SEARCH · Engineering Papers

Results for “training data”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,639 records · Page 91

All sources NASA NTRS DOE OSTI

DOE OSTI2019

Predicting Home Thermal Dynamics Using a Reduced-Order Model and Automatied Real-Time Parameter Estimation

Although smart thermostats have increasingly provided homeowners with abundant operational data related to advanced HVAC control and energy usage management in homes, there is a lack of systematic frameworks that can utilize such data to generate actionable information for advanced home comfort system diagnosis and control. Recognizing that a home thermal model, which is capable of connecting weather and HVAC operational data, is crucial to this framework, this paper introduces a home thermal model that is built upon the standard RC (Resistor-Capacitor) approach and the one virtual envelope assumption to describe the thermal dynamics of a home. A parameter estimation scheme is also developed that enables automatic, sequential, and optimal estimation of the model parameters, i.e., the thermal properties, of a home. Finally, the home thermal model and its parameter estimation scheme are tested using data collected over 63 consecutive days from a test home. In the test, the length of data that can provide enough robustness for parameter estimation is investigated. It is found that 10 days of data are enough, although 20 or more days of data can provide more robust results. In addition, when the model is used to make 12-hour-ahead predictions of indoor temperature the resulting mean, maximum, and 95%-confidence-interval absolute are 0.31 °C, 1.72 °C, and 0.90 °C, respectively, if the model is trained using 10 consecutive days of data, and 0.27 °C, 1.19 °C, and 0.72 °C, respectively, if 20 consecutive days are used. Therefore, the proposed home thermal model requires only a modest amount of training data to yield fairly accurate prediction and is able to perform better with more data.

home thermal model, Building thermal dynamics, ind↗

DOE OSTI2019

Interatomic Potential Model Development: Finite-Temperature Dynamics Machine Learning

Developing an accurate interatomic potential model is a prerequisite for achieving reliable results from classical molecular dynamics (CMD) simulations; however, most of the potentials are biased as specific simulation purposes or conditions are considered in the parameterization. Here, for developing an unbiased potential, a finite-temperature dynamics machine learning (FTD-ML) approach is proposed, and its processes and feasibility are demonstrated using the Buckingham potential model and aluminum (Al) as an example. Compared with conventional machine learning approaches, FTD-ML exhibits three distinguished features: 1) FTD-ML intrinsically incorporates more extensive configurational and conditional space for enhancing the transferability of developed potentials; 2) FTD-ML employs various properties calculated directly from CMD, for ML model training and prediction validation against experimental data instead of first-principles data; 3) FTD-ML is much more computationally cost effective than first-principles simulations, especially when the system size increases over 10 3 atoms as employed in this research for ensuring reliable training data. The Al Buckingham potential developed by the FTD-ML approach exhibits good performance for general simulation purposes. Thus, the FTD-ML approach is expected to contribute to a fast development of interatomic potential model suitable for various simulation purposes and conditions, without limitation of model type, while maintaining experimental-level accuracy.

42 ENGINEERING↗

DOE OSTI2025

Multi-fidelity learning for interatomic potentials: low-level forces and high-level energies are all you need

The promise of machine learning interatomic potentials (MLIPs) has led to an abundance of public quantum mechanical (QM) training datasets. The quality of an MLIP is directly limited by the accuracy of the energies and atomic forces in the training dataset. Unfortunately, most of these datasets are computed with relatively low-accuracy QM methods, e.g. density functional theory with a moderate basis set. Due to the increased computational cost of more accurate QM methods, e.g. coupled-cluster theory with a complete basis set (CBS) extrapolation, most high-accuracy datasets are much smaller and often do not contain atomic forces. The lack of high-accuracy atomic forces is quite troubling, as training with force data greatly improves the stability and quality of the MLIP compared to training to energy alone. Because most datasets are computed with a unique level of theory, traditional single-fidelity (SF) learning is not capable of leveraging the vast amounts of published QM data. In this study, we apply multi-fidelity learning (MFL) to train an MLIP to multiple QM datasets of different levels of accuracy, i.e. levels of fidelity. Specifically, we perform three test cases to demonstrate that MFL with both low-level forces and high-level energies yields an extremely accurate MLIP—far more accurate than a SF MLIP trained solely to high-level energies and almost as accurate as a SF MLIP trained directly to high-level energies and forces. Therefore, MFL greatly alleviates the need for generating large and expensive datasets containing high-accuracy atomic forces and allows for more effective training to existing high-accuracy energy-only datasets. Indeed, low-accuracy atomic forces and high-accuracy energies are all that are needed to achieve a high-accuracy MLIP with MFL.

36 MATERIALS SCIENCE↗

DOE OSTI2026

Describing Point Defect Topology in 2D Energy Materials Through Computer Vision

Point defects such as vacancies and impurity atoms strongly impact the performance of 2D materials. Traditional efforts often rely on manual detection, a process that is time-intensive, prone to human error, and challenging to scale. Here we leverage machine learning (ML) methods to identify and quantify vacancies within 2D transition metal carbides (Ti3C2, MXenes), aiming to expedite detection while improving accuracy. MXenes exhibit valuable defect-defined electrochemical properties, but we currently lack statistical understanding of defect topology needed to fully harness these materials. Here we employ a convolutional neural network for semantic segmentation of experimental MXene images, opening an opportunity to conduct a rigorous statistical study on defect hierarchy while investigating local relaxation in the lattice. We show how the integration of ML can yield fundamental insight into point defects, providing a powerful tool that will play an increasingly crucial role in the future of materials science. ML is often not just a matter of straightforward application, and pretrained models proved ineffective in this case. Instead, we trained our own neural network (NN) and applied data augmentation techniques and fine-tuning to the training dataset. Since labeled microscopy data is often scarce, we developed training data from a previously published wide-frame MXene image, using customized Gaussian fitting to locate atomic positions. Our trained model was then applied to a large dataset of experimental images, enabling a statistical study of defect configurations across three samples prepared with different HF etchant concentrations (5%, 9.1%, and 12.5%), as shown in Fig. 1. This also allowed us to investigate local strain around vacancies, though we find that we are limited by the precision of measurements using high-angle annular dark field (HAADF) images, as shown in Fig. 2. This study demonstrates how ML enables large-scale, quantitative analysis of atomic defects - an otherwise infeasible task with traditional methods. While our NN was specialized for Ti3C2 MXenes, the pipeline we developed provides a foundation for future ML models tailored to other materials. Ultimately, we envision embedding the NN onto the microscope to give real-time feedback to the user. To make this a reality, continued work is necessary to fully understand the NN's capabilities and limitations. This study gets one step closer to our goals of automated experimentation moving away from traditional methods of manual labeling. As ML capabilities advance, we hope to continue adapting and applying these techniques in microscopy.

2D materials↗

DOE OSTI2025

The Effects of Compounded Model Size Reductions on Adversarial Robustness

Recent advances in Edge AI and Tiny Machine Learning (TinyML) have enabled the deployment of machine learning models on resource-constrained environments. However, deploying these models on edge devices, such as micro-controllers, requires significant model footprint reduction through a variety of techniques such as quantization, pruning, and clustering. While these optimization methods offer considerable advantages, they potentially introduce AI-related security vulnerabilities, particularly concerning model robustness with respect to adversarial AI attacks. Prior research has extensively examined the impact of quantization on adversarial robustness; however, the effects of alternative reduction techniques and their combinations remain understudied. This paper investigates the impact of model size reduction techniques on adversarial robustness, when applied individually and combined. We utilized Fast Gradient Sign Method (FGSM) and Projected Gradient Descent (PGD) attacks to generate adversarial perturbations for both training and testing data, and then evaluated the models' accuracy under adversarial training conditions. Our findings revealed that reduction techniques generally diminished robustness; although, combining techniques was not found to make robustness any worse than when applied individually. Moreover, specific techniques can potentially enhance resistance to small size perturbations. This research provides insights into the trade-offs between model size reduction and security, establishing a foundation for future investigations into improving adversarial training techniques and methodologies for maintaining robustness while preserving memory footprint benefits.

Austria, Phillipe [ORNL] (ORCID:0000000236223973)↗

DOE OSTI2024

High-Dimensional Bayesian Optimization via Semi-Supervised Learning with Optimized Unlabeled Data Sampling

We introduce a novel semi-supervised learning approach, named Teacher-Student Bayesian Optimization (TSBO ), integrating the teacher-student paradigm into BO to minimize expensive labeled data queries for the first time. TSBO incorporates a teacher model, an unlabeled data sampler, and a student model. The student is trained on unlabeled data locations generated by the sampler, with pseudo labels predicted by the teacher. The interplay between these three components implements a unique selective regularization to the teacher in the form of student feedback. This scheme enables the teacher to predict high-quality pseudo labels, enhancing the generalization of the GP surrogate model in the search space. To fully exploit TSBO , we propose two optimized unlabeled data samplers to construct effective student feedback that well aligns with the objective of Bayesian optimization. Furthermore, we quantify and leverage the uncertainty of the teacher-student model for the provision of reliable feedback to the teacher in the presence of risky pseudo-label predictions. TSBO demonstrates significantly improved sample-efficiency in several global optimization tasks under tight labeled data budgets. The implementation is available at https://github.com/reminiscenty/TSBO-Official.

Yin, Yuxuan↗

NASA NTRS1985

Measuring Ada as a software development technology in the Software Engineering Laboratory (SEL)

An experiment is in progress to measure the effectiveness of Ada in the National Aeronautics and Space Administration/Goddard Space Flight Center flight dynamics software development environment. The experiment features the parallel development of software in FORTRAN and Ada. The experiment organization, objectives, and status are discussed. Experiences with an Ada training program and data from the development of a 5700-line Ada training exercise are reported.

Agresti, W. W.↗

NASA NTRSDate not supplied

Predictive Workload Model for Air Traffic Controllers during UAM Operations

The effect of airspace factors on air traffic controller (ATC) workload has been an active area of study for almost three decades due to the importance of safety considerations necessary to design and maintain operations. Existing literature has examined several traffic-related (e.g., number of aircraft under control, loss of separation) contributors to ATC workload and proposed mathematical functions to best describe controller response. However, future air traffic continues to increase in complexity with the introduction of urban air mobility (UAM) – or the transportation of humans and cargo using electric vertical takeoff and landing (eVTOL) aircraft. UAM aims to alleviate congestion for existing ground transportation systems and improve mobility within urban centers and other high-demand locations. This shift in the traditional airspace paradigm necessitates an evolved understanding of model use and development for ATC workload prediction. This study aimed to develop an ATC workload forecasting model based on human-in-the-loop (HITL) simulation data for UAM operations at large airports. Data collected from the HITL simulation served as the training and testing data for a Long Short-Term Memory recurrent neural network and enabled time-series forecasting of ATC workload from traffic characteristics. Results demonstrated the potential of LSTM models for forecasting ATC workload 40 minutes into the future and highlighted important considerations for future development.

predictive model↗

NASA NTRSDate not supplied

Predictive Workload Model for Air Traffic Controllers during UAM Operations

predictive model↗

DOE OSTI2023

Propagating Uncertainties in the SALT3 Model-training Process to Cosmological Constraints

Type Ia supernovae (SNe Ia) are standardizable candles that must be modeled empirically to yield cosmological constraints. To understand the robustness of this modeling to variations in the model-training procedure, we build an end-to-end pipeline to test the recently developed SALT3 model. We explore the consequences of removing pre-2000s low-z or poorly calibrated U-band data, adjusting the amount and fidelity of SN Ia spectra, and using a model-independent framework to simulate the training data. We find that the SALT3 model surfaces are improved by having additional spectra and U-band data, and can be shifted by ~5% if host-galaxy contamination is not sufficiently removed from SN spectra. We find that resulting measurements of w are consistent to within 2.5% for all of the training variants explored in this work, with the largest shifts coming from variants that add color-dependent calibration offsets or host-galaxy contamination to the training spectra and those that remove pre-2000s low-z data. These results demonstrate that the SALT3 model-training procedure is largely robust to reasonable variations in the training data, but that additional attention must be paid to the treatment of spectroscopic data in the training process. We also find that the training procedure is sensitive to the color distributions of the input data—the resulting w measurement can be biased by ~2% if the color distribution is not sufficiently wide. Future low-z data, particularly u-band observations and high signal-to-noise ratio SN Ia spectra, will help to significantly improve SN Ia modeling in the coming years.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

Data augmentation for disruption prediction via robust surrogate models

The goal of this work is to generate large statistically representative datasets to train machine learning models for disruption prediction provided by data from few existing discharges. Such a comprehensive training database is important to achieve satisfying and reliable prediction results in artificial neural network classifiers. Here, we aim for a robust augmentation of the training database for multivariate time series data using Student-t process regression. We apply Student-t process regression in a state space formulation via Bayesian filtering to tackle challenges imposed by outliers and noise in the training data set and to reduce the computational complexity. Thus, the method can also be used if the time resolution is high. We use an uncorrelated model for each dimension and impose correlations afterwards via coloring transformations. We demonstrate the efficacy of our approach on plasma diagnostics data of three different disruption classes from the DIII-D tokamak. To evaluate if the distribution of the generated data is similar to the training data, we additionally perform statistical analyses using methods from time series analysis, descriptive statistics, and classic machine learning clustering algorithms.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2019

Summary Report on Synthetic Time History Development and Deployment

This report summarizes research work performed at the Colorado School of Mines (CSM) and George Washington University (GWU) in support of the DOE Office of Nuclear Energy Integrated Energy Systems (IES) program. In particular, the work focuses on improving and strengthening two fundamental steps in the mathematical and computational framework that is currently being developed to optimize the economic performance of Integrated Energy Systems. One of the innovative approaches taken by the IES program is that, rather than optimizing the system for a specific year or week or day, the optimization is performed using stochastic synthetic data. Synthetic data is generated by training artificial intelligence (AI) using available but limited data sets. The synthetic data should possess the same patterns as the original set and the same statistical properties, but never repeat the same set of values exactly. The work from CSM focuses on this first step; namely, investigating a methodology to determine the quality of the synthetic data and the right settings for properly training the AI. Once a large set of trustworthy data can be generated the problem of stochastic optimization can be tackled. The optimal IES configuration is highly dependent on the rare high peaks or deep depths of the net demand (grid electricity demand less variable renewable energy [VRE] production). Using an unbiased random sampling of the synthetic data, these extremes are very difficult to generate and therefore the optimization can be unrepresentative and inaccurate. GWU has started an approach to increase the likelihood of generating these rare events with the proper associated probabilistic weight. Both of these projects represent a step forward in achieving the capability to properly size the IESs and to properly capture the value that these systems can provide to the grid by absorbing volatility of net demand without disproportionately increasing the cost of maintaining grid reliability.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

DOE OSTI2017

Beyond Fine Tuning: Adding capacity to leverage few labels

In this paper we present a technique to train neural network models on small amounts of data. Current methods for training neural networks on small amounts of rich data typically rely on strategies such as fine-tuning a pre-trained neural networks or the use of domain-specific hand-engineered features. Here we take the approach of treating network layers, or entire networks, as modules and combine pre-trained modules with untrained modules, to learn the shift in distributions between data sets. The central impact of using a modular approach comes from adding new representations to a network, as opposed to replacing representations via fine-tuning. Using this technique, we are able surpass results using standard fine-tuning transfer learning approaches, and we are also able to significantly increase performance over such approaches when using smaller amounts of data.

data science↗

DOE OSTI2014

Development and application of a statistical methodology to evaluate the predictive accuracy of building energy baseline models

This paper documents the development and application of a general statistical methodology to assess the accuracy of baseline energy models, focusing on its application to Measurement and Verification (M&V) of whole-building energy savings. The methodology complements the principles addressed in resources such as ASHRAE Guideline 14 and the International Performance Measurement and Verification Protocol. It requires fitting a baseline model to data from a ``training period’’ and using the model to predict total electricity consumption during a subsequent ``prediction period.’’ We illustrate the methodology by evaluating five baseline models using data from 29 buildings. The training period and prediction period were varied, and model predictions of daily, weekly, and monthly energy consumption were compared to meter data to determine model accuracy. Several metrics were used to characterize the accuracy of the predictions, and in some cases the best-performing model as judged by one metric was not the best performer when judged by another metric.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

NASA NTRS1981

Evaluation of SLAR and thematic mapper MSS data for forest cover mapping using computer-aided analysis techniques

A set of training statistics for the 30 meter resolution simulated thematic mapper MSS data was generated based on land use/land cover classes. In addition to this supervised data set, a nonsupervised multicluster block of training statistics is being defined in order to compare the classification results and evaluate the effect of the different training selection methods on classification performance. Two test data sets, defined using a stratified sampling procedure incorporating a grid system with dimensions of 50 lines by 50 columns, and another set based on an analyst supervised set of test fields were used to evaluate the classifications of the TMS data. The supervised training data set generated training statistics, and a per point Gaussian maximum likelihood classification of the 1979 TMS data was obtained. The August 1980 MSS data was radiometrically adjusted. The SAR data was redigitized and the SAR imagery was qualitatively analyzed.

Hoffer, R. M.↗

NASA NTRS1984

Determining Training Device Requirements in Army Aviation Systems

A decision making methodology which applies the systems approach to the training problem is discussed. Training is viewed as a total system instead of a collection of individual devices and unrelated techniques. The core of the methodology is the use of optimization techniques such as the transportation algorithm and multiobjective goal programming with training task and training device specific data. The role of computers, especially automated data bases and computer simulation models, in the development of training programs is also discussed. The approach can provide significant training enhancement and cost savings over the more traditional, intuitive form of training development and device requirements process. While given from an aviation perspective, the methodology is equally applicable to other training development efforts.

Poumade, M. L.↗

NASA NTRS1989

Per-point and per-field contextual classification of multipolarization and multiple incidence angle aircraft L-band radar data

Multipolarized aircraft L-band radar data are classified using two different image classification algorithms: (1) a per-point classifier, and (2) a contextual, or per-field, classifier. Due to the distinct variations in radar backscatter as a function of incidence angle, the data are stratified into three incidence-angle groupings, and training and test data are defined for each stratum. A low-pass digital mean filter with varied window size (i.e., 3x3, 5x5, and 7x7 pixels) is applied to the data prior to the classification. A predominately forested area in northern Florida was the study site. The results obtained by using these image classifiers are then presented and discussed.

Hoffer, Roger M.↗

DOE OSTI2021

Designing alloys with process-mapping AI pre-trained on empirical knowledge

<span style="font-family: Calibri, sans-serif; font-size: 12pt;">Accelerated materials design should match the recent trends in the product development cycles. Materials data analytics can be used to significantly shorten development time of specialized alloys needed for next generation energy applications. However, it faces a challenge of scarce data available for training ML models. Incorporation of the domain knowledge into deep-learning graph structure via fuzzy pre-training and causal process imitation presents a viable approach to developing accurate data-driven models and reliable alloy design tools, with limited datasets. Artificial Intelligence (AI) was used in this study to incorporate such knowledge in the domain-specific computational tool, pyroMind. The tool provides not only novel design ideas but also their interpretation via physics and engineering concepts.</span>

Romanov, Vyacheslav↗