Search NASA⌕ Search

SEARCH · Search NASA

Results for “Model Counting”

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 379 records · Page 21

Signature Analysis Utilizing a Dynamic Molten Salt Reactor Model for MC&A

Moving from traditional fixed fuel nuclear reactor systems to a mobile, dynamic fuel system that has dissolved special nuclear material in a molten salt is a paradigm shift in several respects. One major consideration is how to develop effective nuclear material controls and accounting for these novel reactor systems. The Molten Salt Reactor Experiment is one example of a critical molten salt system and provided a significant reference library based on the documented effort. But it was low thermal power that was not intended to reflect a commercial scale electricity production design. Therefore, to facilitate and assist vendors with domestic licensing considerations, research is underway to identify methods that could be used for domestic safeguards approaches for these novel reactor systems. This research presents the results of a signature analysis of data generated from three simulated scenarios using the molten salt demonstration reactor model defined in the Transient Simulation Framework of Reconfigurable Modules. Each scenario provides 1 hour isotope inventories over a 180 day period. The scenarios investigated provide test cases to examine if direct gamma-ray spectroscopy of the fuel can be used to identify changes comparing a base case (no reactivity control and fixed fission contribution) to an insertion of reactivity (10 pcm no change in fission composition) and a change in fission composition. The analysis demonstrates that monitoring the total count rate in a highly collimated high-resolution photon energy spectrum is sensitive to perturbations imposed into the reactor model. The total photon count rate changes ≈2% for the fission composition change and ≈4.5% for the reactivity insertion compared to the base case. However, both scenarios show an increase in the total photon count rate. The total photon count rate can be used to identify changes due to power (number of fissions) but not due to a change in the material undergoing fission. To distinguish between the two cases of increased power, the photon spectrum would require an intensive analysis technique. A photon peak count rate ratio analysis could be used to identify changes in the fissile material fission generation in the core through identification of a static peak that shows little variation to the source of fission and a highly varying peak. The photon peak strength will ultimately be determined by the isotope’s fission yield. A preliminary analysis investigating the coupling of the isotope’s fission yield and its concentration in the fuel salt derived from the modeling has been performed. A ratio analysis of the photon peak count rates of 140 La to 99 Mo demonstrated that the reactivity insertion creates a distinct difference in the ratio compared to the fission composition change scenario.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Rapid neutron and gamma-ray source localization using machine learning

Rapid localization of radiation sources is critical for applications including nuclear emergency response, safeguards, and security. However, conventional imaging systems such as neutron scatter cameras and Compton cameras depend on rare coincidence events, which often result in long acquisition times. In this work, we address the challenge of rapid source localization by developing a machine learning approach to predict the direction of a single radiation source using only count rates from an array of neutron and gamma-ray detectors. The proposed model is a fully connected neural network (FCNN) trained using Monte Carlo simulation data from a 252 Cf source. The model hyperparameters are optimized with a small set of routine 252 Cf measurements. We benchmarked the performance of the trained and optimized machine learning model using additional 252 Cf , 137 Cs , and PuBe measurements under laboratory conditions with varying source-detector configurations. For these measurements, the machine learning model achieved a mean localization error smaller than 30° with 3 x 10 3 system counts, corresponding to 8 s measurement time for the imaging system used in this work. In this low-statistics regime, the method outperformed traditional scatter-based imaging by more than 75% in localization accuracy for the evaluated measurement configurations. These results demonstrate that a machine learning-based approach can significantly reduce the time required for accurate single-source localization, providing a robust and computationally efficient alternative to traditional imaging systems in time-critical nuclear security and emergency response scenarios.

Gamma-ray imaging↗

Stochastic Gain Degradation in III-V Heterojunction Bipolar Transistors due to Single Particle Displacement Damage

As device dimensions decrease single displacement effects are becoming more important. We measured the gain degradation in III-V Heterojunction Bipolar Transistors due to single particles using a heavy ion microbeam. Two devices with different sizes were irradiated with various ion species ranging from oxygen to gold to study the effect of the irradiation ion mass on the gain change. From the single steps in the inverse gain (which is proportional to the number of defects) we calculated Cumulative Distribution Functions to help determine design margins. The displacement process was modeled using the Marlowe Binary Collision Approximation (BCA) code. The entire structure of the device was modeled and the defects in the base-emitter junction were counted to be compared to the experimental results. While we found good agreement for the large device, we had to modify our model to reach reasonable agreement for the small device.

42 ENGINEERING↗

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↗

A Comparison of Infectious Disease Forecasting Methods across Locations, Diseases, and Time

Accurate infectious disease forecasting can inform efforts to prevent outbreaks and mitigate adverse impacts. This study compares the performance of statistical, machine learning (ML), and deep learning (DL) approaches in forecasting infectious disease incidences across different countries and time intervals. We forecasted three diverse diseases: campylobacteriosis, typhoid, and Q-fever, using a wide variety of features (n = 46) from public datasets, e.g., landscape, climate, and socioeconomic factors. We compared autoregressive statistical models to two tree-based ML models (extreme gradient boosted trees [XGB] and random forest [RF]) and two DL models (multi-layer perceptron and encoder–decoder model). The disease models were trained on data from seven different countries at the region-level between 2009–2017. Forecasting performance of all models was assessed using mean absolute error, root mean square error, and Poisson deviance across Australia, Israel, and the United States for the months of January through August of 2018. The overall model results were compared across diseases as well as various data splits, including country, regions with highest and lowest cases, and the forecasted months out (i.e., nowcasting, short-term, and long-term forecasting). Overall, the XGB models performed the best for all diseases and, in general, tree-based ML models performed the best when looking at data splits. There were a few instances where the statistical or DL models had minutely smaller error metrics for specific subsets of typhoid, which is a disease with very low case counts. Feature importance per disease was measured by using four tree-based ML models (i.e., XGB and RF with and without region name as a feature). The most important feature groups included previous case counts, region name, population counts and density, mortality causes of neonatal to under 5 years of age, sanitation factors, and elevation. This study demonstrates the power of ML approaches to incorporate a wide range of factors to forecast various diseases, regardless of location, more accurately than traditional statistical approaches.

59 BASIC BIOLOGICAL SCIENCES↗

Spatio-temporal Estimates of Disease Transmission Parameters for COVID-19 with a Fully-Coupled, County-Level Model of the United States

Sandia National Laboratories has developed a capability to estimate parameters of epidemiological models from case reporting data to support responses to the COVID-19 pandemic. A differentiating feature of this work is the ability to simultaneously estimate county-specific disease transmission parameters in a nation-wide model that considers mobility between counties. The approach is focused on estimating parameters in a stochastic SEIR model that considers mobility between model patches (i.e., counties) as well as additional infectious compartments. The inference engine developed by Sandia includes (1) reconstruction and (2) transmission parameter inference. Reconstruction involves estimating current population counts within each of the compartments in a modified SEIR model from reported case data. Reconstruction produces input for the inference formulations, and it provides initial conditions that can be used in other modeling and planning efforts. Inference involves the solution of a large-scale optimization problem to estimate the time profiles for the transmission parameters in each county. These provide quantification of changes in the transmission parameter over time (e.g., due to impact of intervention strategies). This capability has been implemented in a Python-based software package, epi_inference, that makes extensive use of Pyomo [5] and IPOPT [10] to formulate and solve the inference formulations.

42 ENGINEERING↗

Bayesian chain graph models to characterize microbe-environment dynamics

Microbiome data require statistical models that can simultaneously decode microbes' reaction to the environment and interactions among microbes. While a multiresponse linear regression model seems like a straight-forward solution, we argue that treating it as a graphical model is problematic given that the regression coefficient matrix does not encode the conditional dependence structure between response and predictor nodes. This observation is especially important in biological settings when we have prior knowledge on the edges from specific experimental interventions that can only be properly encoded under a conditional dependence model. Here, we propose a chain graph model with two sets of nodes (predictors and responses) whose solution yields a graph with edges that indeed represent conditional dependence, thus agreeing with the experimenter's intuition on the average behavior of nodes under treatment. The solution to our model is sparse via the Bayesian linear regression (LASSO). In addition, we propose an adaptive extension so that different shrinkages can be applied to different edges to incorporate edge-specific prior knowledge. Our model is computationally inexpensive through an efficient Gibbs sampling algorithm and can account for binary, counting, and compositional responses via an appropriate hierarchical structure. We test the performance of our model in a variety of simulated datasets, thereby showing superior performance to state-of-the-art approaches. We further apply our model to human gut and soil microbial compositional datasets, and we highlight that CG-LASSO can estimate biologically meaningful network structures in the data.

compositional data↗

From RNNs to Foundation Models: An Empirical Study on Commercial Building Energy Consumption

Accurate short-term energy consumption forecasting for commercial buildings is crucial for smart grid operations. While smart meters and deep learning models enable forecasting using past data from multiple buildings, data heterogeneity from diverse buildings can reduce model performance. The impact of increasing dataset heterogeneity in time series forecasting, while keeping size and model constant, is understudied. We tackle this issue using the ComStock dataset, which provides synthetic energy consumption data for U.S. commercial buildings. Two curated subsets, identical in size and region but differing in building type diversity, are used to assess the performance of various time series forecasting models, including finetuned open-source foundation models (FMs). The results show that dataset heterogeneity and model architecture have a greater impact on post-training forecasting performance than the parameter count. Moreover, despite the higher computational cost, finetuned FMs demonstrate competitive performance compared to base models trained from scratch.

commercial buildings↗

Systematic approach to leptogenesis in nonequilibrium QFT: Self-energy contribution to the CP-violating parameter

In the baryogenesis via leptogenesis scenario the self-energy contribution to the CP-violating parameter plays a very important role. Here, we calculate it in a simple toy model of leptogenesis using the Schwinger-Keldysh/Kadanoff-Baym formalism as starting point. We show that the formalism is free of the double-counting problem typical for the canonical Boltzmann approach. Within the toy model, medium effects increase the CP-violating parameter. In contrast to results obtained earlier in the framework of thermal field theory, the medium corrections are linear in the particle number densities. In the resonant regime quantum corrections lead to modified expressions for the CP-violating parameter and for the decay width. Most notably, in the maximal resonant regime the Boltzmann picture breaks down and an analysis in the full Kadanoff-Baym formalism is required.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Combining biomarker and virus phylogenetic models improves HIV-1 epidemiological source identification

To identify and stop active HIV transmission chains new epidemiological techniques are needed. Here, we describe the development of a multi-biomarker augmentation to phylogenetic inference of the underlying transmission history in a local population. HIV biomarkers are measurable biological quantities that have some relationship to the amount of time someone has been infected with HIV. To train our model, we used five biomarkers based on real data from serological assays, HIV sequence data, and target cell counts in longitudinally followed, untreated patients with known infection times. The biomarkers were modeled with a mixed effects framework to allow for patient specific variation and general trends, and fit to patient data using Markov Chain Monte Carlo (MCMC) methods. Subsequently, the density of the unobserved infection time conditional on observed biomarkers were obtained by integrating out the random effects from the model fit. This probabilistic information about infection times was incorporated into the likelihood function for the transmission history and phylogenetic tree reconstruction, informed by the HIV sequence data. To critically test our methodology, we developed a coalescent-based simulation framework that generates phylogenies and biomarkers given a specific or general transmission history. Testing on many epidemiological scenarios showed that biomarker augmented phylogenetics can reach 90% accuracy under idealized situations. Under realistic within-host HIV-1 evolution, involving substantial within-host diversification and frequent transmission of multiple lineages, the average accuracy was at about 50% in transmission clusters involving 5–50 hosts. Realistic biomarker data added on average 16 percentage points over using the phylogeny alone. Using more biomarkers improved the performance. Shorter temporal spacing between transmission events and increased transmission heterogeneity reduced reconstruction accuracy, but larger clusters were not harder to get right. More sequence data per infected host also improved accuracy. We show that the method is robust to incomplete sampling and that adding biomarkers improves reconstructions of real HIV-1 transmission histories. The technology presented here could allow for better prevention programs by providing data for locally informed and tailored strategies.

60 APPLIED LIFE SCIENCES↗

Dispersion Loss Counteracts Embedding Condensation and Improves Generalization in Small Language Models

Large language models (LLMs) achieve remarkable performance through ever-increasing parameter counts, but scaling incurs steep computational costs. To better understand LLM scaling, we study representational differences between LLMs and their smaller counterparts, with the goal of replicating the representational qualities of larger models in smaller models. We observe a geometric phenomenon which we term embedding condensation, where token embeddings collapse into a narrow cone-like subspace in some language models. Through systematic analyses across multiple Transformer families, we show that small models such as GPT2 and Qwen3-0.6B exhibit severe condensation, whereas larger models such as GPT2-x1 and Qwen3-32B are more resistant to this phenomenon. Additional observations show that embedding condensation is not reliably mitigated by knowledge distillation from larger models. To fight against it, we formulate a dispersion loss that explicitly encourages embedding dispersion during training. Experiments demonstrate that it mitigates condensation, recovers dispersion patterns seen in larger models, and yields performance gains across 10 benchmarks. We believe this work offers a principled path toward improving smaller Transformers without additional parameters.

Xiao, Xi [ORNL] (ORCID:0009000009316982)↗

Energy and particle balance during plasma detachment in a long-leg divertor configuration

Abstract Comprehensive studies of energy and particle balances in the transition to plasma detachment in an alternative divertor configuration with long outer legs are shown. Numerical simulations are performed with the 2D code suite SOLPS 4.3, using a disconnected double null grid with narrow, tightly baffled long poloidal leg divertors at the outer lower target and outer upper target. A particle count scan is performed using the ‘closed gas box’ model, where the tunable parameter in the simulations is the total number of deuterium particles in the simulation space and all other parameters are held fixed, including a constant input power and trace neon impurity radiation, to assess the physics of the transition to detachment in the system as the particle count increases. Three main aspects of the physics of divertor detachment are addressed: the criteria for the local onset of divertor detachment in each of the divertors, the distribution of heat flux and other plasma parameters between the four divertors as each divertor transitions to detachment, and the role of perpendicular transport in the transition to the detached regime. A synergistic mechanism by which the cross-field transport is reduced by factors associated with the onset of plasma recombination effects is identified. These results are compared to the existing understanding of the physics of the transition to plasma detachment in standard divertors.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

On microstates counting in many body polymer quantum systems

Polymer quantum systems are mechanical models quantized in a similar way as loop quantum gravity but in which loops/graphs resembling polymers are replaced by discrete sets of points. Such systems have allowed to study in a simpler context some novel aspects of loop quantum gravity. Although thermal aspects play a crucial role in cosmology and black hole physics little attention has been given to the thermostatistics of many body polymer quantum systems. In this work we explore how the features of a one-dimensional effective polymer gas, affect its microstate counting and hence the corresponding thermodynamical quantities.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Accelerated Immunodeficiency by Anti-CCR5 Treatment in HIV Infection

In 50% of progressing HIV-1 patients, CXCR4-tropic (X4) virus emerges late in infection, often overtaking CCR5-tropic (R5) virus as the dominant viral strain. This ‘‘phenotypic switch’’ is strongly associated with rapidly declining CD4+ T cell counts and AIDS onset, yet its causes remain unknown. Here, we analyze a mathematical model for the mechanism of X4 emergence in late-stage HIV infection and use this analysis to evaluate the utility of a promising new class of antiretroviral drugs—CCR5 inhibitors—in dual R5, X4 infection. The model shows that the R5-to-X4 switch occurs as CD4+ T cell activation levels increase above a threshold and as CD4+ T cell counts decrease below a threshold during late-stage HIV infection. Importantly, the model also shows that highly active antiretroviral therapy (HAART) can inhibit X4 emergence but that monotherapy with CCR5 blockers can accelerate X4 onset and immunodeficiency if X4 infection of memory CD4+ T cells occurs at a high rate. Fortunately, when CXCR4 blockers or HAART are used in conjunction with CCR5 blockers, this risk of accelerated immunodeficiency is eliminated. The results suggest that CCR5 blockers will be more effective when used in combination with CXCR4 blockers and caution against CCR5 blockers in the absence of an effective HAART regimen or during HAART failure.

60 APPLIED LIFE SCIENCES↗

A HYBRID MODEL FOR THE EVOLUTION OF GALAXIES AND ACTIVE GALACTIC NUCLEI IN THE INFRARED

We present a comprehensive investigation of the cosmological evolution of the luminosity function of galaxies and active galactic nuclei (AGNs) in the infrared (IR). Based on the observed dichotomy in the ages of stellar populations of early-type galaxies on one side and late-type galaxies on the other, the model interprets the epoch-dependent luminosity functions at z {>=} 1.5 using a physical approach for the evolution of proto-spheroidal galaxies and of the associated AGNs, while IR galaxies at z < 1.5 are interpreted as being mostly late-type ''cold'' (normal) and ''warm'' (starburst) galaxies. As for proto-spheroids, in addition to the epoch-dependent luminosity functions of stellar and AGN components separately, we have worked out, for the first time, the evolving luminosity functions of these objects as a whole (stellar plus AGN component), taking into account in a self-consistent way the variation with galactic age of the global spectral energy distribution. The model provides a physical explanation for the observed positive evolution of both galaxies and AGNs up to z {approx_equal} 2.5 and for the negative evolution at higher redshifts, for the sharp transition from Euclidean to extremely steep counts at (sub-)millimeter wavelengths, as well as the (sub-)millimeter counts of strongly lensed galaxies that are hard to account for by alternative, physical or phenomenological, approaches. The evolution of late-type galaxies and z < 1.5 AGNs is described using a parametric phenomenological approach. The modeled AGN contributions to the counts and to the cosmic infrared background (CIB) are always sub-dominant. They are maximal at mid-IR wavelengths: the contribution to the 15 and 24 {mu}m counts reaches 20% above 10 and 2 mJy, respectively, while the contributions to the CIB are of 8.6% and of 8.1% at 15 {mu}m and 24 {mu}m, respectively. The model provides a good fit to the multi-wavelength (from the mid-IR to millimeter waves) data on luminosity functions at different redshifts and on number counts (both global and per redshift slices). A prediction of the present model, useful to test it, is a systematic variation with wavelength of the populations dominating the counts and the contributions to the CIB intensity. This implies a specific trend for cross-wavelength CIB power spectra, which is found to be in good agreement with the data.

79 ASTRONOMY AND ASTROPHYSICS↗

High Force Spring Clamp System

This is the Final Technical Report for the work by Hyperboloid LLC using SBIR Grant # DE-SC0019579. The work responded to US DOE SBIR Topics FY2019, Phase 1, Release 1, August 13 2018, Topic: 30. a. (2) SRF cavity joining techniques that create vacuum seals without resorting to the use of bolted flanges in order to minimize particulates. Hyperboloid LLC designed, developed and tested the High Force Spring Clamp System characterized in Patent US 9756715 that is aimed at solving this topic. The patent is held by Thomas Jefferson National Accelerator Facility. The flange bolts are substituted with exceptionally high force, binder-clip-like “C” Clamps. The clamps are opened and applied by hydraulic based tooling called “Clamp Openers” using particle minimization methods. Model 1 Clamp reached the goal of sealing the SRF industry standard hexagonal section, aluminum alloy gasket, leak-free on a research cavity as determined by a superfluid helium challenge, at 2 K for 2 assembly cycles. Descriptions include the ANSYS calculations that determine the size and shape of the clamp, the details of the Clamp Opener construction, particle counts from sealing flanges using both bolts and clamps. Model 2 Clamp, at half the size, needing a softer metallic gasket to be successful, is more readily adaptable to Cryomodule designs. It was tested, but a needs soft gasket development to be successful.

43 PARTICLE ACCELERATORS↗