI2022

Preliminary Enhanced LAW Glass Formulation Algorithm

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2023

Preliminary Enhanced LAW Glass Formulation Algorithm

This report summarizes the Preliminary Enhanced Low-Activity Waste Glass Formulation Algorithm (GFA), its background information, and the calculations it performs. The Preliminary Enhanced Low-Activity Waste GFA is a tool developed in MATLAB to formulate glass for a given waste composition while attempting to maximize waste loading. It is intended for use at the Waste Treatment and Immobilization Plant, where nuclear waste will be vitrified into glass. The formulated glass is required to satisfy several processing and product quality constraints. In addition, calculations must account for associated uncertainties in constraint prediction and measurement. The GFA adheres to Pacific Northwest National Laboratory nuclear quality assurance procedures and has been validated and verified. In this second revision to the report, two constraints: the lithium and zirconia target constraints were amended to Table 3.1 and programmed into the algorithm. Additionally, three sections were added. Section 3.2.1 includes the Enhanced LAW Correlation Rules, an addition to the glass optimality criteria. Section 4.0 discusses a stepped process approach taken to ensure the algorithm can adequately react to differences between its recommended output and the actual measurements and transfers made during plant operation, and calculation of specific radionuclide concentrations. Section 5.2 describes changes to the user interaction with the algorithm program including the addition of a graphical user interface and compilation of the software.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2025

Advancements in NEAMS Tool Capabilities for Multiphysics Simulation of Fast Reactor Core Bowing and Identification of Validation Test Data

Under the U.S. Department of Energy Office of Nuclear Energy Advanced Modeling and Simulation (NEAMS) Program, an integrated multiphysics approach is being developed to model the core bowing phenomena important to liquid metal-cooled fast reactors. Core bowing is an important passive safety mechanism in liquid metal-cooled fast reactors and involves Multiphysics effects including radiation transport, fluid flow, heat transfer, and mechanical response to temperature and flux gradients. This report summarizes recent progress on developing a multiphysics, MOOSE-based workflow to predict core bowing and associated reactivity feedback. Last year, thermal fluids and mechanics were coupled on a multi-assembly benchmark problem based on ABR-1000 design. This year, the reactor physics code Griffin was assessed for readiness of core bowing calculations. Preliminary integration of Griffin’s ring-heterogeneous model with thermal fluids and thermal mechanics solvers was performed. Specifically, thermal-mechanics and reactor physics were coupled for single- and multi-assembly problems, and reactor physics and subchannel methods were coupled for a single assembly model. Finally, the workflow of all three physics was preliminarily demonstrated on a single assembly model. Caveats and future development needed have been identified. To supplement the multiphysics demonstration, verification and assessment efforts of thermos-mechanical capabilities for modeling thermo-mechanical core bowing behavior were continued by analyzing IAEA Verification Problem 5 which includes radiation swelling and creep. Additionally, a small core reactor physics benchmark defined by Japan Atomic Energy Agency (JAEA) was performed to assess neutronics models for estimating reactivity feedback. Finally, Fast Flux Test Facility (FFTF) validation test data for core bowing phenomena has been identified and summarized, with a recommended path forward for validation once this capability is mature.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2025

Vibrational signatures of S Vacancies in monolayer tungsten disulfide

S vacancies in WS 2 affect material performance but a difficult to detect. In this work I use theoretical simulations to predict the vibrational frequencies associated with S vacancies in monolayer WS 2 . I report the phonon spectrum and band structure of the isolated monolayer and the frequencies that are introduced to the spectrum when a vacancy is created.

36 MATERIALS SCIENCE↗

DOE OSTI2025

A LENS on DUNE-PRISM: Characterizing a Neutrino Beam with Off-Axis Measurements

Upcoming precision long-baseline neutrino oscillation experiments will be severely limited by the large systematic uncertainties associated with neutrino flux predictions and neutrino--nucleus cross sections. A promising remedy is the PRISM (Precision Reaction Independent Spectrum Measurement) technique, whereby the near detector measures the neutrino spectrum at different angles with respect to the beam axis. These measurements are then linearly combined into a prediction of the oscillated neutrino flux at the far detector. This prediction is data-driven, but still dependent on some theoretical knowledge about the neutrino flux. In this paper, we study to what extent off-axis measurements themselves can be used to directly constrain neutrino flux models. In particular, we use them to extract separately the fluxes and spectra of different meson species in the beam. We call this measurement LENS (Lateral Extraction of Neutrino Spectra). Second, we demonstrate how the thus improved flux model helps to further constrain the far detector flux prediction, thereby ultimately improving oscillation measurements.

FOS: Physical sciences↗

DOE OSTI2025

Knowledge Oriented Graph Unified Transformer (KOGUT) v0.1

KOGUT — Knowledge Oriented Graph Unified Transformer KOGUT implements the Relational Graph Transformer (RelGT) architecture for knowledge graph link prediction in biological domains, with a primary focus on microbial growth media prediction. While the original RelGT (arXiv:2505.10960) targets relational tables, time series, and multi-table databases, KOGUT adapts this architecture for heterogeneous biological knowledge graphs, providing first-in-class AI predictive models for microbial cultivation. Key Adaptations Beyond Original RelGT: - Knowledge Graph Focus: Applied to biological KGs with semantic node types (taxa, chemicals, media, phenotypes, environments) versus generic relational database tables, trained on the KG-Microbe knowledge graph (1.3M entities, 2.9M edges, 24 relation types). - Multimodal Node Encoding: Integrates node labels, categories, descriptions, and synonyms from KG metadata through learned embedding layers—adapting relational column features to graph node attributes with textual semantics. - Extended K-Hop Subgraph Strategy: Optimized neighborhood sampling (3-hop default, configurable up to 200 nodes) tuned for sparse biological networks, building on the original local-global attention framework with biological relation preservation. - Biolink Predicate Preservation: Type-specific transformations for 24 biological edge semantics (occurs_in, consumes, produces, has_phenotype, subclass_of) beyond standard relational foreign keys, enabling multi-relation link prediction. - Inductive Learning Support: Enables zero-shot predictions for novel taxa through feature-based embeddings (temperature, oxygen requirements, gram stain, cell shape), extending the original transductive relational benchmark scope to uncultured microorganisms. CheapSOTA Performance Optimizations (This Distribution): - VQ-EMA Centroid Attention: Vector quantization with exponential moving average for improved global context modeling (+5-10% MRR improvement). - HDF5 Precomputed Data Loading: One-time preprocessing of k-hop subgraphs to eliminate redundant graph traversals (2-5× training speedup). - Distributed Data Parallel Training: Multi-GPU support for scaling to larger knowledge graphs (tested on 4× NVIDIA A100 GPUs at NERSC Perlmutter). - Mixed Precision Training: Automatic mixed precision (AMP) for memory efficiency and faster training. Advantages Over Standard Knowledge Graph Embedding Models: Combines RelGT's proven multi-element tokenization (features, type, hop, structure) with graph-native biological representations, enabling interpretable link prediction across heterogeneous entities that standard embedding models (TransE, RotatE, ComplEx) and table-based transformers cannot directly model. Achieves near-perfect performance on microbial growth media prediction (MRR: 0.9966, Precision@1: 0.9932, Hit@10: 1.0000) while maintaining explainability through attention-based reasoning over biological pathways. Training Data: - KG-Microbe merged knowledge graph: 1,379,337 nodes, 2,960,472 edges - 24 biological relation types including taxonomic hierarchies, metabolic interactions, phenotype associations, and environmental relationships - Primary prediction task: Growth media suitability for microbial taxa (biolink:occurs_in, 50K edges) - Multi-relation capability: Predicts links for any of the 24 relation types, including chemical consumption/production, phenotype associations, and taxonomic classification Citation: Original RelGT Architecture: Dwivedi et al., "Relational Graph Transformer", arXiv:2505.10960, 2025 KOGUT Implementation: Knowledge Oriented Graph Unified Transformer for Microbial Growth Media Prediction Developed at Lawrence Berkeley National Laboratory (LBNL) Trained on NERSC Perlmutter supercomputer

Joachimiak, Marcin [Lawrence Berkeley National Lab↗

DOE OSTI2023

Mechanisms and Impacts of Earth System Tipping Elements

Tipping elements are components of the Earth system which may respond nonlinearly to anthropogenic climate change by transitioning toward substantially different long-term states upon passing key thresholds or “tipping points.” In some cases, such changes could produce additional greenhouse gas emissions or radiative forcing that could compound global warming. Improved understanding of tipping elements is important for predicting future climate risks and their impacts. Here we review mechanisms, predictions, impacts, and knowledge gaps associated with 10 notable Earth system components proposed to be tipping elements. We evaluate which tipping elements are approaching critical thresholds and whether shifts may manifest rapidly or over longer timescales. Some tipping elements have a higher risk of crossing tipping points under middle-of-the-road emissions pathways and will possibly affect major ecosystems, climate patterns, and/or carbon cycling within the 21st century. However, literature assessing different emissions scenarios indicates a strong potential to reduce impacts associated with many tipping elements through climate change mitigation. The studies synthesized in our review suggest most tipping elements do not possess the potential for abrupt future change within years, and some proposed tipping elements may not exhibit tipping behavior, rather responding more predictably and directly to the magnitude of forcing. Nevertheless, uncertainties remain associated with many tipping elements, highlighting an acute need for further research and modeling to better constrain risks.

58 GEOSCIENCES↗

DOE OSTI2022

Incorporating polar field data for improved solar flare prediction

In this paper, we consider incorporating data associated with the sun’s north and south polar field strengths to improve solar flare prediction performance using machine learning models. When used to supplement local data from active regions on the photospheric magnetic field of the sun, the polar field data provides global information to the predictor. While such global features have been previously proposed for predicting the next solar cycle’s intensity, in this paper we propose using them to help classify individual solar flares. We conduct experiments using HMI data employing four different machine learning algorithms that can exploit polar field information. Additionally, we propose a novel probabilistic mixture of experts model that can simply and effectively incorporate polar field data and provide on-par prediction performance with state-of-the-art solar flare prediction algorithms such as the Recurrent Neural Network (RNN). Our experimental results indicate the usefulness of the polar field data for solar flare prediction, which can improve Heidke Skill Score (HSS2) by as much as 10.1%.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Comparing generator predictions of transverse kinematic imbalance in neutrino-argon scattering

The largest uncertainties in estimating neutrino-nucleus interaction cross sections lie in the incomplete understanding of nuclear effects. A powerful tool to study nuclear effects is Transverse Kinematic Imbalance. This paper presents the first detailed comparison of the predictions of multiple event generators for distributions associated with Transverse Kinematic Imbalance for neutrino interactions on argon. Predictions for muon neutrinos interacting with an argon target are obtained using four standard neutrino event generation tools (GENIE, NuWro, GiBUU and NEUT). Example opportunities for discrimination between nuclear models leveraging future measurements are highlighted. The predictions shown in this paper are motivated by studying muon neutrinos from the Fermilab Booster Neutrino Beam interacting at the location of the MicroBooNE liquid argon time projection chamber, but the methods directly apply to other accelerator-based liquid argon neutrino experiments such as SBND, ICARUS and DUNE.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2025

Tunturi virus isolates and metagenome-assembled viral genomes provide insights into the virome of Acidobacteriota in Arctic tundra soils

Arctic soils are climate-critical areas, where microorganisms play crucial roles in nutrient cycling processes. Acidobacteriota are phylogenetically and physiologically diverse bacteria that are abundant and active in Arctic tundra soils. Still, surprisingly little is known about acidobacterial viruses in general and those residing in the Arctic in particular. Here, we applied both culture-dependent and -independent methods to study the virome of Acidobacteriota in Arctic soils. Five virus isolates, Tunturi 1–5, were obtained from Arctic tundra soils, Kilpisjärvi, Finland (69°N), using Tunturiibacter spp. strains originating from the same area as hosts. The new virus isolates have tailed particles with podo- (Tunturi 1, 2, 3), sipho- (Tunturi 4), or myovirus-like (Tunturi 5) morphologies. The dsDNA genomes of the viral isolates are 63–98 kbp long, except Tunturi 5, which is a jumbo phage with a 309-kbp genome. Tunturi 1 and Tunturi 2 share 88% overall nucleotide identity, while the other three are not related to one another. For over half of the open reading frames in Tunturi genomes, no functions could be predicted. To further assess the Acidobacteriota-associated viral diversity in Kilpisjärvi soils, bulk metagenomes from the same soils were explored and a total of 1881 viral operational taxonomic units (vOTUs) were bioinformatically predicted. Almost all vOTUs (98%) were assigned to the class Caudoviricetes. For 125 vOTUs, including five (near-)complete ones, Acidobacteriota hosts were predicted. Acidobacteriota-linked vOTUs were abundant across sites, especially in fens. Terriglobia-associated proviruses were observed in Kilpisjärvi soils, being related to proviruses from distant soils and other biomes. Approximately genus- or higher-level similarities were found between the Tunturi viruses, Kilpisjärvi vOTUs, and other soil vOTUs, suggesting some shared groups of Acidobacteriota viruses across soils. This study provides acidobacterial virus isolates as laboratory models for future research and adds insights into the diversity of viral communities associated with Acidobacteriota in tundra soils. Predicted virus-host links and viral gene functions suggest various interactions between viruses and their host microorganisms. Largely unknown sequences in the isolates and metagenome-assembled viral genomes highlight a need for more extensive sampling of Arctic soils to better understand viral functions and contributions to ecosystem-wide cycling processes in the Arctic.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2022

NaroNet: Discovery of tumor microenvironment elements from highly multiplexed images

Understanding the spatial interactions between the elements of the tumor microenvironment -i.e. tumor cells. fibroblasts, immune cells- and how these interactions relate to the diagnosis or prognosis of a tumor is one of the goals of computational pathology. We present NaroNet, a deep learning framework that models the multi-scale tumor microenvironment from multiplex-stained cancer tissue images and provides patient-level interpretable predictions using a seamless end-to-end learning pipeline. Trained only with multiplex-stained tissue images and their corresponding patient-level clinical labels, NaroNet unsupervisedly learns which cell phenotypes, cell neighborhoods, and neighborhood interactions have the highest influence to predict the correct label. To this end, NaroNet incorporates several novel and state-of-the-art deep learning techniques, such as patch-level contrastive learning, multi-level graph embeddings, a novel max-sum pooling operation, or a metric that quantifies the relevance that each microenvironment element has in the individual predictions. We validate NaroNet using synthetic data simulating multiplex-immunostained images where a patient label is artificially associated to the -adjustable- probabilistic incidence of different microenvironment elements. We then apply our model to two sets of images of human cancer tissues: 336 seven-color multiplex-immunostained images from 12 high-grade endometrial cancer patients; and 382 35-plex mass cytometry images from 215 breast cancer patients. In both synthetic and real datasets, NaroNet provides outstanding predictions of relevant clinical information while associating those predictions to the presence of specific microenvironment elements.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2022

Composition design of high-entropy alloys with deep sets learning

High entropy alloys (HEAs) are an important material class in the development of next-generation structural materials, but the astronomically large composition space cannot be efficiently explored by experiments or first-principles calculations. Machine learning (ML) methods might address this challenge, but ML of HEAs has been hindered by the scarcity of HEA property data. In this work, the EMTO-CPA method was used to generate a large HEA dataset (spanning a composition space of 14 elements) containing 7086 cubic HEA structures with structural properties, 1911 of which have the complete elastic tensor calculated. The elastic property dataset was used to train a ML model with the Deep Sets architecture. The Deep Sets model has better predictive performance and generalizability compared to other ML models. Association rule mining was applied to the model predictions to describe the compositional dependence of HEA elastic properties and to demonstrate the potential for data-driven alloy design.

36 MATERIALS SCIENCE↗

DOE OSTI2020

Characterization of Rare-earth Doped Thermal Barrier Coatings for Phosphor Thermometry

Thermal Barrier Coatings have been extensively used to protect and insulate the metallic components in turbine engines from extreme environments to achieve higher turbine inlet temperatures, resulting in an increase in combined engine cycle efficiency, lowering NOx emissions and fuel consumption. Additionally, as the major failure mechanisms determining lifetime are thermally activated during engine operation, uncertainty in temperature measurements affects lifetime prediction. Early detection of Thermal Barrier Coatings spallation symptoms, associated with local delamination induced temperature changes, can reduce forced engine outages. Further improvements are envisioned with the development of more reliable measurement techniques. For this purpose, Phosphor Thermometry has been considered a promising method for precision monitoring of turbine blade coatings that contain embedded phosphor dopants. In this work, a modified four-flux Kubelka-Munk model for luminescent rare-earth doped Thermal Barrier Coating configurations supported the design and the fabrication of viable sensing coatings for potential industrial implementation. Phosphor Thermometry instrumentation was developed with the objective of expanding capabilities of the technique, using a synchronized acquisition method on sensing coatings. Experiments on an innovative Erbium-Europium co-doped Yttria-Stabilized Zirconia coating have demonstrated the effectiveness of this advanced Phosphor Thermometry instrument with enhanced sensitivity and extended temperature range. Delamination progression monitoring was achieved for the first time using a Thermal Barrier Coating configuration subjected to indentation that includes a thin luminescent layer deposited on the surface and associating the multi-layer configuration with a predictive model that evaluates the advancement of the degradation of the coatings. To further ensure the integrity of the phosphor doped coatings, synchrotron X-ray diffraction was performed to characterize the specific residual strains and thermal expansion of the materials, determining mechanical robustness and compatibility with state-of-the-art Thermal Barrier Coatings. The outcomes of this work pave the way for in-situ temperature measurements on phosphor-doped turbine blade coatings with increased performance and accuracy.

Fouliard, Quentin↗

DOE OSTI2026

HydraGNN_Predictive_GFM_2026 - Ensemble of predictive graph foundation models for atomistic materials modeling

This release contains data and parameters of HydraGNN-based graph foundation models trained as a result of the work published in the pre-print "Exascale Multi-Task Graph Foundation Models for Imbalanced, Multi-Fidelity Atomistic Data" by M. Lupo Pasini et al. (https://arxiv.org/abs/2604.15380). We jointly train on 16 open first-principles datasets (544+ million structures covering 85+ elements) using a multi-task architecture with per-dataset heads and a scalable ADIOS2/DDStore data pipeline. On Frontier, we execute six large-scale DeepHyper hyperparameter optimization campaigns in FP64 and promote the top-performing message-passing models to sustained 2,048-node training, yielding a PaiNN-based lead model. The version of HydraGNN used to generate the outputs provided in this release is HydraGNN v5.0 (https://github.com/ORNL/HydraGNN/releases/tag/v5.0) The list of datasets used for the training of the graph foundation model is the following: 1) Alexandria [1] 2) ANI1x [2] 3) MPTrj [3] 4) Open Catalyst 2020 (OC20) [4] 5) Open Catalyst 2022 (OC22) [5] 6) Open Catalyst 2025 (OC25) [6] 7) Open Direct ir Capture 2023 (ODAC23) [7] 8) Open Materials 2024 (OMat24) [8] 9) Open Molecules 2025 (OMol25) [9] 10) OMol25-neutral (subset of OMol25 that contains only molecules with zero total charge) 11) OMol25-non-neutral (subset of OMol25 that contains only molecules with non-zero total charge) 12) Open Polymers 2026 (OPoly2026) [10] 13) Nabla2DFT [11] 14) QCML [12] 15) QM7X [reference 13] 16) transition1x [14] Dataset references: [1] J. Schmidt et al., “A dataset of 175k stable and metastable materials calculated with the PBEsol and SCAN functionals,” Scientific Data, vol. 9, p. 64, 2022. [2] J. S. Smith et al., “The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules,” Scientific Data, vol. 7, p. 134, 2020. [Online]. Available: https: //www.nature.com/articles/s41597-020-0473-z [3] A. Jain et al., “Commentary: The Materials Project: A materials genome approach to accelerating materials innovation,” APL Materials, vol. 1, no. 1, p. 011002, 07 2013. [Online]. Available: https://doi.org/10.1063/1.4812323 [4] L. Chanussot et al., “Open catalyst 2020 (oc20) dataset and community challenges,” ACS Catalysis, vol. 11, no. 10, pp. 6059–6072, 2021. [Online]. Available: https://doi.org/10.1021/acscatal.0c04525 [5] K. Tran et al., “Open catalyst 2022 (oc22) dataset and challenges for oxidation electrocatalysts,” ACS Catalysis, vol. 13, no. 5, pp. 3066–3084, 2023. [Online]. Available: https://doi.org/10.1021/acscatal.2c05426 [6] S. J. Sahoo et al., “The open catalyst 2025 (oc25) dataset and models for solid-liquid interfaces,” arXiv preprint arXiv:2509.17862, 2025. [Online]. Available: https://arxiv.org/abs/2509.17862 [7] A. Sriram et al., “The open DAC 2023 dataset and challenges for sorbent discovery in direct air capture,” ACS Central Science, vol. 10, no. 5, pp. 923–941, 2024. [8] L. Barroso-Luque et al., “Open materials 2024 (omat24) inorganic materials dataset and models,” 2024. [Online]. Available: https://arxiv.org/abs/2410.12771 [9] D. S. Levine et al., “The open molecules 2025 (OMol25) dataset, evaluations, and models,” 2025. [Online]. Available: https://arxiv.org/abs/2505.08762 [10] D. S. Levine et al., The open polymers 2026 (OPoly26) dataset and evaluations,” arXiv preprint arXiv:2512.23117, 2025. [Online]. Available: https://arxiv.org/abs/2512.23117 [11] K. Khrabrov et al., “Nabla2dft: A universal quantum chemistry dataset of drug-like molecules and a benchmark for neural network potentials,” in NeurIPS 2024 Datasets and Benchmarks Track, 2024. [Online]. Available: https://openreview.net/forum?id=ElUrNM9U8c [12] S. Ganscha et al., “The QCML dataset, quantum chemistry reference data from 33.5M DFT and 14.7B semi-empirical calculations,” Scientific Data, vol. 12, p. 406, 2025. [13] J. Hoja et al., “QM7-X, a comprehensive dataset of quantum-mechanical properties spanning the chemical space of small organic molecules,” Scientific Data, vol. 8, p. 43, 2021. [Online]. Available: https://www.nature.com/articles/s41597-021-00812-2 [14] M. Schreiner et al., “Transition1x - a dataset for building generalizable reactive machine learning potentials,” Scientific Data, vol. 9, p. 779, 2022. The folder "datasets_ADIOS2_format" contains the set of pre-processed datasets in Adaptable I/O System (ADIOS) format (https://www.exascaleproject.org/research-project/adios/) that have been used for the development and training of GFMs in this work. The "datasets_ADIOS2_format" directory contains 2 sub-directories, one for the version "v1" of the datasets and one for the version "v2" of the datasets. The version "v1" of the datasets provides values of the total energy as they are extracted from the original data as it was released by the respective institutions. The version "v2" of the datasets provides values of the energy that have been realigned. The realignment was performed by training a linear regression model that predicts the total energy as a function of the chemical composition of the atomistic structure, and then subtract such prediction from the original value of the total energy. Both folders "v1" and "v2" contain 16 sub-directories, each corresponding to an ADIOS2-formatted dataset The folder "DeepHyper-results" contains the configurational files and model's parameters for all the 186 HPO trials that were successfully completed by the scalable hyperparameter optimization (HPO) runs on Frontier. The content of the folder "DeepHyper-results" I structured as follows: 1) task-list.txt: list of mpnn name, jobid, and deephyper task id 2) gfm_${MPNN}_${JOBID}_0.${TASKID}: run directory with checkpoint files 3) gfm_${MPNN}: deephyper summary directory (*.csv) for each specific MPNN type 4) deephyper-experiment-${JOBID}: output and error logs for each job The file "deephyper-sorted.csv" contains the details of each HydraGNN model built and tested by HPO, obtained by merging the (*.csv) filed from each HPO run executed. Out of all the HPO trials, we selected 10 to continue the training of the respective HydraGNN models. Due to limited computational budget available in the LRN070 allocation we could not complete the training till convergence for all these 10 selected models. The folder "models" contains multiple sub-folders, one per each HydraGNN model trained. Each model sub-folder contains the parameters of each HydraGNN model, with multiple checkpoint-restarts. The list of sub-folders are as follows: 1) multidataset_hpo-BEST1-fp64 2) multidataset_hpo-BEST2-fp64 3) multidataset_hpo-BEST3-fp64 4) multidataset_hpo-BEST4-fp64 5) multidataset_hpo-BEST5-fp64 6) multidataset_hpo-BEST6-fp64 7) multidataset_hpo-BEST7-fp64 8) multidataset_hpo-BEST8-fp64 9) multidataset_hpo-BEST9-fp64 10) multidataset_hpo-BEST10-fp64 Within each one of these folders, additional auxiliary log files are provided with descriptions about how the training proceeded. The lead PaiNN-model is contained inside "multidataset_hpo-BEST6-fp64". The file "mlp_branch_weights" contains the parameters of the multi-layer perceptron (MLP) used to reconcile the predictions of the 16 output decoding heads of the HydragNN architectures. The MLP takes in input the chemical composition of the atomistic structure and predicts averaging weights to linearly mix the predictions of each output decoding head toward consolidating them into a single one. The folder "1.1billion-structure-inference" contains 1.1 billion atomistic structures randomly generated. Each structures is associated with energy and forces predicted with the lead-PaiNN model combined with the MLP model for reconciliation of the multi-branch predictions generated by the 16 output decoding heads. The folder "1.1billion-structure-inference" contains 9,300 (*.tar.gz) subdirectories, one per Frontier compute node used to execute the inference at exascale. Once uncompressed, each (*.tar.gz) subdirectory contains an ADIOS2 (*.bp) file container, where each atomistic structure is stored as a PyTorch-Geometric Data object. The file "export_dataset_environment_variables.sh" contains the environment variables that need to be set before running the HydraGNN code to reproduce the results provided in this dataset release. The code that can be used to load the ADIOS2 files, load HydraGNN models, and run inference is available at: https://github.com/ORNL/HydraGNN/releases/tag/v5.0

36 MATERIALS SCIENCE↗