Search NASA⌕ Search

SEARCH · Search NASA

Results for “embedding model”

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 127 records · Page 7

Machine learning approaches for integrating multi-omics data to expand microbiome annotation (Final Technical Report)

We fulfilled all original three aims of the proposal. Following the earlier release (during the first phase of the project at Montana) of software that identifies and fills gaps in the annotation of metabolic proteins within bacterial genomes, we have nearly completed a second gap-filling tool that improves accuracy and explainability. We completed software for alignment-based annotation of protein coding DNA, allowing for coding frameshifts caused by sequencing error. Finally, we completed a neural embedding model for identifying similarities between protein sequences based on amino-wise latent vectors.

59 BASIC BIOLOGICAL SCIENCES↗

Facial Named Entity Recognition by Attention-Based Graph Convolutional Neural Network

In the realm of facial recognition and analysis, the ability to accurately cluster large datasets of facial images stands as a cornerstone for various applications, ranging from security surveillance to user biometric identification. This project evolves a novel approach to facial data clustering by embedding facial images into a high-dimensional vector space using an advanced embedding model trained on separate data and assumes a graph-like structure on the high-dimensional vectors. We find our method works significantly better than common shallow methods.

97 MATHEMATICS AND COMPUTING↗

Knowledge Graph Entity Linking using Graph Embeddings

Details the use of a custom embedding model on knowledge graphs to aid in downstream natural language processing (NLP) models for Derivative Classification Assist. Motivations, algorithms, and results were discussed.

Mahesh, Aarav [Sandia National Laboratories (SNL-N↗

FPGA-accelerated SpeckleNN with SNL for real-time X-ray single-particle imaging

We present the implementation of a specialized version of our previously published unified embedding model, SpeckleNN, for real-time speckle pattern classification in X-ray Single-Particle Imaging (SPI), using the SLAC Neural Network Library (SNL) on an FPGA platform. This hardware realization transitions SpeckleNN from a prototypic model into a practical edge solution, optimized for running inference near the detector in high-throughput X-ray free-electron laser (XFEL) facilities, such as those found at the Linac Coherent Light Source (LCLS). To address the resource constraints inherent in FPGAs, we developed a more specialized version of SpeckleNN. The original model, which was designed for broader classification across multiple biological samples, comprised ~5.6 million parameters. The new implementation, while reducing the parameter count to 64.6K (a 98.8% reduction), focuses on maintaining the model's essential functionality for real-time operation, achieving an accuracy of 90%. Furthermore, we compressed the latent space from 128 to 50 dimensions. This implementation was demonstrated on the KCU1500 FPGA board, utilizing 71% of available DSPs, 75% of LUTs, and 48% of FFs, with an average power consumption of 9.4W according to the Vivado post-implementation report. The FPGA performed inference on a single image with a latency of 45.015 microseconds at a 200 MHz clock rate. In comparison, running the same inference on an NVIDIA A100 GPU resulted in an average power consumption of ~73W and an image processing latency of around 400 microseconds. Our FPGA-accelerated version of SpeckleNN demonstrated significant improvements, achieving an 8.9 × speedup and a 7.8 × reduction in power consumption compared to the GPU implementation. Key advancements include model specialization and dynamic weight loading through SNL, which eliminates the need for time-consuming FPGA design re-synthesis, allowing fast and continuous deployment of models (re)trained online. These innovations enable real-time adaptive classification and efficient vetoing of speckle patterns, making SpeckleNN more suited for deployment in XFEL facilities. This implementation has the potential to significantly accelerate SPI experiments and enhance adaptability to evolving experimental conditions.

47 OTHER INSTRUMENTATION↗

Investigation of the \(\gamma \)-ray Properties of the \(2^+\) States in \(^{14}\)C

The properties of the 2$^+_1$ and 2$^+_2$ excited states in 14 C were studied in an experiment conducted at Argonne National Laboratory. A 9 Be( 6 Li,pγ) fusion-evaporation reaction and the GRETINA-ORRUBA setup were employed to populate states of 14 C and detect γ-particle coincidence events. Finally, the precise determination of the 2$^+_1$ level energy, complemented by the estimation of the γ-ray branch of the 2$^+_2$ near-threshold state, will serve as a benchmark to test the Shell Model Embedded in the Continuum calculations.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Constrained Block Nonlinear Neural Dynamical Models

Neural network modules conditioned by known priors can be effectively trained and combined to represent systems with nonlinear dynamics. This work explores a novel formulation for data-efficient learning of deep control-oriented nonlinear dynamical models by embedding local model structure and constraints. The proposed method consists of neural network blocks that represent input, state, and output dynamics with constraints placed on the network weights and system variables. For handling partially observable dynamical systems, we utilize a state observer neural network to estimate the states of the system's latent dynamics. We evaluate the performance of the proposed architecture and training methods on system identification tasks for three nonlinear systems: a continuous stirred tank reactor, a two tank interacting system, and an aerodynamics body. Models optimized with a few thousand system state observations accurately represent system dynamics in open loop simulation over thousands of time steps from a single set of initial conditions. Experimental results demonstrate an order of magnitude reduction in open-loop simulation mean squared error for our constrained, block-structured neural models when compared to traditional unstructured and unconstrained neural network models.

Skomski, Elliott↗

An Introduction to Word Embeddings and Language Models

Language models have advanced at a phenomenal pace over the past decade. This document provides a short introduction to terminology, word embeddings (aka low-dimensional representations), and popular large-scale language models (LMs). Word embeddings are used to represent words as numerical vectors and are context-independent, meaning a word can only have a single representation (e.g., club can only be club sandwich, not golf club ). Language models can determine the probability of a given sequence of words occurring in a sentence and can provide context to distinguish between words and phrases that sound similar. LMs are context-dependent (e.g., club can be club sandwich or golf club ) and largely fall in two main classes – autoregressive and autoencoding models. Autoregressive models are pretrained on the classic language modeling task: guess the next token having read all the previous ones. Those models can be fine-tuned and achieve great results on many tasks, the most natural application is text generation. A typical example of such models is GPT, but others include GPT-2, GPT-3, CTLR, TRANSFORMER-XL, REFORMER, XLNET. Autoencoding models are pretrained by corrupting the input tokens in some way and trying to reconstruct the original sentence. They can be fine-tuned and achieve great results on many tasks such as text generation, but their most natural application is sentence classification or token classification. A typical example of such models is BERT, but others include ROBERTA, ALBERT, XML, XML-ROBERTA, FLAUBERT AND LONGFORMER.

97 MATHEMATICS AND COMPUTING↗

Experiments and Modeling of Proppant Embedment and Fracture Conductivity for the Caney Shale, Oklahoma, USA

ABSTRACT: The ultimate aim of hydraulic fracturing is to have a long and conductive flow path that extends from the wellbore into the formation. The effective fracture length is part of a hydraulically propped fracture which contributes to production. The difficulty in achieving economical production targets from shale reservoirs is at the forefront in many exploration companies. Fracture conductivity loss is related to; proppant embedment under depletion, proppant crushing, damage as a result of fracturing fluid, fines migration and proppant-pack permeability-damage are some of the factors that contribute to production decline after hydraulic fracturing in shale reservoirs. The Caney Shale is a calcareous organic-rich mudrock. Various studies have investigated the effect that clay on shale well productivity, however, there is currently no literature on the Caney shale in relation to horizontal wells; all available literature exists in vertical wells as well as on formations of the Caney that are shallow in comparison to an emerging play which is twice the depth. In this paper we investigate stress-dependent fracture conductivity of proppant-filled fractures and proppant embedment in Caney shale through laboratory and modeling studies. API fracture conductivity tests were conducted using 2% KCl on five locations within the Caney shale that consisted of selecting three brittle(reservoir) zones and two ductile zones. Confining pressures range from 1,000 psi to 12,000 psi at 210°F. Conductivity, permeability as well as embedment were measured during the test. Our experimental results have confirmed that improved fracture conductivity is attributed to; proppant size, the increase in porosity of the proppant pack, closure pressure changes and the reduction in fracture conductivity are a function of many factors such as fracture closure stress. The findings from this study could help the stimulation design by providing new insights into the critical factors that are to be determined to facilitate the choice of proppants as well as fracturing fluids for long term production and recovery from shale reservoirs. 1 INTRODUCTION The development of low permeability formations, like shales, has been aided by hydraulic fracturing of horizontal wells (Radonjic et al., 2020). Hydraulic fracturing fluid is injected at a high pressure to induce tensile fractures that can link to and stimulate natural fractures (Katende et al., 2021a,b). Preserving adequate conductivity in hydraulic fractures over the life of the wells is required for economic production; nevertheless, conserving such conductivity can be difficult in some circumstances, particularly in soft, clay-rich formations (Wang et al., 2021). Proppant particles help to keep the fractures open when the pumping stops and the fracturing fluid returns to the wellbore, producing one or more propped hydraulic fractures of varying length, breadth, and height (Katende et al., 2021a). The proppant pack within the hydraulic fracture boosts well output by providing a greater permeability flowpath for hydrocarbons (Duenckel et al., 2016). Proppant in the fracture is under complicated stress conditions, and the interplay between the rock formation and the proppant pack has a significant impact on proppant-pack permeability (Karazincir et al., 2019). Proppant may be embedded (Katende et al., 2021a) in the rock or crushed into small pieces if the proppant size and strength characteristics are not specified appropriately, resulting in a loss in proppant-pack permeability and fracture aperture, and consequently a fall in well output.

Katende, A.↗

A hybrid embedded discrete fracture model for simulating tight porous media with complex fracture systems

Numerical simulation considering complex fracture networks in tight oil reservoir becomes a hot issue during these years. Recently, embedded discrete fracture model (EDFM) shows great advantages over the DFM approaches in both calculation efficiency and simplicity of simulation workflow. EDFM has been widely used in tight oil/gas natural energy or enhanced oil recovery (EOR) development. If a large number of small-scale natural fractures exist in reservoir, the calculation efficiency of EDFM method may decrease. A new hybrid simulation model combining the EDFM and dual porosity (DP) model is proposed. The large-scale fractures are dealt with EDFM method while small scale fractures are dealt with DP model. Considering the low permeability of matrix, transient transfer between matrix and natural fracture is considered. The proposed hybrid model is validated compared with the local grid refinement (LGR) model and EDFM-MINC model, and the difference between transient transfer and pseudo-steady-state transfer effect on production is presented. Finally, two test cases are presented, and the sensitivity of some key parameters are investigated. Result shows that the new simulation model has higher calculation efficiency than LGR and EDFM-MINC models. Meanwhile, it honors a good accuracy. The pseudo-steady-state transfer between matrix and natural fracture leads to big error in tight oil production performance analysis both in single phase and two phases flow. Low matrix permeability and big matrix size increase the production difference between pseudo-steady-state transfer and transient transfer models. The proposed model can simulate reservoirs considering stochastic permeability distribution.

58 GEOSCIENCES↗

HybridSystemsSimulations.jl - Solving the Merchant Collocated Facilities with JuMP

The development of new clean-generation technologies also leads to new plant-level architectures that combine several generation and storage assets behind the point of connection. These co-located generation resources (Hybrid Systems) primarily operate as merchant assets that employ automated market bidding models and internal Energy Management Systems (EMS) to comply with the operator's signals. Formulating an optimal bidding model requires embedding the EMS control model into the bidding algorithm, resulting in a bi-level optimization problem. In this presentation, we first showcase using JuMP to formulate and solve this problem effectively for multiple merchant systems and the bidding outcomes considering different model formulations. Second, the bidding outcomes are later integrated into a PowerSimulations.jl (also built with JuMP) simulation to study the system-level effects of the various merchant bidding and the interactions between market-clearing models and the embedded EMS model. We will showcase simulations conducted in the RTS system considering different levels of merchant hybrid systems participation. The presentation provides the following specific insights on JuMP usage: 1) the Formulation of specialized bi-level problems with custom cuts to solve the merchant hybrid system bidding problem; 2) the integration of a modular model within a complex simulation workflow supported by JuMP in PowerSimulations.jl; 3) Accelerating the solution of power systems operations simulation that employ agent optimization problems using JuMP.

energy markets↗

Interface and embedded joint methods for modeling dynamic fracture opening by explosive products [Two methods for modeling dynamic fracture opening by explosive products]

Two computational approaches are proposed in the paper to model dynamic fracture opening by explosive products. The first method assumes that the fractures may be modeled using flow elements embedded along the mesh lines. This method models crack opening in a straightforward way by splitting the nodes of the computational grid. It can account for crack branching; however, the crack directions are constrained by existing mesh faces, which may lead to mesh dependence. Also, the stress in flow elements is calculated explicitly separate from the surrounding solid elements that can impose additional limits on the time step stability condition for explicit integration. The second approach uses embedded flow elements to model the cracks. Typical thickness of the cracks is much smaller than the element size. Therefore, gas pressure in the cracks is assumed to be in stress equilibrium with the element stress. To achieve this, the crack thickness and the state of the gas is updated simultaneously with the state of the solid element which contains the crack. Therefore, the time step is controlled by the explicit solver applied for the solid and does not depend on the thickness of the crack. In conclusion, the main disadvantage of the second approach is due to the complexity of modeling multiple intersecting cracks, which go through the same element. We discuss the areas of possible applications of these 2 methods and the ways to improve and enhance them for future practical applications.

58 GEOSCIENCES↗

Complete asymptotically safe embedding of the standard model

We introduce and discuss the “Tetrad Model”, a large color/flavor embedding of the Standard model which has an interacting ultraviolet fixed point. It is demonstrated that its extended-Pati-Salam symmetry is broken radiatively via the Coleman-Weinberg mechanism, while the remaining electroweak symmetry is broken when mass-squared terms run negative. In the IR the theory yields just the Standard Model, augmented by the fact that the Higgs fields carry the same generation indices as the matter fields. It is also shown that the Higgs mass-squareds develop a hierarchical structure in the IR, from a UV theory that is asymptotically flavor symmetric, opening up an interesting direction for explaining the emergence of the observed flavor structure.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Parallel simulated annealing with embedded machine learning and multifidelity models for reactor core design

This paper presents extensions to a penalty-free, parallel simulated annealing (SA) algorithm for multi-constrained combinatorial optimization with the aim of embedding multi-fidelity physics models into the annealing procedure. The method uses a low-fidelity, quickly executing model for rapid design space exploration and a high-fidelity model for detailed constraint resolution and on-the-fly bias correction. Machine learning models updated within the annealing procedure were used to bridge the gap between the multi-fidelity models, which led to accurate rapid exploration and efficient detailed constraint resolution. A software implementation of the new multi-fidelity optimization methods, called ML-PSA, was demonstrated on a continuous multi-fidelity optimization problem and a constrained combinatorial PWR lattice design problem. These problems demonstrate some of the features, parallel performance characteristics, and extensible nature of the multi-fidelity SA methods. This paper shows that the developed software and procedure are a general optimization tool that can be applied to a wide variety of scientific and engineering design optimization applications. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Stable Machine‐Learning Parameterization of Subgrid Processes in a Comprehensive Atmospheric Model Learned From Embedded Convection‐Permitting Simulations

Modern climate projections often suffer from inadequate spatial and temporal resolution due to computational limitations, resulting in inaccurate representations of sub-grid processes. A promising technique to address this is the multiscale modeling framework (MMF), which embeds a kilometer-resolution cloud-resolving model (CRM) within each atmospheric column of a host climate model to replace traditional convection and cloud parameterizations. Machine learning offers a unique opportunity to make MMF more accessible by emulating the embedded CRM and reducing its substantial computational cost. Although many studies have demonstrated proof-of-concept success of achieving stable hybrid simulations, it remains a challenge to achieve near operational-level success with real geography and comprehensive variable emulation that includes, for example, explicit cloud condensate coupling. In this study, we present a stable hybrid model capable of integrating for at least 5 years with near operational-level complexity, including coarse-grid geography, seasonality, explicit cloud condensate and wind predictions, and land coupling. Our model demonstrates skillful online performance, achieving a 5-year zonal mean tropospheric temperature bias within 2 K, water vapor bias within 1 g/kg, and a precipitation root mean square error of 0.96 mm/day. Key factors contributing to our online performance include an expressive U-Net architecture and physical thermodynamic constraints for microphysics. With microphysical constraints mitigating unrealistic cloud formation, our work is the first to demonstrate realistic multi-year cloud condensate climatology under the MMF framework. Despite these advances, online diagnostics reveal persistent biases in certain regions, highlighting the need for innovative strategies to further optimize online performance.

Hu, Zeyuan [NVIDIA Corporation, Santa Clara, CA (U↗

Data from a four-day long microcosm experiment addressing the destabilization of artificial mineral-associated organic matter by model root exudates embedded in a soil matrix from the Rocky Mountain Biological Laboratory (Gothic, CO, USA), 2019

This dataset provides data collected during a four-day long laboratory soil microcosm experiment testing the efficacy of root exudate-driven mineral-associated organic matter destabilization. This dataset contains four data files in comma-separate values (*.csv). The files provide the metadata and the experimental results on microbial respiration, MAOM-derived respiration, and sequential mineral-extractions. This data was used to produce the figures in Bölscher et al., 2026. The results of the experiment can be found in the open access article Bölscher et al., 2026 (https://doi.org/10.1016/j.soilbio.2026.110276). Abstract: Mineral-associated organic matter (MAOM) is often considered stable, but root exudates can destabilize MAOM via various pathways. Theory and model system studies suggest that direct MAOM destabilization by strong ligands, like oxalic acid, or reducing agents, like catechol, is more effective than indirect, microbial-mediated MAOM destabilization, stimulated by less reactive compounds like glucose. Here, we demonstrate that the presence of a soil matrix alters the efficacy of exudate-driven MAOM destabilization pathways. Glucose and catechol destabilized significantly greater amounts of MAOM from ferrihydrite and aluminum hydroxide (Al (OH)3) embedded in a soil matrix than oxalic acid. Our findings indicate that indirect, microbial-mediated MAOM destabilization may play a larger role than direct MAOM destabilization in soil environments.

Destabilization↗

Comparative study of adaptive variational quantum eigensolvers for multi-orbital impurity models

Hybrid quantum-classical embedding methods for correlated materials simulations provide a path towards potential quantum advantage. However, the required quantum resources arising from the multi-band nature of d and f electron materials remain largely unexplored. Here we compare the performance of different variational quantum eigensolvers in ground state preparation for interacting multi-orbital embedding impurity models, which is the computationally most demanding step in quantum embedding theories. Focusing on adaptive algorithms and models with 8 spin-orbitals, we show that state preparation with fidelities better than 99.9% can be achieved using about 2 14 shots per measurement circuit. When including gate noise, we observe that parameter optimizations can still be performed if the two-qubit gate error lies below 10 -3 , which is slightly smaller than current hardware levels. Finally, we measure the ground state energy on IBM and Quantinuum hardware using a converged adaptive ansatz and obtain a relative error of 0.7%.

42 ENGINEERING↗

Efficient Simulation of Cascading Outages Using an Energy Function-Embedded Quasi-Steady-State Model

Here, this paper proposed an energy function-embedded quasi-steady-state model for efficient simulation of cascading outages on a power grid while addressing transient stability concerns. Compared to quasi-steady-state models, the proposed model incorporates short-term dynamic simulation and an energy function method to efficiently evaluate the transient stability of a power grid together with outage propagation without transient stability simulation. Cascading outage simulation using the proposed model conducts three steps for each disturbance such as a line outage. First, it performs time-domain simulation for a short term to obtain a post-disturbance trajectory. Second, along the trajectory, the system state with the local maximum potential energy is found and used as the initial point to search for a relevant unstable equilibrium by Newton's method. Third, the transient energy margin is estimated based on this unstable equilibrium to predict an out-of-step condition with generators. The proposed energy function-embedded quasi-steady-state model is tested in terms of its accuracy and time performance on an NPCC 140-bus power system and compared to a quasi-steady-state model embedding transient stability simulation.

Guo, Zhenping [Univ. of Tennessee, Knoxville, TN (↗