Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Graph processing”

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 829 records · Page 46
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

Using mixed methods to construct and analyze a participatory agent-based model of a complex Zimbabwean agro-pastoral system

Complex social-ecological systems can be difficult to study and manage. Simulation models can facilitate exploration of system behavior under novel conditions, and participatory modeling can involve stakeholders in developing appropriate management processes. Participatory modeling already typically involves qualitative structural validation of models with stakeholders, but with increased data and more sophisticated models, quantitative behavioral validation may be possible as well. In this study, we created a novel agent-basedmodel applied to a specific context: Zimbabwean non-governmental organization the Muonde Trust has been collecting data on their agro-pastoral system for the last 35 years and had concerns about land-use planning and the effectiveness of management interventions in the face of climate change. We collaboratively created an agent-based model of their system using their data archive, qualitatively calibrating it to the observed behavior of the real system without tuning any parameters to match specific quantitative outputs. We then behaviorally validated the model using quantitative community-based data and conducted a sensitivity analysis to determine the relative impact of underlying parameter assumptions, Indigenous management interventions, and different rainfall variation scenarios. We found that our process resulted in a model which was successfully structurally validated and sufficiently realistic to be useful for Muonde researchers as a discussion tool. The model was inconsistently behaviorally validated, however, with some model variables matching field data better than others. We observed increased model system instability due to increasing variability in underlying drivers (rainfall), and also due to management interventions that broke feedbacks between the components of the system. Interventions that smoothed year-to-year variation rather than exaggerating it tended to improve sustainability. The Muonde trust has used the model to successfully advocate to local leaders for changes in land-use planning policy that will increase the sustainability of their system.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2025

Causal discovery from data assisted by large language models

Knowledge-driven discovery of novel materials necessitates the development of causal models for property emergence. While in the classical physical paradigm, the causal relationships are deduced based on physical principles or via experiment, the rapid accumulation of observational data necessitates learning causal relationships between dissimilar aspects of material structure and functionalities based on observations. For this, it is essential to integrate experimental data with prior domain knowledge. Here, we demonstrate this approach by combining high-resolution scanning transmission electron microscopy data with insights derived from large language models (LLMs). By applying ChatGPT to domain-specific literature, such as arXiv papers on ferroelectrics, and combining the obtained information with data-driven causal discovery, we construct adjacency matrices for directed acyclic graphs that map the causal relationships between structural, chemical, and polarization degrees of freedom in Sm-doped BiFeO 3 . This approach enables us to hypothesize how synthesis conditions influence material properties and guides experimental validation. Furthermore, the ultimate objective of this work is to develop a unified framework that integrates LLM-driven literature analysis with data-driven discovery, facilitating the precise engineering of ferroelectric materials by establishing clear connections between synthesis conditions and their resulting material properties.

Causal inference↗
DOE OSTI2025

wa-hls4ml and lui-gnn: A benchmark and GNN-based surrogate model for hls4ml resource and latency estimation

As machine learning (ML) increasingly serves as a tool for addressing real-time challenges in scientific applications, the development of advanced tooling has significantly reduced the time required to iterate on various designs. These advancements have solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks, such as model synthesis, are now becoming limiting factors in the rapid iteration of designs. To reduce these emerging constraints, multiple efforts are being launched toward designing an ML-based surrogate model that estimates resource usage of synthesized accelerator architectures. This model would reduce the design iteration time, especially when designing within a set of given hardware constraints. This approach shows considerable potential, but as it stands, the effort is early and would benefit from coordination and standardization to assist future work as it emerges. We introduce wa-hls4ml, a benchmark for ML accelerator resource and latency estimation, and its corresponding initial dataset of more than 100,000 fully connected neural networks, all synthesized using hls4ml and targeting Xilinx FPGAs. In addition to the resource utilization and latency data provided, the dataset includes generated artifacts and log files for many of the synthesized neural networks, in order to support future research in ML-based code generation. The benchmark evaluates the performance of resource and latency predictors against several common ML model architectures, primarily originating from scientific domains, as exemplar models, as well as the average performance across a subset of the dataset. We measure the performance of a given predictor model through multiple metrics, including $R^2$ score and SMAPE on regression tasks, as well as inference time to further characterize the estimator under test. Additionally, we introduce the latency/utilization inference graph neural network (lui-gnn), a surrogate model that uses a graph neural network to represent input architectures in the form of a directed graph. This graph representation allows for a diverse set of model architectures to all be effectively handled by a surrogate model. We present the architecture and performance of the model, as evaluated by the new proposed benchmark, including SMAPE, $R^2$ score, and inference times, and find that lui-gnn generally predicts latency and utilization for the 75\% quantile within several percent of the synthesized resources on the synthetic test dataset, indicating that this approach of estimating resource and latency via a surrogate models has promise and warrants further research.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2025

Explainable Synthesizability Prediction of Inorganic Crystal Polymorphs Using Large Language Models

Abstract We evaluate the ability of machine learning to predict whether a hypothetical crystal structure can be synthesized and explain those predictions to scientists. Fine‐tuned large language models (LLMs) trained on a human‐readable text description of the target crystal structure perform comparably to previous bespoke convolutional graph neural network methods, but better prediction quality can be achieved by training a positive‐unlabeled learning model on a text‐embedding representation of the structure. An LLM‐based workflow can then be used to generate human‐readable explanations for the types of factors governing synthesizability, extract the underlying physical rules, and assess the veracity of those rules. These explanations can guide chemists in modifying or optimizing non‐synthesizable hypothetical structures to make them more feasible for materials design.

Kim, Seongmin [Department of Chemical and Biologic↗
DOE OSTI2025

Explainable Synthesizability Prediction of Inorganic Crystal Polymorphs Using Large Language Models

Abstract We evaluate the ability of machine learning to predict whether a hypothetical crystal structure can be synthesized and explain those predictions to scientists. Fine‐tuned large language models (LLMs) trained on a human‐readable text description of the target crystal structure perform comparably to previous bespoke convolutional graph neural network methods, but better prediction quality can be achieved by training a positive‐unlabeled learning model on a text‐embedding representation of the structure. An LLM‐based workflow can then be used to generate human‐readable explanations for the types of factors governing synthesizability, extract the underlying physical rules, and assess the veracity of those rules. These explanations can guide chemists in modifying or optimizing non‐synthesizable hypothetical structures to make them more feasible for materials design.

Kim, Seongmin [Department of Chemical and Biologic↗
DOE OSTI2020

Temporal Analysis and Scene Change Detection in Multispectral Overhead Imagery

Scene change detection can be a tedious and time consuming process especially when concerning large geographical areas, and the process can be even more cumbersome when analyzing changes in an area over large spans of time. Developing a useful way to help analysts recognize at what points in time significant changes to a scene have occurred can allow them to better focus their efforts in characterizing events. Applications include: Facility monitoring, Construction chronology, Monitoring of vehicle/aircraft activity, Characterization of larger sequences of events. In large areas exceeding hundreds to thousands of square kilometers in size, it can be difficult localizing when scene changes have occurred. Analysts can spend hours going through imagery to try to identify new construction, monitor facility activities, monitor vehicle movement, etc. where the object of interest may only be a few square meters. Our goal is to help cut down this time by giving analysts change maps with hot spots of change, allowing them to focus on regions that have experienced actual change in time frames they're interested in. Additionally, by combining these change maps into layers within a data cube, analysts can examine the change maps from a temporal perspective, allowing events to be characterized over spans of time. By opening the data cube in an imaging software capable of separating the layers, we can analyze the change maps sequentially, allowing us to examine scene changes occurring over time. As an example, we examined overhead imagery from Planet Labs of what appears to be a parking lot on Fort Irwin over the course of a year using ENVI, a geospatial satellite imagery analysis software. Using ENVI, we generate a graph of changes over time, and notice a particular segment near the end of our analysis window where no changes are detected. Examination of the actual satellite imagery reveals that during this time span, the parking lot was empty. This could be due to facility shutdown for maintenance or upgrades, or possibly even total workforce/vehicle fleet movement. Information like this could help analysts better characterize events, as well as to help create clearer timelines in larger sequences of events. Workflow steps: - Collect multiple maps of the same AOI (Area of Interest) during a time span of interest; - Generate change maps from AOI maps; - Generate data cube from change maps. An analyst can use the data cube to help inspect an AOI for activities within a time span of interest. If an event of interest is discovered, the analyst can then refer to the maps corresponding to the appropriate dates and times in the data cube to see precisely what is transpiring. The biggest objective being worked on is improving the change detection methodology employed. We currently use PCA-EM (Principal Component Analysis with Expectation Maximization), but we are currently focusing on implementing IR-MAD (Iteratively Reweighted Multivariate Alteration Detection) to be used in conjunction with PCA-EM in an effort to decrease false positivity and noise in the change maps we generate.

42 ENGINEERING↗
NASA NTRS2005

ASK Magazine; No. 21

THIS ISSUE FEATURES A VISUAL DEPICTION OF THE ACADEMY of Program and Project Leadership (APPL). I imagine a variety of initial reactions to the drawing. One might be, "What is a cartoon doing in a magazine about project management?" Or perhaps, "Wow, nice colors-and fun." Another may be to closely search the image for signs, symbols and meaning. Still another, to read a new level of innovation and creativity into the picture. Undoubtedly, some readers will raise questions about the cost. Of course, any reaction is a sign of engagement. The stronger, the more energized the emotional and cognitive processing, the better. It is a sign of attention and interaction. For I've heard it said, "You only need to worry if they don t care one way or the other." So what is the point of the picture? To stimulate interest, raise questions, promote discussion, and maybe raise a smile.. .That, at least, was my initial reaction when I was introduced to the work of Nancy Hegedus, who helps to create these drawings for Root Learning Inc. At the NASA PM Conference, I was first shown the work Nancy had been doing with the help of Goddard s Knowledge Management Architect, Dr. Ed Rogers. I was immediately drawn into the power of visualization as a tool for more effective learning, communicating, and conveying complex knowledge concepts. We need new tools in today s world, where information and data overwhelms by sheer volume. There are articles, pamphlets, communications, and white papers-all aiming to convince and influence. Reactions to these tend to be either avoidance or mind-numbing, heavy-eyed consent; the message never registers or enters the soul. That s one of the reasons that APPL s Knowledge Sharing Initiative (KSI) has turned to storytelling as a memorable way of transfer- ring knowledge, inspiring imitation of best practices, and spurring reflection. ASK Magazine s recent fourth birthday marks an important milestone in APPL s continuing quest to provide ongoing support to project managers and to promote mission success. And similar to storytelling, the power of visualization is receiving increasing attention in recent years as a way to stimulate engagement. Pictures and visual graphs are viewed as one of the most effective ways for displaying, describing, and generating discussion about quantitative and technically complex information. Prototypes, models, and simulations are considered essential for stimulating innovation through open and engaging discussions. There has also been extensive writing on the use of visual graphics, pictures, and cartoons to facilitate memory, creativity, openness, attention-and even well-being. For many of these reasons, I am excited to have a colorful visual depiction of the APPL world included in ASK. Without the addition of text or slides, the intent is to invite people into the world of the APPL mission-as well as its products, services, customers, and partners- in a fun and engaging manner. As project leaders strive to find ways to encourage engagement, learning, and transmission of knowledge, traditional technologies are proving to be as valuable as modern technologies. (But for those who want more information in the form of texts and slide presentations, we certainly have an abundance of those as well.)

Laufer, Alexander↗
DOE OSTI2025

RhizoGrid Indexed Sorghum Rhizosphere Multi-Omics

PerCon SFA project data dentification of spatially resolved biomarkers of drought in Sorghum bicolor rhizosphere molecular-microbe interactions using a novel root cartography "RhizoGrid" system for sampling plants under drought and control conditions across 10 equally sized root zone environments (4 quadrants each). Each quadrant was sampled and processed for 16S amplicon, metabolomics, and X-ray computed tomography (XCT). Data download includes experimental metadata and results files for 16S rRNA sequence analysis of microbial community assembly (processed data files), liquid chromatography mass spectrometry (LC-MS) metabolomics analysis of microbial community root exudates (processed data files), X-ray computed tomography (XCT) spatial gradient analysis (raw and processed data files) of microbial community composition, and related computational modeling outputs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Efficient Parallelization of Irregular Applications on GPU Architectures

With the enlarging computation capacity of general Graphics Processing Units (GPUs), leveraging GPUs to accelerate parallel applications has become a critical topic in academia and industry. However, a wide range of irregular applications with the computation-/memory-intensive nature cannot easily achieve high GPU utilization. The challenges mainly involve the following aspects: first, data dependence leads to coarse-grained kernel and inefficient parallelism; second, heavy GPU memory usage may cause frequent memory evictions and extra overhead of I/O; third, specific computation patterns produce memory redundancies; last, workload balance and data reusability conjunctly benefit the overall performance, but there may exist a dynamic trade-off between them. Targeting these challenges, this dissertation proposes multiple optimizations to accelerate two real-world applications: many-body correlation functions to simulate nuclear physics in a large-scale scientific system; the other is the eALS-based matrix factorization recommendation system. To accelerate the calculations of many-body correlation functions, this dissertation presents three frameworks in GPU memory management and multi-GPU scheduling. Firstly, an optimized systematic GPU memory management framework, MemHC, utilizes a series of new memory reduction designs in GPU memory allocation, CPU/GPU communications, and GPU memory oversubscription. Secondly, an enhanced multi-GPU scheduling framework, MICCO, particularly by taking both data dimension (e.g., data reuse and data eviction) and computation dimension into account. MICCO designs a heuristic scheduling algorithm and a machine learning-based regression model to generate the optimal settings of a proposed new concept to manage the trade-off. Thirdly, a locality-aware multi-GPU scheduling framework. This scheduler leverages pipeline batch generation with a looking-ahead strategy by building local dependency graphs for memory transfer reduction and better data reuse, achieving up to 79.92% memory cost reduction and 1.67x speedup. To parallelize the eALS-based recommendation system, this dissertation proposes an efficient CPU/GPU heterogeneous recommendation system, HEALS. HEALS employs newly designed architecture-adaptive data formats to achieve load balance and good data locality on CPU and GPU. To mitigate the data dependence, HEALS presents a CPU/GPU collaboration model for both task parallelism and data parallelism with multiple kernel computation optimizations. In summary, this dissertation efficiently accelerates two typical irregular applications on GPUs by building four frameworks, including CPU/GPU collaboration, GPU memory management, and multi-GPU scheduling.

Wang, Qihan↗
DOE OSTI2024

The circular bioeconomy: a driver for system integration

Background: Human and earth system modeling, traditionally centered on the interplay between the energy system and the atmosphere, are facing a paradigm shift. The Intergovernmental Panel on Climate Change’s mandate for comprehensive, cross-sectoral climate action emphasizes avoiding the vulnerabilities of narrow sectoral approaches. Our study explores the circular bioeconomy, highlighting the intricate interconnections among agriculture, forestry, aquaculture, technological advancements, and ecological recycling. Collectively, these sectors play a pivotal role in supplying essential resources to meet the food, material, and energy needs of a growing global population. We pose the pertinent question of what it takes to integrate these multifaceted sectors into a new era of holistic systems thinking and planning. Results: The foundation for discussion is provided by a novel graphical representation encompassing statistical data on food, materials, energy flows, and circularity. This representation aids in constructing an inventory of technological advancements and climate actions that have the potential to significantly reshape the structure and scale of the economic metabolism in the coming decades. In this context, the three dominant mega-trends—population dynamics, economic developments, and the climate crisis—compel us to address the potential consequences of the identified actions, all of which fall under the four categories of substitution, efficiency, sufficiency, and reliability measures. Substitution and efficiency measures currently dominate systems modeling. Including novel bio-based processes and circularity aspects might require only expanded system boundaries. Conversely, paradigm shifts in systems engineering are expected to center on sufficiency and reliability actions. Effectively assessing the impact of sufficiency measures will necessitate substantial progress in inter- and transdisciplinary collaboration, primarily due to their non-technological nature. In addition, placing emphasis on modeling the reliability and resilience of transformation pathways represents a distinct and emerging frontier that highlights the significance of an integrated network of networks. Conclusions: Existing and emerging circular bioeconomy practices can serve as prime examples of system integration. These practices facilitate the interconnection of complex biomass supply chain networks with other networks encompassing feedstock-independent renewable power, hydrogen, CO 2 , water, and other biotic, abiotic, and intangible resources. Elevating the prominence of these connectors will empower policymakers to steer the amplification of synergies and mitigation of tradeoffs among systems, sectors, and goals.

09 BIOMASS FUELS↗
NASA NTRS2011

Superpixel-Augmented Endmember Detection for Hyperspectral Images

Superpixels are homogeneous image regions comprised of several contiguous pixels. They are produced by shattering the image into contiguous, homogeneous regions that each cover between 20 and 100 image pixels. The segmentation aims for a many-to-one mapping from superpixels to image features; each image feature could contain several superpixels, but each superpixel occupies no more than one image feature. This conservative segmentation is relatively easy to automate in a robust fashion. Superpixel processing is related to the more general idea of improving hyperspectral analysis through spatial constraints, which can recognize subtle features at or below the level of noise by exploiting the fact that their spectral signatures are found in neighboring pixels. Recent work has explored spatial constraints for endmember extraction, showing significant advantages over techniques that ignore pixels relative positions. Methods such as AMEE (automated morphological endmember extraction) express spatial influence using fixed isometric relationships a local square window or Euclidean distance in pixel coordinates. In other words, two pixels covariances are based on their spatial proximity, but are independent of their absolute location in the scene. These isometric spatial constraints are most appropriate when spectral variation is smooth and constant over the image. Superpixels are simple to implement, efficient to compute, and are empirically effective. They can be used as a preprocessing step with any desired endmember extraction technique. Superpixels also have a solid theoretical basis in the hyperspectral linear mixing model, making them a principled approach for improving endmember extraction. Unlike existing approaches, superpixels can accommodate non-isometric covariance between image pixels (characteristic of discrete image features separated by step discontinuities). These kinds of image features are common in natural scenes. Analysts can substitute superpixels for image pixels during endmember analysis that leverages the spatial contiguity of scene features to enhance subtle spectral features. Superpixels define populations of image pixels that are independent samples from each image feature, permitting robust estimation of spectral properties, and reducing measurement noise in proportion to the area of the superpixel. This permits improved endmember extraction, and enables automated search for novel and constituent minerals in very noisy, hyperspatial images. This innovation begins with a graph-based segmentation based on the work of Felzenszwalb et al., but then expands their approach to the hyperspectral image domain with a Euclidean distance metric. Then, the mean spectrum of each segment is computed, and the resulting data cloud is used as input into sequential maximum angle convex cone (SMACC) endmember extraction.

Thompson, David R.↗
DOE OSTI2023

Local Decomposition of Hexahedral Singular Nodes into Singular Curves

Hexahedral (hex) meshing is a long studied topic in geometry processing with many challenging associated problems. Hex meshes vary from structured to unstructured depending on application or domain of interest. Fully structured meshes require that all interior mesh edges be adjacent to four hexes each. Edges failing this criteria are singular and indicate an unstructured hex mesh. Singular edges join together into singular curves that either form closed cycles, end on the mesh boundary, or end at a singular node, a complex junction of more than two singular curves. Hex meshes with more complex singular nodes tend to have more distorted elements and smaller scaled Jacobian values. In this work, we study the topology of singular nodes. We show that all eight of the most common singular nodes are decomposable into just singular curves. We further show that all singular nodes, regardless of edge valence, are locally decomposable. Finally we demonstrate these decompositions on hex meshes, thereby decreasing their distortion and converting all singular nodes into singular curves. In conclusion, with this decomposition, the enigmatic complexity of 3D singular nodes becomes effectively 2D.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2020

The matrix product approximation for the dynamic cavity method

Stochastic dynamics of classical degrees of freedom, defined on vertices of locally tree-like graphs, can be studied in the framework of the dynamic cavity method which is exact for tree graphs. Such models correspond for example to spin-glass systems, Boolean networks, neural networks, and other technical, biological, and social networks. The central objects in the cavity method are edge messages—conditional probabilities of two vertex variable trajectories. In this paper, we discuss a rather pedagogical derivation for the dynamic cavity method, give a detailed account of the novel matrix product edge message (MPEM) algorithm for the solution of the dynamic cavity equation as introduced, and present optimizations and extensions. Matrix product approximations of the edge messages are constructed recursively in an iteration over time. Computation costs and precision can be tuned by controlling the matrix dimensions of the MPEM in truncations. Without truncations, the dynamics is exact. Data for Glauber–Ising dynamics shows a linear growth of computation costs in time. In contrast to Monte Carlo simulations, the approach has a much better error scaling. Hence, it gives for example access to low probability events and decaying observables like temporal correlations. Here, we discuss optimized truncation schemes and an extension that allows to capture models which have a continuum time limit.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2025

Graph-Learning-Assisted State and Event Tracking for Solar-Penetrated Power Grids with Heterogeneous Data Sources

Unlike transmission systems, distribution systems do not typically contain sufficient metering to enable real-time state estimation. The lack of sufficient real-time measurements prohibits accurate and timely monitoring of the state of distribution systems. As a result, control and optimal operation of distribution systems, especially those containing large numbers of renewable generation units are not possible without proper data and information about the current state of the system. The main motivation of this project is to address this shortcoming by developing an approach which provides “predicted” real-time measurements so that they can be used to execute a distribution system state estimator. Thus, the objective of the project is to make the distribution systems fully observable, such that the hosting capacity for solar generation can be accurately estimated, and unnecessary solar curtailments can be avoided. In order to accomplish this goal, the project investigated the use of a grid-model-informed machine learning (ML) tool which integrates heterogeneous data streams obtained from AMI meters, SCADA as well as PMU measurements and created synchronous measurement snapshots for the state estimator (SE); and developed a hybrid robust SE which provides not only accurate state estimates but also real-time feedback for the ML model refinement.

14 SOLAR ENERGY↗
NASA NTRS2004

Optical Calibration Process Developed for Neural-Network-Based Optical Nondestructive Evaluation Method

A completely optical calibration process has been developed at Glenn for calibrating a neural-network-based nondestructive evaluation (NDE) method. The NDE method itself detects very small changes in the characteristic patterns or vibration mode shapes of vibrating structures as discussed in many references. The mode shapes or characteristic patterns are recorded using television or electronic holography and change when a structure experiences, for example, cracking, debonds, or variations in fastener properties. An artificial neural network can be trained to be very sensitive to changes in the mode shapes, but quantifying or calibrating that sensitivity in a consistent, meaningful, and deliverable manner has been challenging. The standard calibration approach has been difficult to implement, where the response to damage of the trained neural network is compared with the responses of vibration-measurement sensors. In particular, the vibration-measurement sensors are intrusive, insufficiently sensitive, and not numerous enough. In response to these difficulties, a completely optical alternative to the standard calibration approach was proposed and tested successfully. Specifically, the vibration mode to be monitored for structural damage was intentionally contaminated with known amounts of another mode, and the response of the trained neural network was measured as a function of the peak-to-peak amplitude of the contaminating mode. The neural network calibration technique essentially uses the vibration mode shapes of the undamaged structure as standards against which the changed mode shapes are compared. The published response of the network can be made nearly independent of the contaminating mode, if enough vibration modes are used to train the net. The sensitivity of the neural network can be adjusted for the environment in which the test is to be conducted. The response of a neural network trained with measured vibration patterns for use on a vibration isolation table in the presence of various sources of laboratory noise is shown. The output of the neural network is called the degradable classification index. The curve was generated by a simultaneous comparison of means, and it shows a peak-to-peak sensitivity of about 100 nm. The following graph uses model generated data from a compressor blade to show that much higher sensitivities are possible when the environment can be controlled better. The peak-to-peak sensitivity here is about 20 nm. The training procedure was modified for the second graph, and the data were subjected to an intensity-dependent transformation called folding. All the measurements for this approach to calibration were optical. The peak-to-peak amplitudes of the vibration modes were measured using heterodyne interferometry, and the modes themselves were recorded using television (electronic) holography.

Decker, Arthur J.↗
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↗
NASA NTRS2013

Issues and Strategies in Solving Multidisciplinary Optimization Problems

Optimization research at NASA Glenn Research Center has addressed the design of structures, aircraft and airbreathing propulsion engines. The accumulated multidisciplinary design activity is collected under a testbed entitled COMETBOARDS. Several issues were encountered during the solution of the problems. Four issues and the strategies adapted for their resolution are discussed. This is followed by a discussion on analytical methods that is limited to structural design application. An optimization process can lead to an inefficient local solution. This deficiency was encountered during design of an engine component. The limitation was overcome through an augmentation of animation into optimization. Optimum solutions obtained were infeasible for aircraft and airbreathing propulsion engine problems. Alleviation of this deficiency required a cascading of multiple algorithms. Profile optimization of a beam produced an irregular shape. Engineering intuition restored the regular shape for the beam. The solution obtained for a cylindrical shell by a subproblem strategy converged to a design that can be difficult to manufacture. Resolution of this issue remains a challenge. The issues and resolutions are illustrated through a set of problems: Design of an engine component, Synthesis of a subsonic aircraft, Operation optimization of a supersonic engine, Design of a wave-rotor-topping device, Profile optimization of a cantilever beam, and Design of a cylindrical shell. This chapter provides a cursory account of the issues. Cited references provide detailed discussion on the topics. Design of a structure can also be generated by traditional method and the stochastic design concept. Merits and limitations of the three methods (traditional method, optimization method and stochastic concept) are illustrated. In the traditional method, the constraints are manipulated to obtain the design and weight is back calculated. In design optimization, the weight of a structure becomes the merit function with constraints imposed on failure modes and an optimization algorithm is used to generate the solution. Stochastic design concept accounts for uncertainties in loads, material properties, and other parameters and solution is obtained by solving a design optimization problem for a specified reliability. Acceptable solutions can be produced by all the three methods. The variation in the weight calculated by the methods was found to be modest. Some variation was noticed in designs calculated by the methods. The variation may be attributed to structural indeterminacy. It is prudent to develop design by all three methods prior to its fabrication. The traditional design method can be improved when the simplified sensitivities of the behavior constraint is used. Such sensitivity can reduce design calculations and may have a potential to unify the traditional and optimization methods. Weight versus reliability traced out an inverted-S-shaped graph. The center of the graph corresponded to mean valued design. A heavy design with weight approaching infinity could be produced for a near-zero rate of failure. Weight can be reduced to a small value for a most failure-prone design. Probabilistic modeling of load and material properties remained a challenge.

Patnaik, Surya↗
DOE OSTI2025

Machine learning for single-ended event reconstruction in PROSPECT experiment

The Precision Reactor Oscillation and Spectrum Experiment, PROSPECT, was a segmented antineutrino detector that successfully operated at the High Flux Isotope Reactor in Oak Ridge, TN, during its 2018 run. Despite challenges with photomultiplier tube base failures affecting some segments, innovative machine learning approaches were employed to perform position and energy reconstruction, and particle classification. This work highlights the effectiveness of convolutional neural networks and graph convolutional networks in enhancing data analysis. By leveraging these techniques, a 3.3% increase in effective statistics was achieved compared to traditional methods, showcasing their potential to improve analysis performance. Furthermore, these machine learning methodologies offer promising applications for other segmented particle detectors, underscoring their versatility and impact.

47 OTHER INSTRUMENTATION↗