Search NASA⌕ Search

SEARCH · Search NASA

Results for “Benchmark dataset”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2021

Dataset of low global warming potential refrigerant refrigeration system for fault detection and diagnostics

Abstract HVAC and refrigeration system fault detection and diagnostics (FDD) has attracted extensive studies for decades; however, FDD of supermarket refrigeration systems has not gained significant attention. Supermarkets consume around 50 kWh/ft 2 of electricity annually. The biggest consumer of energy in a supermarket is its refrigeration system, which accounts for 40%–60% of its total electricity usage and is equivalent to about 2%–3% of the total energy consumed by commercial buildings in the United States. Also, the supermarket refrigeration system is one of the biggest consumers of refrigerants. Reducing refrigerant usage or using environmentally friendly alternatives can result in significant climate benefits. A challenge is the lack of publicly available data sets to benchmark the system performance and record the faulted performance. This paper identifies common faults of supermarket refrigeration systems and conducts an experimental study to collect the faulted performance data and analyze these faults. This work provides a foundation for future research on the development of FDD methods and field automated FDD implementation.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗
DOE OSTI2023

CholecTriplet2021: A benchmark challenge for surgical action triplet recognition

Context-aware decision support in the operating room can foster surgical safety and efficiency by leveraging real-time feedback from surgical workflow analysis. Most existing works recognize surgical activities at a coarse-grained level, such as phases, steps or events, leaving out fine-grained interaction details about the surgical activity; yet those are needed for more helpful AI assistance in the operating room. Recognizing surgical actions as triplets of ‹ instrument, verb, target › combination delivers more comprehensive details about the activities taking place in surgical videos. This paper presents CholecTriplet2021: an endoscopic vision challenge organized at MICCAI 2021 for the recognition of surgical action triplets in laparoscopic videos. Here, the challenge granted private access to the large-scale CholecT50 dataset, which is annotated with action triplet information. In this paper, we present the challenge setup and the assessment of the state-of-the-art deep learning methods proposed by the participants during the challenge. Here, a total of 4 baseline methods from the challenge organizers and 19 new deep learning algorithms from the competing teams are presented to recognize surgical action triplets directly from surgical videos, achieving mean average precision (mAP) ranging from 4.2% to 38.1%. This study also analyzes the significance of the results obtained by the presented approaches, performs a thorough methodological comparison between them, in-depth result analysis, and proposes a novel ensemble method for enhanced recognition. Our analysis shows that surgical workflow analysis is not yet solved, and also highlights interesting directions for future research on fine-grained surgical activity recognition which is of utmost importance for the development of AI in surgery.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2026

Data from: "Towards CONUS-Wide ML-Augmented Conceptually-Interpretable Modeling of Catchment-Scale Precipitation-Storage-Runoff Dynamics"

This data package was generated to support the manuscript “Towards CONUS-Wide Machine Learning-Augmented Conceptually Interpretable Modeling of Catchment-Scale Precipitation-Storage-Runoff Dynamics.” It provides input files, model outputs, plotting data, scripts, notebooks, and documentation used to develop, evaluate, and reproduce Mass-Conserving Perceptron (MCP)-based hydrologic modeling experiments across 513 selected Catchment Attributes and Meteorology for Large-sample Studies in the United States (CAMELS-US) basins. The files are organized by modeling component and analysis purpose, including rainfall–runoff experiments, snow module experiments, coupled hydrologic-snow experiments, Long Short-Term Memory (LSTM) benchmark results, model skill metrics, initialization and epoch records, cell-state normalization files, Akaike Information Criterion (AIC)-based model comparison files, and data used to generate manuscript figures. Tabular files can be opened using standard spreadsheet software or Python/R data-analysis tools. Python scripts, Jupyter notebooks, and selected MATLAB scripts are included for model execution, postprocessing, plotting, and statistical analysis. Quality assurance and quality control were conducted through the source-data selection and modeling workflow. Meteorological forcing, streamflow, and static catchment attributes were derived from the CAMELS-US dataset, and snow water equivalent data were derived from the University of Arizona (UA) Snow Water Equivalent dataset. Selected basins and time periods were screened during the associated research workflow to avoid missing observations or poor-quality cases. Static geospatial features were processed primarily using Quantum Geographic Information System (QGIS) and Geospatial Data Abstraction Library (GDAL) workflows. Additional details are provided in the associated manuscript and documentation.

ESS-DIVE CSV File Formatting Guidelines Reporting ↗
DOE OSTI2024

Benchmarking machine learning interatomic potentials via phonon anharmonicity

Abstract Machine learning approaches have recently emerged as powerful tools to probe structure-property relationships in crystals and molecules. Specifically, machine learning interatomic potentials (MLIPs) can accurately reproduce first-principles data at a cost similar to that of conventional interatomic potential approaches. While MLIPs have been extensively tested across various classes of materials and molecules, a clear characterization of the anharmonic terms encoded in the MLIPs is lacking. Here, we benchmark popular MLIPs using the anharmonic vibrational Hamiltonian of ThO 2 in the fluorite crystal structure, which was constructed from density functional theory (DFT) using our highly accurate and efficient irreducible derivative methods. The anharmonic Hamiltonian was used to generate molecular dynamics (MD) trajectories, which were used to train three classes of MLIPs: Gaussian approximation potentials, artificial neural networks (ANN), and graph neural networks (GNN). The results were assessed by directly comparing phonons and their interactions, as well as phonon linewidths, phonon lineshifts, and thermal conductivity. The models were also trained on a DFT MD dataset, demonstrating good agreement up to fifth-order for the ANN and GNN. Our analysis demonstrates that MLIPs have great potential for accurately characterizing anharmonicity in materials systems at a fraction of the cost of conventional first principles-based approaches.

interatomic potentials↗
DOE OSTI2026

Forest aboveground biomass estimation through integration of sentinel-2 and PALSAR-2 time series: assessing models trained on GEDI and field inventory benchmarks

Accurate and spatially explicit forest Aboveground Biomass (AGB) mapping through remote sensing is critical for quantifying terrestrial carbon stocks and informing effective forest management strategies. However, AGB estimation in dense forests with complex terrain remains challenging due to satellite sensor signal saturation problem (saturation issue occurs in high biomass forests), structural complexity, and limited ground truth for calibration. This study presents a novel framework that integrates multi-temporal Sentinel-2 optical imagery, ALOS PALSAR-2 Synthetic Aperture Radar (SAR) data, and topographic variables with explainable Machine Learning to map AGB across mountainous forests within subtropical and temperate oceanic climate zones of Mexico. We evaluate the effects of temporal granularity and sensor synergy by comparing multiple temporal inputs and sensor configurations (Sentinel-2, PALSAR-2, and their fusion), and assess model performance using two reference datasets: NASA GEDI LiDAR-derived biomass and Mexico’s National Forest and Soil Inventory (INFyS). Our results showed that models trained on INFyS consistently outperformed those trained on GEDI, highlighting limitations in GEDI’s reliability in biomass estimates within this study region. Furthermore, the integration of Sentinel-2 and PALSAR-2 provided improved predictions compared to single-sensor models, particularly when combined with temporally explicit yearly statistics. The best-performing model, which was trained on INFyS data, and considered both Sentinel-2 and PALSAR-2 yearly statistics, as well as topographic variables, achieved an R2 of 0.64, RMSE of 51.10 Mg/ha, and relative RMSE (rRMSE) of 58.69%. Explainable ML analysis identified Sentinel-2 spectral indices and topographic features as key predictors, while PALSAR-2 metrics provided complementary information, partially mitigating saturation effects in high-biomass areas. Specifically, integrating both sensors substantially improved AGB estimation in high biomass forest (≥200 Mg/ha), yielding 98% gains over optical-only model, with resulting estimates exceeding GEDI L4B by 29% and ESA-CCI-BIOMASS by 174%. Terrain-stratified analysis indicated close agreement with GEDI in low-slope areas, with increasing divergence as slope steepness increased, while estimates remained consistently higher than ESA-CCI-BIOMASS across all slope classes. The proposed approach advances multi-sensor fusion and temporal feature engineering for AGB mapping using open-access satellite datasets, providing a scalable and reproducible framework for annual biomass monitoring in topographically complex mountainous forests. The resulting 25 m resolution biomass product has the potential to provide spatially detailed information for forest monitoring and may support applications in carbon accounting and forest management.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Reverse-mode differentiation in arbitrary tensor network format: with application to supervised learning.

This paper describes an efficient reverse-mode differentiation algorithm for contraction operations of tensor networks that may have arbitrary and unconventional network topologies. The approach leverages the tensor contraction tree of Evenbly and Pfeifer (2014), which provides an instruction set for the contraction sequence of a network. We show that this tree can be efficiently leveraged for differentiation of a full tensor network contraction using a recursive scheme that exploits (1) the bilinear property of contraction and (2) the property that trees have single path from root to leaves. While differentiation of tensor-tensor contraction is already possible in most automatic differentiation packages, we show that exploiting these two additional properties in the specific context of contraction sequences can improve efficiency. Following a description of the algorithm and computational complexity analysis, we investigate its utility for gradient-based supervised learning for low-rank function recovery and for fitting real-world unstructured datasets. We demonstrate improved performance over alternating least-squares optimization approaches and the capability to handle heterogeneous and arbitrary tensor network formats. When compared to alternating minimization algorithms, we find that the gradient-based approach requires a smaller oversampling ratio (number of samples compared to number model parameters) for recovery. This increased efficiency extends to fitting unstructured data of varying dimensionality and when employing a variety of tensor network formats. Here, we show improved learning using the hierarchical Tucker method over the tensor-train in high-dimensional settings on a number of benchmark problems.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Assessment of the ParFlow–CLM CONUS 1.0 integrated hydrologic model: evaluation of hyper-resolution water balance components across the contiguous United States

Abstract. Recent advancements in computational efficiency and Earth system modeling have awarded hydrologists with increasingly high-resolution models of terrestrial hydrology, which are paramount to understanding and predicting complex fluxes of moisture and energy. Continental-scale hydrologic simulations are, in particular, of interest to the hydrologic community for numerous societal, scientific, and operational benefits. The coupled hydrology–land surface model ParFlow–CLM configured over the continental United States (PFCONUS) has been employed in previous literature to study scale-dependent connections between water table depth, topography, recharge, and evapotranspiration, as well as to explore impacts of anthropogenic aquifer depletion to the water and energy balance. These studies have allowed for an unprecedented process-based understanding of the continental water cycle at high resolution. Here, we provide the most comprehensive evaluation of PFCONUS version 1.0 (PFCONUSv1) performance to date by comparing numerous modeled water balance components with thousands of in situ observations and several remote sensing products and using a range of statistical performance metrics for evaluation. PFCONUSv1 comparisons with these datasets are a promising indicator of model fidelity and ability to reproduce the continental-scale water balance at high resolution. Areas for improvement are identified, such as a positive streamflow bias at gauges in the eastern Great Plains, a shallow water table bias over many areas of the model domain, and low bias in seasonal total water storage amplitude, especially for the Ohio, Missouri, and Arkansas River basins. We discuss several potential sources for model bias and suggest that minimizing error in topographic processing and meteorological forcing would considerably improve model performance. Results here provide a benchmark and guidance for further PFCONUS model development, and they highlight the importance of concurrently evaluating all hydrologic components and fluxes to provide a multivariate, holistic validation of the complete modeled water balance.

O'Neill, Mary M. F. (ORCID:0000000221455165)↗
DOE OSTI2023

Calibration strategy of the PROSPECT-II detector with external and intrinsic sources

This paper presents an energy calibration scheme for an upgraded reactor antineutrino detector for the Precision Reactor Oscillation and Spectrum Experiment (PROSPECT). The PROSPECT collaboration is preparing an upgraded detector, PROSPECT-II (P-II), to advance capabilities for the investigation of fundamental neutrino physics, fission processes and associated reactor neutrino flux, and nuclear security applications. P-II will expand the statistical power of the original PROSPECT (P-I) dataset by at least an order of magnitude. The new design builds upon previous P-I design and focuses on improving the detector robustness and long-term stability to enable multi-year operation at one or more sites. The new design optimizes the fiducial volume by elimination of dead space previously occupied by internal calibration channels, which in turn necessitates the external deployment. In this paper, we describe a calibration strategy for P-II. Here, the expected performance of externally deployed calibration sources is evaluated using P-I data and a well-benchmarked simulation package by varying detector segmentation configurations in the analysis. The proposed external calibration scheme delivers a compatible energy scale model and achieves comparable performance with the inclusion of an additional AmBe neutron source, in comparison to the previous internal arrangement. Most importantly, the estimated uncertainty contribution from the external energy scale calibration model meets the precision requirements of the P-II experiment.

47 OTHER INSTRUMENTATION↗
DOE OSTI2025

UBW (USLCI-Brightway2) [SWR-25-169]

Life cycle inventory (LCI) data are critical for robust life cycle assessment (LCA), yet many widely used datasets such as the U.S. Life Cycle Inventory (USLCI) are not natively compatible with advanced modeling frameworks like Brightway2. This work presents an automated pipeline to transform USLCI data into a fully functional Brightway2 project. The workflow performs systematic data cleaning, resolves duplicate process and exchange identifiers, and applies allocation to multi-output processes. Technosphere and biosphere flows are harmonized through unit conversions and a bridge mapping to the biosphere3 database, with comprehensive logging of missing flows and cutoff issues. The resulting Brightway2 database is validated using matrix diagnostics to ensure consistency of the technosphere, and is benchmarked via life cycle impact assessment (LCIA) methods such as ReCiPe and IPCC GWP. Outputs include reproducible CSV exports of corrected processes, elementary flows, characterization factors, and LCIA results, alongside backup utilities for project sharing. This pipeline lowers barriers for integrating USLCI data into open-source LCA workflows, enabling reproducible, validated LCA inventories within the Brightway 2 framework.

Ghosh, Tapajyoti [National Laboratory of the Rocki↗
DOE OSTI2025

Conditional Pseudo-Reversible Normalizing Flow for Surrogate Modeling in Quantifying Uncertainty Propagation

We introduce a conditional pseudo-reversible normalizing flow (PR-NF) that directly learns conditional probability distributions from noisy physical models to efficiently quantify both forward and inverse uncertainty propagation. Traditional surrogate modeling approaches approximate only the deterministic component of physical models, requiring separate noise characterization and computationally expensive sampling methods for inverse problems. Here, in this work, we develop the conditional PR-NF model to directly learn and efficiently generate samples from the conditional probability density functions (PDFs). The training process utilizes dataset consisting of input-output pairs without requiring prior knowledge about the noise and the function. Once trained, our model efficiently generates samples from conditional PDFs for any input within the training domain. Moreover, the pseudo-reversibility feature allows for the use of fully connected neural network architectures, which simplifies the implementation and enables theoretical analysis. We provide a rigorous convergence analysis of the conditional PR-NF model, showing its ability to converge to the target conditional PDF using the Kullback−Leibler divergence. To demonstrate the effectiveness of our method, we apply it to several benchmark tests and a real-world geologic carbon storage problem.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Benchmarking Demand Flexibility in Commercial Buildings and Flattening the Duck – Addressing Baseline and Commissioning Challenges

With the transition from our traditional electric grid to a cleaner grid with renewable power generation, there is a need to enable building loads to be flexible. Load shedding and shifting will be essential for flattening the “Duck” for decarbonization. This paper explored the trend in the timing of DR events as a reflection of the grid’s needs using recent four years of event data from 203 retail stores in 11 states. The events are becoming significantly shorter with 2-hour duration being the most popular; shifting to late afternoon and early evening is another trend beyond California. Benchmarking will be essential for accounting DF as a reliable grid resource. This paper addresses a challenging aspect of benchmarking – inaccuracies in counterfactual baseline methods can introduce significant DF metrics variations in addition to weather and building characteristics related factors. The conventional “10/10” with adjustment baseline method has inherent limitation by design for load shifting applications. Therefore, it is imperative to identify alternative methods. This study compared three hourly regression baseline methods with “10/10” methods using two groups of commercial buildings that participated in DR programs: (1) 121 big-box retail stores, and (2) 11 office buildings in CA. The 14-day hourly outdoor temperature regression method was found to produce least error in the tested datasets and is promising for load shifting. The paper also pointed out that commissioning issues can also be a significant barrier for achieving consistent DF performance, which building managers and utilities should be aware of.

Liu, Jingjing↗
DOE OSTI2024

Energy dataset of Frontier supercomputer for waste heat recovery

The Hewlett Packard Enterprise–Cray EX Frontier is the world’s first and fastest exascale supercomputer, hosted at the Oak Ridge Leadership Computing Facility in Tennessee, United States. Frontier is a significant electricity consumer, drawing 8–30 MW; this massive energy demand produces significant waste heat, requiring extensive cooling measures. Although harnessing this waste heat for campus heating is a sustainability goal at Oak Ridge National Laboratory (ORNL), the 30 °C–38 °C waste heat temperature poses compatibility issues with standard HVAC systems. Heat pump systems, prevalent in residential settings and some industries, can efficiently upgrade low-quality heat to usable energy for buildings. Thus, heat pump technology powered by renewable electricity offers an efficient, cost-effective solution for substantial waste heat recovery. However, a major challenge is the absence of benchmark data on high-performance computing (HPC) heat generation and waste heat profiles. This paper reports power demand and waste heat measurements from an ORNL HPC data centre, aiming to guide future research on optimizing waste heat recovery in large-scale data centres, especially those of HPC calibre.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2024

A Comparison between Invariant and Equivariant Classical and Quantum Graph Neural Networks

Machine learning algorithms are heavily relied on to understand the vast amounts of data from high-energy particle collisions at the CERN Large Hadron Collider (LHC). The data from such collision events can naturally be represented with graph structures. Therefore, deep geometric methods, such as graph neural networks (GNNs), have been leveraged for various data analysis tasks in high-energy physics. One typical task is jet tagging, where jets are viewed as point clouds with distinct features and edge connections between their constituent particles. The increasing size and complexity of the LHC particle datasets, as well as the computational models used for their analysis, have greatly motivated the development of alternative fast and efficient computational paradigms such as quantum computation. In addition, to enhance the validity and robustness of deep networks, we can leverage the fundamental symmetries present in the data through the use of invariant inputs and equivariant layers. In this paper, we provide a fair and comprehensive comparison of classical graph neural networks (GNNs) and equivariant graph neural networks (EGNNs) and their quantum counterparts: quantum graph neural networks (QGNNs) and equivariant quantum graph neural networks (EQGNN). The four architectures were benchmarked on a binary classification task to classify the parton-level particle initiating the jet. Based on their area under the curve (AUC) scores, the quantum networks were found to outperform the classical networks. However, seeing the computational advantage of quantum networks in practice may have to wait for the further development of quantum technology and its associated application programming interfaces (APIs).

Forestano, Roy T. (ORCID:0000000203552076)↗
DOE OSTI2024

A combinatorially complete epistatic fitness landscape in an enzyme active site

Protein engineering often targets binding pockets or active sites which are enriched in epistasis—nonadditive interactions between amino acid substitutions—and where the combined effects of multiple single substitutions are difficult to predict. Few existing sequence-fitness datasets capture epistasis at large scale, especially for enzyme catalysis, limiting the development and assessment of model-guided enzyme engineering approaches. We present here a combinatorially complete, 160,000-variant fitness landscape across four residues in the active site of an enzyme. Assaying the native reaction of a thermostable β-subunit of tryptophan synthase (TrpB) in a nonnative environment yielded a landscape characterized by significant epistasis and many local optima. These effects prevent simulated directed evolution approaches from efficiently reaching the global optimum. There is nonetheless wide variability in the effectiveness of different directed evolution approaches, which together provide experimental benchmarks for computational and machine learning workflows. The most-fit TrpB variants contain a substitution that is nearly absent in natural TrpB sequences—a result that conservation-based predictions would not capture. Thus, although fitness prediction using evolutionary data can enrich in more-active variants, these approaches struggle to identify and differentiate among the most-active variants, even for this near-native function. Overall, this work presents a large-scale testing ground for model-guided enzyme engineering and suggests that efficient navigation of epistatic fitness landscapes can be improved by advances in both machine learning and physical modeling.

biocatalysis↗
DOE OSTI2023

Search for single production of vector-like T quarks decaying into Ht or Zt in pp collisions at $\sqrt{s}$ = 13 TeV with the ATLAS detector

This paper describes a search for the single production of an up-type vector-like quark (T) decaying as T → Ht or T → Zt. The search utilises a dataset of pp collisions at $\sqrt{s}$ = 13 TeV collected with the ATLAS detector during the 2015–2018 data-taking period of the Large Hadron Collider, corresponding to an integrated luminosity of 139 fb ₋1 . Data are analysed in final states containing a single lepton with multiple jets and b-jets. The presence of boosted heavy resonances in the event is exploited to discriminate the signal from the Standard Model background. No significant excess above the Standard Model expectation is observed, and 95% CL upper limits are set on the production cross section of T quarks in different decay channels. The results are interpreted in several benchmark scenarios to set limits on the mass and universal coupling strength (κ) of the vector-like quark. For singlet T quarks, κ values above 0.53 are excluded for all masses below 2.3 TeV. At a mass of 1.6 TeV, κ values as low as 0.35 are excluded. For T quarks in the doublet scenario, where the production cross section is much lower, κ values above 0.72 are excluded for all masses below 1.7 TeV, and this exclusion is extended to κ above 0.55 for low masses around 1.0 TeV.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2024

Community Geothermal: Energy, Cost, and Carbon Modeling for District Design - Ann Arbor, MI

This data includes results on an analysis of existing and projected energy, cost, and carbon for the City of Ann Arbor - District Geothermal Design and Deployment to Equitably Decarbonize Low Income Neighborhoods in Ann Arbor project. The scope of the project includes designing and implementing a geothermal district heating and cooling system that reduces thermal heating and cooling load by 75% and greenhouse gas emissions by 40% in the project area (262 households, 6 commercial buildings). The existing neighborhood was modeled using Design Builder, an EnergyPlus software, to understand the current energy load. The energy model was then flipped to reflect the designed district geothermal heating and cooling system to project the effect on energy, carbon, and cost. This dataset includes the analysis files utilized and created for this study. There are 3 categories of data: 1) existing/benchmarking, 2) energy modeling, and 3) post processed calculations. This follows the methodology and process of the project team, which is fully explained in file 00_Technical Economic Environmental Assessment. All uses of data are referenced throughout this assessment to their respective files included below.

15 GEOTHERMAL ENERGY↗
DOE OSTI2020

Predicting molecular dipole moments by combining atomic partial charges and atomic dipoles

The molecular dipole moment ( μ ) is a central quantity in chemistry. It is essential in predicting infrared and sum-frequency generation spectra as well as induction and long-range electrostatic interactions. Furthermore, it can be extracted directly—via the ground state electron density—from high-level quantum mechanical calculations, making it an ideal target for machine learning (ML). Here, we choose to represent this quantity with a physically inspired ML model that captures two distinct physical effects: local atomic polarization is captured within the symmetry-adapted Gaussian process regression framework which assigns a (vector) dipole moment to each atom, while the movement of charge across the entire molecule is captured by assigning a partial (scalar) charge to each atom. The resulting “MuML” models are fitted together to reproduce molecular μ computed using high-level coupled-cluster theory and density functional theory (DFT) on the QM7b dataset, achieving more accurate results due to the physics-based combination of these complementary terms. The combined model shows excellent transferability when applied to a showcase dataset of larger and more complex molecules, approaching the accuracy of DFT at a small fraction of the computational cost. We also demonstrate that the uncertainty in the predictions can be estimated reliably using a calibrated committee model. The ultimate performance of the models—and the optimal weighting of their combination—depends, however, on the details of the system at hand, with the scalar model being clearly superior when describing large molecules whose dipole is almost entirely generated by charge separation. These observations point to the importance of simultaneously accounting for the local and non-local effects that contribute to μ ; furthermore, they define a challenging task to benchmark future models, particularly those aimed at the description of condensed phases.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

HAPPA: A Modular Platform for HPC Application Resilience Analysis with LLMs Embedded

High-performance computing (HPC) systems are increasingly vulnerable to soft errors, which pose significant challenges in maintaining computational accuracy and reliability. Predicting the resilience of HPC applications to these errors is crucial for robust code protection and detailed resilience analysis. In this study, we present HAppA, a modular platform designed for HPC Application Resilience Analysis. Embedding Large Language Models (LLMs), HAppA addresses understanding the context information of long code sequences typical in HPC applications. HAppA implements a novel code representation module that chunks the code into fixed-size segments and aggregates the embeddings of these segments. Three aggregation methods have been explored: MeanPooling, MaxPooling, and LSTM-based techniques. We built a DAtaset for REsilience analysis using Fault Injection (FI), named DARE. Using our DARE dataset, HAppA is trained for regression prediction tasks. Our evaluation results demonstrate the predictive accuracy of HAppA compared to other models, particularly noting that the LSTM-based aggregation method -- HAppA-LSTM -- achieves a mean squared error (MSE) of 0.078 for SDC prediction, surpassing the existing state-of-the-art PARIS model, which recorded an MSE of 0.1172. Additionally, HAppA with the KeyBERT model extracts a list of keywords representing the source code. A comprehensive importance analysis of these keywords further elucidates the code patterns contributing to the error rate. These findings highlight the effectiveness of HAppA in analyzing the resilience of HPC applications and establish a new benchmark for predictive accuracy in resilience.

Jiang, Hailong [Kent State University]↗