SEARCH · Search NASA

Results for “latent feature models”

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 19 records

All sources NASA NTRS DOE OSTI

DOE OSTI2021

The Attraction Indian Buffet Distribution

We propose the attraction Indian buffet distribution (AIBD), a distribution for binary feature matrices influenced by pairwise similarity information. Binary feature matrices are used in Bayesian models to uncover latent variables (i.e., features) that explain observed data. The Indian buffet process (IBP) is a popular exchangeable prior distribution for latent feature matrices. In the presence of additional information, however, the exchangeability assumption is not reasonable or desirable. The AIBD can incorporate pairwise similarity information, yet it preserves many properties of the IBP, including the distribution of the total number of features. Thus, much of the interpretation and intuition that one has for the IBP directly carries over to the AIBD. A temperature parameter controls the degree to which the similarity information affects feature-sharing between observations. Unlike other nonexchangeable distributions for feature allocations, the probability mass function of the AIBD has a tractable normalizing constant, making posterior inference on hyperparameters straight-forward using standard MCMC methods. A novel posterior sampling algorithm is proposed for the IBP and the AIBD. We demonstrate the feasibility of the AIBD as a prior distribution in feature allocation models and compare the performance of competing methods in simulations and an application.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2025

An attention-based neural ordinary differential equation framework for modeling inelastic processes

To preserve strictly conservative behavior as well as model the variety of dissipative behavior displayed by solid materials, we propose a significant enhancement to the internal state variable-neural ordinary differential equation (ISV-NODE) framework. In this data-driven, physics-constrained modeling framework internal states are inferred rather than prescribed. The ISV-NODE consists of: (a) a stress model dependent on observable deformation and inferred internal state, and (b) a model of the evolution of the internal states. The enhancements to ISV-NODE proposed in this work are multifold: (a) a partially input convex neural network stress potential provides polyconvexity in terms of observed strain while leaving the inferred state unconstrained, and (b) an internal state flow model uses common latent features to inform novel attention-based gating and drives the flow of internal state only in dissipative regimes. We demonstrated that this architecture can accurately model dissipative and conservative behavior across an isotropic, isothermal elastic-viscoelastic-elastoplastic spectrum with three exemplars, while maintaining fundamental principles by design.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

INSURE: An Information Theory iNspired diSentanglement and pURification modEl for Domain Generalization

Domain Generalization (DG) aims to learn a generalizable model on the unseen target domain by only training on the multiple observed source domains. Although a variety of DG methods have focused on extracting domain-invariant features, the domain-specific class-relevant features have attracted attention and been argued to benefit generalization to the unseen target domain. To take into account the class-relevant domain-specific information, in this paper we propose an Information theory iNspired diSentanglement and pURification modEl (INSURE) to explicitly disentangle the latent features to obtain sufficient and compact (necessary) class-relevant feature for generalization to the unseen domain. Specifically, we first propose an information theory inspired loss function to ensure the disentangled class-relevant features contain sufficient class label information and the other disentangled auxiliary feature has sufficient domain information. Additionally, we further propose a paired purification loss function to let the auxiliary feature discard all the class-relevant information and thus the class-relevant feature will contain sufficient and compact (necessary) class-relevant information. Moreover, instead of using multiple encoders, we propose to use a learnable binary mask as our disentangler to make the disentanglement more efficient and make the disentangled features complementary to each other. We conduct extensive experiments on five widely used DG benchmark datasets including PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet. The proposed INSURE achieves state-of-the-art performance. We also empirically show that domain-specific class-relevant features are beneficial for domain generalization. The code is available at https://github.com/yuxi120407/INSURE .

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Data-driven model for divertor plasma detachment prediction

We present a fast and accurate data-driven surrogate model for divertor plasma detachment prediction leveraging the latent feature space concept in machine learning research. Our approach involves constructing and training two neural networks: an autoencoder that finds a proper latent space representation (LSR) of plasma state by compressing the multi-modal diagnostic measurements and a forward model using multi-layer perception (MLP) that projects a set of plasma control parameters to its corresponding LSR. By combining the forward model and the decoder network from autoencoder, this new data-driven surrogate model is able to predict a consistent set of diagnostic measurements based on a few plasma control parameters. In order to ensure that the crucial detachment physics is correctly captured, highly efficient 1D UEDGE model is used to generate training and validation data in this study. The benchmark between the data-driven surrogate model and UEDGE simulations shows that our surrogate model is capable of providing accurate detachment prediction (usually within a few per cent relative error margin) but with at least four orders of magnitude speed-up, indicating that performance-wise, it has the potential to facilitate integrated tokamak design and plasma control. Comparing with the widely used two-point model and/or two-point model formatting, the new data-driven model features additional detachment front prediction and can be easily extended to incorporate richer physics. This study demonstrates that the complicated divertor and scrape-off-layer plasma state has a low-dimensional representation in latent space. Understanding plasma dynamics in latent space and utilising this knowledge could open a new path for plasma control in magnetic fusion energy research.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

NASA NTRS1979

Sensible and latent heating of the atmosphere as inferred from DST-6 data

The average distribution of convective latent heating, boundary layer sensible heat flux, and vertical velocity are determined for the winter 1976 DST period from GLAS model diagnostics. Key features are the regions of intense latent heating over Brazil, Central Africa, and Indonesia; and the regions of strong sensible heating due to air mass modification over the North Atlantic and North Pacific Oceans.

Herman, G. F.↗

DOE OSTI2021

A machine learning-based interaction force model for non-spherical and irregular particles in low Reynolds number incompressible flows

In this study, interaction force of non-spherical particles in low Reynolds number gas-solid flow is investigated by neural network approaches. An artificial neural network (ANN) model is developed to correlate the non-spherical particle shape and the flow conditions with the interaction force. To define the particle shape, spherical harmonic expansion is applied. Furthermore, variational autoencoder model is then used to extract latent geometric features. The latent vector is utilized as an input with the Reynolds number for the ANN. The interaction force data, which is used as output data of the ANN, is obtained by particle resolved direct numerical simulation for 5200 non-spherical particles. The proposed model enables unsupervised extraction for non-spherical particle shapes and accurate predictions on the interaction force without heavy computation. This study provides the model that can explain complicated shapes of particles and be applied to a large scale, computational fluid dynamics simulation.

01 COAL, LIGNITE, AND PEAT↗

NASA NTRS1979

The use of satellite estimates of rainfall in the initialization of a mesoscale hurricane model

The use of rainfall rate data obtained from satellites to improve the initial specification of the hurricane circulation is discussed with particular reference to the techniques for incorporating rainfall data into high-resolution hurricane models. The variation of latent heating in a numerical model of a steady state tropical cyclone is considered to ascertain what features of the storm motion can be deduced from a given field of latent heat release. A real-data example involving a mesoscale model for hurricane Eloise (1975) has demonstrated that satellite measurements of rainfall may be beneficial to numerical hurricane prediction.

Fiorino, M.↗

DOE OSTI2022

Manifold Learning-Based Polynomial Chaos Expansions for High-Dimensional Surrogate Models

In this work we introduce a manifold learning-based method for uncertainty quantification (UQ) in systems describing complex spatiotemporal processes. Our first objective is to identify the embedding of a set of high-dimensional data representing quantities of interest of the computational or analytical model. For this purpose, we employ Grassmannian diffusion maps, a two-step nonlinear dimension reduction technique which allows us to reduce the dimensionality of the data and identify meaningful geometric descriptions in a parsimonious and inexpensive manner. Polynomial chaos expansion is then used to construct a mapping between the stochastic input parameters and the diffusion coordinates of the reduced space. An adaptive clustering technique is proposed to identify an optimal number of clusters of points in the latent space. The similarity of points allows us to construct a number of geometric harmonic emulators which are finally utilized as a set of inexpensive pretrained models to perform an inverse map of realizations of latent features to the ambient space and thus perform accurate out-of-sample predictions. Thus, the proposed method acts as an encoder-decoder system which is able to automatically handle very high-dimensional data while simultaneously operating successfully in the small-data regime. The method is demonstrated on two benchmark problems and on a system of advection-diffusion-reaction equations which model a first-order chemical reaction between two species. In all test cases, the proposed method is able to achieve highly accurate approximations which ultimately lead to the significant acceleration of UQ tasks.

42 ENGINEERING↗

DOE OSTI2022

Deep learning for drag force modelling in dilute, poly-dispersed particle-laden flows with irregular-shaped particles

Here, this study applies machine learning-based approaches to develop a drag force model for irregular-shaped particles in incompressible flows. The particle-laden flows are studied through an in-house particle-resolved direct numerical simulation (PR-DNS) at low-intermediate Reynolds numbers (Re). We utilize the PR-DNS to obtain drag force coefficients and flow fields of single particles. A variational auto-encoder model is used to obtain latent vectors to represent the geometrical features of the particles, and artificial neural networks (ANN) are developed to predict drag force coefficients and flow fields of the single particles. This study applies a pairwise interaction extended point-particle (PIEP) model to obtain the coefficients assuming the flow fields of neighboring particles can be linearly superposed. The PIEP method shows significant improvement on prediction for a few neighbored particles. In addition, the results reveal R2 scores of 0.56-0.62 and errors of 9.1-10.0 % for the dilute, polydispersed systems with a volume fraction of 0.5 %.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Quantification of Lanthanides in a Molten Salt Reactor Surrogate Off-Gas Stream Using Laser-Induced Breakdown Spectroscopy

To enable the deployment of molten salt reactor technology, the development of off-gas treatment systems and advanced monitoring tools capable of operating with high temperatures and radiation fields while delivering near real-time information is necessary. Here, we aim to fulfill this requirement and proposes laser-induced breakdown spectroscopy (LIBS) for monitoring molten salt aerosol streams. A sheath gas measuring method was developed to protect optical elements from aerosol particles and to ensure a relatively constant aerosol stream for measurement. An aqueous system was studied to demonstrate the utility of LIBS for monitoring possible fission products in an aerosol stream: Gd, Nd, and Sm up to 2000 parts per million (ppm). A calibration model was built using partial least squares (PLS) regression with five, six, and nine latent variables for Gd, Nd, and Sm, respectively. This calibration model successfully estimated the concentrations of three test samples, which were validated with inductively charged plasma optical emission spectroscopy measurements at a 99.9% confidence interval. To enhance these models, a genetic algorithm was used to filter the spectra before entering the PLS model, thereby limiting the spectral features being regressed to those with greater correlations to concentration. This allowed for the number of latent variables used in the PLS models to be reduced to four, three, and three for Gd, Nd, and Sm, respectively. Finally, the genetic algorithm-filtered PLS models were used to predict the concentrations of the aerosol stream on a real-time dataset and resulted in a 73%, 18%, and 25% improvement in root mean squared error of prediction compared to the original PLS models developed.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

NASA NTRS1988

Hybrid transfinite element modeling/analysis of nonlinear heat conduction problems involving phase change

The present paper describes the applicability of hybrid transfinite element modeling/analysis formulations for nonlinear heat conduction problems involving phase change. The methodology is based on application of transform approaches and classical Galerkin schemes with finite element formulations to maintain the modeling versatility and numerical features for computational analysis. In addition, in conjunction with the above, the effects due to latent heat are modeled using enthalpy formulations to enable a physically realistic approximation to be dealt computationally for materials exhibiting phase change within a narrow band of temperatures. Pertinent details of the approach and computational scheme adapted are described in technical detail. Numerical test cases of comparative nature are presented to demonstrate the applicability of the proposed formulations for numerical modeling/analysis of nonlinear heat conduction problems involving phase change.

Tamma, Kumar K.↗

NASA NTRS2024

Transfer-AE: A novel autoencoder-based impact detection model for structural digital twin

Accurately detecting the location and intensity of impacts is crucial for ensuring structural safety. Currently, AI-based structural impact detection methods are widely used for their excellent detection accuracy. However, their generalization capability is limited by the scenarios present in the training data. Many complex and dangerous impact scenarios are difficult to conduct real-world experiments on to collect sufficient samples. To capture all impact scenarios and fully leverage the advantages of AI-based detection technologies, advanced methods involve combining real-world structural monitoring data with corresponding numerical models to construct digital twins. These methods continuously refine the created numerical models with limited real-world data and provide diverse impact scenarios through numerical model simulations. However, there are inevitable differences between digital models and physical models that are challenging to correct through mechanical means. This discrepancy in data distribution between the two models significantly hinders the application of digital twin technology in impact/event identification tasks. To address this challenge, this study proposes a novel model based on autoencoders, named Transfer-AE. Transfer-AE encodes the common features of digital twins in the latent space to bridge the uncertainty gap at a macro scale between numerical models and physical models and synchronously fits the magnitude and location of the impact load in the decoder. This enables consistent detection results for the same impact event, whether the sample comes from the numerical model or the physical model. Transfer-AE includes two operating modes: Mode 1 has a fixed computational complexity with stable inference speed, but the training cost and difficulty increase with data distribution. Mode 2's computational complexity increases with data distribution, but it has a fixed training cost and speed. In both cases involving the geodesic dome structure simulating a deep space habitat and the IASC-ASCE benchmark structure, Transfer-AE demonstrated the best performance in impact localization and quantification tasks compared to mainstream domain-adaptive transfer models.

Chengjia Han↗

DOE OSTI2024

Unsupervised physics-informed disentanglement of multimodal data

Here, we introduce physics-informed multimodal autoencoders (PIMA) - a variational inference framework for discovering shared information in multimodal datasets. Individual modalities are embedded into a shared latent space and fused through a product-of-experts formulation, enabling a Gaussian mixture prior to identify shared features. Sampling from clusters allows cross-modal generative modeling, with a mixture-of-experts decoder that imposes inductive biases from prior scientific knowledge and thereby imparts structured disentanglement of the latent space. This approach enables cross-modal inference and the discovery of features in high-dimensional heterogeneous datasets. Consequently, this approach provides a means to discover fingerprints in multimodal scientific datasets and to avoid traditional bottlenecks related to high-fidelity measurement and characterization of scientific datasets.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Highly sensitive 2D X-ray absorption spectroscopy via physics informed machine learning

Abstract Improving the spatial and spectral resolution of 2D X-ray near-edge absorption structure (XANES) has been a decade-long pursuit to probe local chemical reactions at the nanoscale. However, the poor signal-to-noise ratio in the measured images poses significant challenges in quantitative analysis, especially when the element of interest is at a low concentration. In this work, we developed a post-imaging processing method using deep neural network to reliably improve the signal-to-noise ratio in the XANES images. The proposed neural network model could be trained to adapt to new datasets by incorporating the physical features inherent in the latent space of the XANES images and self-supervised to detect new features in the images and achieve self-consistency. Two examples are presented in this work to illustrate the model’s robustness in determining the valence states of Ni and Co in the LiNi x Mn y Co 1-x-y O 2 systems with high confidence.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Active Learning Surrogates for Integrating Electron Microscopy and Computational Insights from Simulations in Autonomous Experiments

Artificial Intelligence (AI) combined with simulations and experiments has great potential to accelerate scientific discovery across technology and pharmaceuticals. However, the gap between simulations and experiments is challenging due to disparities in time and scale, making it difficult to estimate properties like energy and electronic states from experiments, and to provide feedback based on theoretical insights.Our research addresses the challenge by developing unique deep kernel based surrogate models that learns from microscopic images, mapping structural features to energy differences from defect formation. We start with full-training using simulated images to determine optimal settings, establishing a baseline for active learning. Using these settings from the baseline, active learning is trained, and predicts structures along simulation trajectories based on uncertainty and energetic stability, thus reducing data requirements, simulation time and computational costs. The results demonstrate that the model achieves a low average error margin of approximately 0.03 meV, indicating good performance. To enhance feature extraction and reconstruction capabilities, we developed an autoencoder-decoder as additional surrogate to create latent space to capture essential features, enabling precise comparisons between simulations and experiments. The results from this model achieved a reconstruction loss of around 0.2 and accurately reconstructed molecular structures.Overall, this work advances the steering of experiments through computational simulations by employing a surrogate models that actively predicts the trajectories of structural evolution, achieving time-to-solution comparable to experimental measurements.

Saranathan, Gayathri [Hewlett-Packard]↗

DOE OSTI2022

A Probabilistic Autoencoder for Type Ia Supernova Spectral Time Series

We construct a physically parameterized probabilistic autoencoder (PAE) to learn the intrinsic diversity of Type Ia supernovae (SNe Ia) from a sparse set of spectral time series. The PAE is a two-stage generative model, composed of an autoencoder that is interpreted probabilistically after training using a normalizing flow. We demonstrate that the PAE learns a low-dimensional latent space that captures the nonlinear range of features that exists within the population and can accurately model the spectral evolution of SNe Ia across the full range of wavelength and observation times directly from the data. By introducing a correlation penalty term and multistage training setup alongside our physically parameterized network, we show that intrinsic and extrinsic modes of variability can be separated during training, removing the need for the additional models to perform magnitude standardization. We then use our PAE in a number of downstream tasks on SNe Ia for increasingly precise cosmological analyses, including the automatic detection of SN outliers, the generation of samples consistent with the data distribution, and solving the inverse problem in the presence of noisy and incomplete data to constrain cosmological distance measurements. We find that the optimal number of intrinsic model parameters appears to be three, in line with previous studies, and show that we can standardize our test sample of SNe Ia with an rms of 0.091 ± 0.010 mag, which corresponds to 0.074 ± 0.010 mag if peculiar velocity contributions are removed.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2022

High-Temperature Latent Heat Storage System using Transportable Heat Pipes for Versatile Integration with Emerging Microreactors

With the growing interest in the development of micro nuclear reactors and their application to microgrid, high-temperature thermal energy storage (TES) has also been receiving great attention as a technology to reinforce the energy utilization and profitability of this class of reactors. Given the importance and research needs of the TES technology, a research project is underway at Idaho National Laboratory (INL) to develop high-temperature TES that can be easily interfaced with emerging micro nuclear reactors. Currently under development is the Heat pipe-Integrated Thermal Battery (HITB), a novel latent heat storage design that enables flexible and robust integration with any design of currently considered micro nuclear reactors. The development of detailed design based on the proposed TES concept, HITB, requires not only development and validation of modeling and simulation capabilities, but also research that spans thermal performance analysis, component- and system-scale design optimization, and experimental demonstrations including high-temperature feasibility testing. The current research efforts (and plans) are divided into the following four areas: (i) Thermal and structural modeling and analysis at both system- and component-scales for evaluating performance and conducting design optimization, (ii) Small-scale experimental demonstrations for the key components of the proposed HITB design, (iii) Modelica-based dynamic model development for developing a control strategy under dynamic operation within integrated energy systems configuration, (iv) Integrated experimental testing and demonstration. This paper introduces the overall research efforts that have been and are being conducted to develop and demonstrate the novel design of the high-temperature latent heat storage device, HITB. The key design features, ongoing research efforts including the development of modeling and simulation capabilities, small-scale demonstrations of the HITB’s key functionalities prior to scaled-up experimental demonstration, are discussed. Then, this paper concludes with a remark on the future research plans for this project.

25 ENERGY STORAGE↗

NASA NTRS1995

A numerical model of nonmigrating diurnal tides between the surface and 65 km

Observations of surface pressure and middle atmosphere temperatures and winds indicate that a substantial nonmigrating component is present in the diurnal tide. The nonmigrating tides, which propagate with a zonal phase speed that is different from the earth's rotation, are attributed to the diurnal heating of geographically fixed sources. In this study we utilize a classical tidal model to examine the propagation characteristics of diurnal tides. The global fields of tropospheric sensible, radiative, and latent heating used to drive the model are supplied from summer and winter diurnal climatologies of the National Center for Atmospheric Research (NCAR) Community Climate Model (CCM2). A novel aspect of this study is the focus on the relative importance of the nonmigrating components. The classical model successfully reproduces many observed features of the low-latitude diurnal surface pressure tides. In the middle atmosphere, the simulated migrating (or sun-synchronous) tide shows qualitative agreement with November-March Limb Infrared Monitor of the Stratosphere (LIMS) observations. Tropospheric solar heating is clearly the dominant driving force for the migrating tide, with secondary contributions from boundary-layer sensible heating and tropospheric latent heat release. The leading modes of the zonal mean tide are also driven chiefly by tropospheric solar heating. The higher-order modes of the zonal mean and eastward propagating tides may be attributed to the joint effects of tropospheric solar heating, sensible heating, and latent heat release. The LIMS and other data reveal features that cannot be explained or examined within the context of the classical model used in the present study. These include upward phase propagation, vertical attenuation, and temporal variations in the migrating diurnal tide.

Lieberman, Ruth S.↗