Search NASA⌕ Search

SEARCH · Search NASA

Results for “Contrastive learning”

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 217 records · Page 12
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

An Evaluation of Representation Learning Methods in Particle Physics Foundation Models

We present a systematic evaluation of representation learning objectives for particle physics within a unified framework. Our study employs a shared transformer-based particle-cloud encoder with standardized preprocessing, matched sampling, and a consistent evaluation protocol on a jet classification dataset. We compare contrastive (supervised and self-supervised), masked particle modeling, and generative reconstruction objectives under a common training regimen. In addition, we introduce targeted supervised architectural modifications that achieve state-of-the-art performance on benchmark evaluations. This controlled comparison isolates the contributions of the learning objective, highlights their respective strengths and limitations, and provides reproducible baselines. We position this work as a reference point for the future development of foundation models in particle physics, enabling more transparent and robust progress across the community.

Chen, Michael [Caltech]↗
DOE OSTI2023

Context-aware learning of hierarchies of low-fidelity models for multi-fidelity uncertainty quantification

Multi-fidelity Monte Carlo methods leverage low-fidelity and surrogate models for variance reduction to make tractable uncertainty quantification even when numerically simulating the physical systems of interest with high-fidelity models is computationally expensive. This work proposes a context-aware multi-fidelity Monte Carlo method that optimally balances the costs of training low-fidelity models with the costs of Monte Carlo sampling. It generalizes the previously developed context-aware bi-fidelity Monte Carlo method to hierarchies of multiple models and to more general types of low-fidelity models. When training low-fidelity models, the proposed approach takes into account the context in which the learned low-fidelity models will be used, namely for variance reduction in Monte Carlo estimation, which allows it to find optimal trade-offs between training and sampling to minimize upper bounds of the mean-squared errors of the estimators for given computational budgets. This is in stark contrast to traditional surrogate modeling and model reduction techniques that construct low-fidelity models with the primary goal of approximating well the high-fidelity model outputs and typically ignore the context in which the learned models will be used in upstream tasks. Further, the proposed context-aware multi-fidelity Monte Carlo method applies to hierarchies of a wide range of types of low-fidelity models such as sparse-grid and deep-network models. Numerical experiments with the gyrokinetic simulation code Gene show speedups of up to two orders of magnitude compared to standard estimators when quantifying uncertainties in small-scale fluctuations in confined plasma in fusion reactors. This corresponds to a runtime reduction from 72 days to four hours on one node of the Lonestar6 supercomputer at the Texas Advanced Computing Center.

42 ENGINEERING↗
DOE OSTI2023

An interlaboratory comparison of mid-infrared spectra acquisition: Instruments and procedures matter

Diffuse reflectance spectroscopy has been extensively employed to deliver timely and cost-effective predictions of a number of soil properties. However, although several soil spectral laboratories have been established worldwide, the distinct characteristics of instruments and operations still hamper further integration and interoperability across mid-infrared (MIR) soil spectral libraries. In this study, we conducted a large-scale ring trial experiment to understand the lab-to-lab variability of multiple MIR instruments. By developing a systematic evaluation of different mathematical treatments with modeling algorithms, including regular preprocessing and spectral standardization, we quantified and evaluated instruments' dissimilarity and how this impacts internal and shared model performance. We found that all instruments delivered good predictions when calibrated internally using the same instruments' characteristics and standard operating procedures by solely relying on regular spectral preprocessing that accounts for light scattering and multiplicative/additive effects, e.g., using standard normal variate (SNV). When performing model transfer from a large public library (the USDA NSSCKSSL MIR library) to secondary instruments, good performance was also achieved by regular preprocessing (e. g., SNV) if both instruments shared the same manufacturer. However, significant differences between the KSSL MIR library and contrasting ring trial instruments responses were evident and confirmed by a semi-unsupervised spectral clustering. For heavily contrasting setups, spectral standardization was necessary before transferring prediction models. Non-linear model types like Cubist and memory-based learning delivered more precise estimates because they seemed to be less sensitive to spectral variations than global partial least square regression. In summary, the results from this study can assist new laboratories in building spectroscopy capacity utilizing existing MIR spectral libraries and support the recent global efforts to make soil spectroscopy universally accessible with centralized or shared operating procedures.

58 GEOSCIENCES↗
NASA NTRS2007

CFD to Flight: Some Recent Success Stories of X-plane Design to Flight Test at the NASA Dryden Flight Research Center

Several examples from the past decade of success stories involving the design and ight test of three true X-planes will be described: in particular, X-plane design techniques that relied heavily upon computational fluid dynamics (CFD). Three specific examples chosen from the authors personal experience are presented: the X-36 Tailless Fighter Agility Research Aircraft, the X-45A Unmanned Combat Air Vehicle, and, most recently, the X-48B Blended Wing Body Demonstrator Aircraft. An overview will be presented of the uses of CFD analysis, comparisons and contrasts with wind tunnel testing, and information derived from the CFD analysis that directly related to successful flight test. Some lessons learned on the proper application, and misapplication, of CFD are illustrated. Finally, some highlights of the flight-test results of the three example X-planes will be presented. This overview paper will discuss some of the authors experience with taking an aircraft shape from early concept and three-dimensional modeling through CFD analysis, wind tunnel testing, further re ned CFD analysis, and, finally, flight. An overview of the key roles in which CFD plays well during this process, and some other roles in which it does not, are discussed. How wind tunnel testing complements, calibrates, and verifies CFD analysis is also covered. Lessons learned on where CFD results can be misleading are also given. Strengths and weaknesses of the various types of ow solvers, including panel methods, Euler, and Navier-Stokes techniques, are discussed. The paper concludes with the three specific examples, including some flight test video footage of the X-36, the X-45A, and the X-48B.

Cosentino, Gary B.↗
NASA NTRS2007

CFD to Flight: Some Recent Success Stories of X-Plane Design to Flight Test at the NASA Dryden Flight Research Center

Several examples from the past decade of success stories involving the design and flight test of three true X-planes will be described: in particular, X-plane design techniques that relied heavily upon computational fluid dynamics (CFD). Three specific examples chosen from the author s personal experience are presented: the X-36 Tailless Fighter Agility Research Aircraft, the X-45A Unmanned Combat Air Vehicle, and, most recently, the X-48B Blended Wing Body Demonstrator Aircraft. An overview will be presented of the uses of CFD analysis, comparisons and contrasts with wind tunnel testing, and information derived from the CFD analysis that directly related to successful flight test. Some lessons learned on the proper application, and misapplication, of CFD are illustrated. Finally, some highlights of the flight-test results of the three example X-planes will be presented. This overview paper will discuss some of the author s experience with taking an aircraft shape from early concept and three-dimensional modeling through CFD analysis, wind tunnel testing, further refined CFD analysis, and, finally, flight. An overview of the key roles in which CFD plays well during this process, and some other roles in which it does not, are discussed. How wind tunnel testing complements, calibrates, and verifies CFD analysis is also covered. Lessons learned on where CFD results can be misleading are also given. Strengths and weaknesses of the various types of flow solvers, including panel methods, Euler, and Navier-Stokes techniques, are discussed. The paper concludes with the three specific examples, including some flight test video footage of the X-36, the X-45A, and the X-48B.

Cosentino, Gary B.↗
NASA NTRS2009

Active Learning with Irrelevant Examples

An improved active learning method has been devised for training data classifiers. One example of a data classifier is the algorithm used by the United States Postal Service since the 1960s to recognize scans of handwritten digits for processing zip codes. Active learning algorithms enable rapid training with minimal investment of time on the part of human experts to provide training examples consisting of correctly classified (labeled) input data. They function by identifying which examples would be most profitable for a human expert to label. The goal is to maximize classifier accuracy while minimizing the number of examples the expert must label. Although there are several well-established methods for active learning, they may not operate well when irrelevant examples are present in the data set. That is, they may select an item for labeling that the expert simply cannot assign to any of the valid classes. In the context of classifying handwritten digits, the irrelevant items may include stray marks, smudges, and mis-scans. Querying the expert about these items results in wasted time or erroneous labels, if the expert is forced to assign the item to one of the valid classes. In contrast, the new algorithm provides a specific mechanism for avoiding querying the irrelevant items. This algorithm has two components: an active learner (which could be a conventional active learning algorithm) and a relevance classifier. The combination of these components yields a method, denoted Relevance Bias, that enables the active learner to avoid querying irrelevant data so as to increase its learning rate and efficiency when irrelevant items are present. The algorithm collects irrelevant data in a set of rejected examples, then trains the relevance classifier to distinguish between labeled (relevant) training examples and the rejected ones. The active learner combines its ranking of the items with the probability that they are relevant to yield a final decision about which item to present to the expert for labeling. Experiments on several data sets have demonstrated that the Relevance Bias approach significantly decreases the number of irrelevant items queried and also accelerates learning speed.

Wagstaff, Kiri↗
DOE OSTI2023

Few-Shot Learning Enables Population-Scale Analysis of Leaf Traits in Populus trichocarpa

Plant phenotyping is typically a time-consuming and expensive endeavor, requiring large groups of researchers to meticulously measure biologically relevant plant traits, and is the main bottleneck in understanding plant adaptation and the genetic architecture underlying complex traits at population scale. In this work, we address these challenges by leveraging few-shot learning with convolutional neural networks to segment the leaf body and visible venation of 2,906 Populus trichocarpa leaf images obtained in the field. In contrast to previous methods, our approach (a) does not require experimental or image preprocessing, (b) uses the raw RGB images at full resolution, and (c) requires very few samples for training (e.g., just 8 images for vein segmentation). Traits relating to leaf morphology and vein topology are extracted from the resulting segmentations using traditional open-source image-processing tools, validated using real-world physical measurements, and used to conduct a genome-wide association study to identify genes controlling the traits. In this way, the current work is designed to provide the plant phenotyping community with (a) methods for fast and accurate image-based feature extraction that require minimal training data and (b) a new population-scale dataset, including 68 different leaf phenotypes, for domain scientists and machine learning researchers. All of the few-shot learning code, data, and results are made publicly available.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Data-Driven Kinetic Reaction Networks for Separation Chemistry

Understanding complex, multistep chemical reactions at the molecular level is a major challenge whose solution would greatly benefit the design and optimization of numerous chemical processes. The separation of rare-earth (4f) and actinide (5f) elements is an example where improving our chemical understanding is important for designing and optimizing new chemistries, even with a limited number of observations. Here, in this work, we leverage data-driven artificial intelligence and machine-learning approaches to develop kinetic reaction networks that describe the liquid–liquid extraction mechanism of uranium using N,N-di-2-ethylhexyl-isobutyramide (DEHiBA). Specifically, we compare and contrast the properties of two classes of models: (1) purely data-driven models that are regularized using chemistry-agnostic, L1 regression and (2) chemistry-informed models that are regularized using relative reaction energies provided by quantum mechanical calculations. We observe that purely data-driven models are unbiased, simple, and accurate in their predictions of experimental measurements when provided with sufficient data but are difficult to fully constrain and interpret. In contrast, chemistry-informed models exhibit significantly improved chemical interpretability and consistency, providing a detailed description of the separation process while achieving high accuracy through ensemble averaging. Overall, the dominant species predicted to be extracted into the organic phase is UO 2 (NO 3 ) 2 (DEHiBA) 2 , agreeing with experimental slope analysis, thermodynamic modeling, EXAFS, and crystal structures. This work demonstrates that leveraging the fundamental structure of the problem can lead to efficient learning schemes that provide both accurate predictions and chemical insights at a low computational cost.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
NASA NTRS1997

Analytical, Experimental, and Modelling Studies of Lunar and Terrestrial Rocks

The goal of our research has been to understand the paths and the processes of planetary evolution that produced planetary surface materials as we find them. Most of our work has been on lunar materials and processes. We have done studies that obtain geological knowledge from detailed examination of regolith materials and we have reported implications for future sample-collecting and on-surface robotic sensing missions. Our approach has been to study a suite of materials that we have chosen in order to answer specific geologic questions. We continue this work under NAG5-4172. The foundation of our work has been the study of materials with precise chemical and petrographic analyses, emphasizing analysis for trace chemical elements. We have used quantitative models as tests to account for the chemical compositions and mineralogical properties of the materials in terms of regolith processes and igneous processes. We have done experiments as needed to provide values for geochemical parameters used in the models. Our models take explicitly into account the physical as well as the chemical processes that produced or modified the materials. Our approach to planetary geoscience owes much to our experience in terrestrial geoscience, where samples can be collected in field context and sampling sites revisited if necessary. Through studies of terrestrial analog materials, we have tested our ideas about the origins of lunar materials. We have been mainly concerned with the materials of the lunar highland regolith, their properties, their modes of origin, their provenance, and how to extrapolate from their characteristics to learn about the origin and evolution of the Moon's early igneous crust. From this work a modified model for the Moon's structure and evolution is emerging, one of globally asymmetric differentiation of the crust and mantle to produce a crust consisting mainly of ferroan and magnesian igneous rocks containing on average 70-80% plagioclase, with a large, mafic, trace-element-rich geochemical province, and a regolith that globally contains trace-element-rich material distributed from this province by the Imbrium basin-forming impact. This contrasts with earlier models of a concentrically zoned Moon with a crust of ferroan anorthosite overlying a layer of urKREEP overlying ultramafic cumulates. From this work, we have learned lessons useful for developing strategies for studying regolith materials that help to maximize the information available about both the evolution of the regolith and the igneous differentiation of the planet. We believe these lessons are useful in developing strategies for on-surface geological, mineralogical, and geochemical studies, as well. The main results of our work are given in the following brief summaries of major tasks. Detailed accounts of these results have been submitted in the annual progress reports.

Haskin, Larry A.↗
DOE OSTI2022

Empirical relationships between environmental factors and soil organic carbon produce comparable prediction accuracy as the Machine Learning

Accurate representation of environmental controllers of soil organic carbon (SOC) stocks in Earth System Model (ESM) land models could reduce uncertainties in future carbon-climate feedback projections. Using empirical relationships between environmental factors and SOC stocks to evaluate land models can help modelers understand prediction biases beyond what can be achieved with the observed SOC stocks alone. In this study, we used 31 observed environmental factors, field SOC observations (n = 6,213) from the continental US, and two Machine Learning approaches [Random Forest (RF) and Generalized Additive Modeling (GAM)] to (1) select important environmental predictors of SOC stocks, (2) derive empirical relationships between environmental factors and SOC stocks, and (3) use the derived relationships to predict SOC stocks and compare the prediction accuracy of simpler model developed with the machine learning predictions. Out of the 31 environmental factors we investigated, 12 were identified as important predictors of SOC stocks by the RF approach. In contrast, the GAM approach identified six (of those 12) environmental factors as important controllers of SOC stocks: potential evapotranspiration, normalized difference vegetation index, soil drainage condition, precipitation, elevation, and net primary productivity. The GAM approach showed minimal SOC predictive importance of the remaining six environmental factors identified by the RF approach. Our derived empirical relations produced comparable prediction accuracy as the GAM and RF approach using only a subset of environmental factors. The empirical relationships we derived using the GAM approach can serve as important benchmarks to evaluate environmental control representations of SOC stocks in ESMs, which could reduce uncertainty in predicting future carbon-climate feedbacks.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2022

Safety-assured, real-time neural active fault management for resilient microgrids integration

Federated-learning-based active fault management (AFM) is devised to achieve real-time safety assurance for microgrids and the main grid during faults. AFM was originally formulated as a distributed optimization problem. Here, federated learning is used to train each microgrid's network with training data achieved from distributed optimization. The main contribution of this work is to replace the optimization-based AFM control algorithm with a learning-based AFM control algorithm. The replacement transfers computation from online to offline. With this replacement, the control algorithm can meet real-time requirements for a system with dozens of microgrids. By contrast, distributed-optimization-based fault management can output reference values fast enough for a system with several microgrids. More microgrids, however, lead to more computation time with optimization-based method. Distributed-optimization-based fault management would fail real-time requirements for a system with dozens of microgrids. Controller hardware-in-the-loop real-time simulations demonstrate that learning-based AFM can output reference values within 10 ms irrespective of the number of microgrids.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2024

Voltage Mining for (De)lithiation-Stabilized Cathodes and a Machine Learning Model for Li-Ion Cathode Voltage

Advances in lithium-metal anodes have inspired interest in discovery of Li-free cathodes, most of which are natively found in their charged state. This is in contrast to today's commercial lithium-ion battery cathodes, which are more stable in their discharged state. In this study, we combine calculated cathode voltage information from both categories of cathode materials, covering 5577 and 2423 total unique structure pairs, respectively. The resulting voltage distributions with respect to the redox pairs and anion types for both classes of compounds emphasize design principles for high-voltage cathodes, which favor later Period 4 transition metals in their higher oxidation states and more electronegative anions like fluorine or polyanion groups. Generally, cathodes that are found in their charged, delithiated state are shown to exhibit voltages lower than those that are most stable in their lithiated state, in agreement with thermodynamic expectations. Deviations from this trend are found to originate from different anion distributions between redox pairs. In addition, a machine learning model for voltage prediction based on chemical formulas is trained and shows state-of-the-art performance when compared to two established composition-based ML models for material properties predictions, Roost and CrabNet.

25 ENERGY STORAGE↗
DOE OSTI2020

Deep learning of interface structures from simulated 4D STEM data: cation intermixing vs. roughening ∗

Abstract Interface structures in complex oxides remain an active area of condensed matter physics research, largely enabled by recent advances in scanning transmission electron microscopy (STEM). Yet the nature of the STEM contrast in which the structure is projected along the given direction precludes separation of possible structural models. Here, we utilize deep convolutional neural networks (DCNN) trained on simulated 4D STEM datasets to predict structural descriptors of interfaces. We focus on the widely studied interface between LaAlO 3 and SrTiO 3 , using dynamical diffraction theory and leveraging high performance computing to simulate thousands of possible 4D STEM datasets to train the DCNN to learn properties of the underlying structures on which the simulations are based. We test the DCNN on simulated data and show that it is possible (with >95% accuracy) to identify a physically rough from a chemically diffuse interface and create a DCNN regression model to predict step positions. We quantify the applicability of the model to different thicknesses and the transferability of the approach. The method shown here is general and can be applied for any inverse imaging problem where forward models are present.

42 ENGINEERING↗
NASA NTRS2015

Advanced Bayesian Method for Planetary Surface Navigation

Autonomous Exploration, Inc., has developed an advanced Bayesian statistical inference method that leverages current computing technology to produce a highly accurate surface navigation system. The method combines dense stereo vision and high-speed optical flow to implement visual odometry (VO) to track faster rover movements. The Bayesian VO technique improves performance by using all image information rather than corner features only. The method determines what can be learned from each image pixel and weighs the information accordingly. This capability improves performance in shadowed areas that yield only low-contrast images. The error characteristics of the visual processing are complementary to those of a low-cost inertial measurement unit (IMU), so the combination of the two capabilities provides highly accurate navigation. The method increases NASA mission productivity by enabling faster rover speed and accuracy. On Earth, the technology will permit operation of robots and autonomous vehicles in areas where the Global Positioning System (GPS) is degraded or unavailable.

Center, Julian↗
DOE OSTI2021

Seismic Spatial Gradients and Machine Learning-Based Classifiers for Explosion Monitoring (LDRD 218327)

This final report summarizes the work completed under the Laboratory Directed Research and Development (LDRD) project “Seismic Spatial Gradients as a Machine Learning-Based Classifier for Explosion Monitoring.” The overarching goal of the project was to explore the efficacy of using machine learning-based classification algorithms where the input data are the spatial gradient of the seismic wavefield collected at a single point on the Earth’s surface. The methods that I describe here are in direct contrast to conventional methods of seismic discrimination which typically rely on a spatially extended network of instruments and physics-based wavefield attributes such as, for example, the ratio between $\textit{P}$ and $\textit{S}$ waves. Rather, we use the spatial gradient of the seismic wavefield observed at a single point on the Earth’s surface and data processing approaches inspired by the machine learning community. We tested two algorithms, a neural network and a modified version of principal component analysis termed Spectrally Filtered Principal Component Analysis (SFPCA). To test these algorithms, we first conducted a series of numerical tests using synthetic data and then conducted a small-scale controlled field experiment. The tests using synthetic data showed that both algorithms had high success rates on gradiometric data, even when simulated noise was added to the signal. Furthermore, we found that using seismic spatial gradients increased the performance of our discrimination algorithms when compared to using just the traditional translational motion seismic data. The tests with field data also showed a high degree of discriminative success.

58 GEOSCIENCES↗
DOE OSTI2026

Machine learning enables reconstruction of past fire regimes from charcoal-derived fire intensity and fuel composition

Background Fire is a foundational ecological process that shapes ecosystem structure, diversity, and resilience. Quantifying paleofire regime attributes such as frequency, severity, and intensity is essential for understanding the historical range of variability in fire behavior and its ecological effects. While frequency and severity are often reconstructed in paleofire studies, quantitative reconstructions of fire intensity remain limited. Recent work has shown that maximum pyrolysis temperature—a proxy for fire intensity—and plant species type can be inferred from charcoal using transmission Fourier-transform infrared (FTIR) spectroscopy. However, the sample preparation for transmission FTIR is destructive and time-consuming, limiting application and reuse of materials for other analyses. We evaluated reflectance FTIR spectroscopy as a non-destructive alternative for reconstructing combustion temperature and plant species from laboratory-generated charcoal. We also examined the influence of contrasting airflow environments (ambient air versus nitrogen-rich) on pyrolysis temperature and plant species reconstruction prediction accuracies and compared predictive performance between a novel, neural network–based deep learning model with the traditional modern analogue technique (MAT) using k-nearest neighbor functions. As proof of concept, we apply our enhanced methodology to ancient charcoal to demonstrate applicability at improving long-term fire regime reconstructions and the ability to link paleofire records with contemporary fire ecology. Results Our analysis shows that transmission and reflectance FTIR spectra yield comparable spectral profiles. However, sample preparation for reflectance FTIR is minimal and non-destructive, unlike transmission FTIR which is destructive. We demonstrate that oxygen environments improved reconstruction accuracy relative to nitrogen-rich conditions. Finally, our deep learning neural network (DL) achieved testing accuracies of 98.7% for temperature and 96.2% for species identification, outperforming MAT’s k-NN approach (89.8% and 65.9%, respectively). A Shapley importance analysis identified 5 key spectral regions that greatly influenced the model’s temperature or species categorization. When applied to ancient charcoal, our results show historic fires from the most recent past primarily burned at low intensities (400–500 °C), reflective of natural fire regimes in ponderosa pine forests. Our results corroborate charcoal morphology data that suggests all ancient charcoal originated from burned woody plant types. Conclusions By combining reflectance FTIR spectroscopy with a deep learning approach, we provide the first accuracies high enough to confidently identify both species and temperature from laboratory-produced charcoal, improving quantitative reconstructions of fire intensity and fuel composition from paleofire records. This opens a wide range of research into the link between fire and larger drivers (i.e., climate or human) and greater ecological understanding of fire regimes beyond that of burn scars or recent observations. These methodological improvements have direct relevance for fire management by improving interpretation of historical fire behavior, informing fuel–fire relationships, and providing a scalable analytical framework applicable to both long-term ecological studies and contemporary fire science.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2024

Insights into the structure and dynamics of K + ions at the muscovite–water interface from machine learning potential simulations

The surfaces of many minerals are covered by naturally occurring cations that become partially hydrated and can be replaced by hydronium or other cations when the surface is exposed to water or an aqueous solution. These ion exchange processes are relevant to various chemical and transport phenomena, yet elucidating their microscopic details is challenging for both experiments and simulations. Here, in this work, we make a first step in this direction by investigating the behavior of the native K + ions at the interface between neat water and the muscovite mica (001) surface with ab-initio -based machine learning molecular dynamics and enhanced sampling simulations. Our results show that the desorption of the surface K + ions in pure ion-free water has a significant free energy barrier irrespective of their local surface arrangement. In contrast, facile K + diffusion between mica’s ditrigonal cavities characterized by different Al/Si orderings is observed. This behavior suggests that the K + ions may favor a dynamic disordered surface arrangement rather than complete desorption when exposed to deionized water.

Ab-initio molecular dynamics↗