Search NASA⌕ Search

SEARCH · Search NASA

Results for “sparse 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.

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

At least 847 records · Page 47
All sourcesNASA NTRSDOE OSTI
NASA NTRS2017

Coupled Data Assimilation for Integrated Earth System Analysis and Prediction: Goals, Challenges, and Recommendations

The purpose of this report is to identify fundamental issues for coupled data assimilation (CDA), such as gaps in science and limitations in forecasting systems, in order to provide guidance to the World Meteorological Organization (WMO) on how to facilitate more rapid progress internationally. Coupled Earth system modeling provides the opportunity to extend skillful atmospheric forecasts beyond the traditional two-week barrier by extracting skill from low-frequency state components such as the land, ocean, and sea ice. More generally, coupled models are needed to support seamless prediction systems that span timescales from weather, subseasonal to seasonal (S2S), multiyear, and decadal. Therefore, initialization methods are needed for coupled Earth system models, either applied to each individual component (called Weakly Coupled Data Assimilation - WCDA) or applied the coupled Earth system model as a whole (called Strongly Coupled Data Assimilation - SCDA). Using CDA, in which model forecasts and potentially the state estimation are performed jointly, each model domain benefits from observations in other domains either directly using error covariance information known at the time of the analysis (SCDA), or indirectly through flux interactions at the model boundaries (WCDA). Because the non-atmospheric domains are generally under-observed compared to the atmosphere, CDA provides a significant advantage over single-domain analyses. Next, we provide a synopsis of goals, challenges, and recommendations to advance CDA: Goals: (a) Extend predictive skill beyond the current capability of NWP (e.g. as demonstrated by improving forecast skill scores), (b) produce physically consistent initial conditions for coupled numerical prediction systems and reanalyses (including consistent fluxes at the domain interfaces), (c) make best use of existing observations by allowing observations from each domain to influence and improve the full earth system analysis, (d) develop a robust observation-based identification and understanding of mechanisms that determine the variability of weather and climate, (e) identify critical weaknesses in coupled models and the earth observing system, (f) generate full-field estimates of unobserved or sparsely observed variables, (g) improve the estimation of the external forcings causing changes to climate, (h) transition successes from idealized CDA experiments to real-world applications. Challenges: (a) Modeling at the interfaces between interacting components of coupled Earth system models may be inadequate for estimating uncertainty or error covariances between domains, (b) current data assimilation methods may be insufficient to simultaneously analyze domains containing multiple spatiotemporal scales of interest, (c) there is no standardization of observation data or their delivery systems across domains, (d) the size and complexity of many large-scale coupled Earth system models makes it is difficult to accurately represent uncertainty due to model parameters and coupling parameters, (e) model errors lead to local biases that can transfer between the different Earth system components and lead to coupled model biases and long-term model drift, (e) information propagation across model components with different spatiotemporal scales is extremely complicated, and must be improved in current coupled modeling frameworks, (h) there is insufficient knowledge on how to represent evolving errors in non-atmospheric model components (e.g. as sea ice, land and ocean) on the timescales of NWP.

CDA↗
NASA NTRS2008

On Combining Thermal-Infrared and Radio-Occultation Data of Saturn's Atmosphere

Radio-occultation and thermal-infrared measurements are complementary investigations for sounding planetary atmospheres. The vertical resolution afforded by radio occultations is typically approximately 1 km or better, whereas that from infrared sounding is often comparable to a scale height. On the other hand, an instrument like CIRS can easily generate global maps of temperature and composition, whereas occultation soundings are usually distributed more sparsely. The starting point for radio-occultation inversions is determining the residual Doppler-shifted frequency, that is the shift in frequency from what it would be in the absence of the atmosphere. Hence the positions and relative velocities of the spacecraft, target atmosphere, and DSN receiving station must be known to high accuracy. It is not surprising that the inversions can be susceptible to sources of systematic errors. Stratospheric temperature profiles on Titan retrieved from Cassini radio occultations were found to be very susceptible to errors in the reconstructed spacecraft velocities (approximately equal to 1 mm/s). Here the ability to adjust the spacecraft ephemeris so that the profiles matched those retrieved from CIRS limb sounding proved to be critical in mitigating this error. A similar procedure can be used for Saturn, although the sensitivity of its retrieved profiles to this type of error seems to be smaller. One issue that has appeared in inverting the Cassini occultations by Saturn is the uncertainty in its equatorial bulge, that is, the shape in its iso-density surfaces at low latitudes. Typically one approximates that surface as a geopotential surface by assuming a barotropic atmosphere. However, the recent controversy in the equatorial winds, i.e., whether they changed between the Voyager (1981) era and later (after 1996) epochs of Cassini and some Hubble observations, has made it difficult to know the exact shape of the surface, and it leads to uncertainties in the retrieved temperature profiles of one to a few kelvins. This propagates into errors in the retrieved helium abundance, which makes use of thermal-infrared spectra and synthetic spectra computed with retrieved radio-occultation temperature profiles. The highest abundances are retrieved with the faster Voyager-era winds, but even these abundances are somewhat smaller than those retrieved from the thermal-infrared data alone (albeit with larger formal errors). The helium abundance determination is most sensitive to temperatures in the upper troposphere. Further progress may include matching the radio-occultation profiles with those from CIRS limb sounding in the upper stratosphere.

Flasar, F. M.↗
NASA NTRS2005

Construction of Covariance Functions with Variable Length Fields

This article focuses on construction, directly in physical space, of three-dimensional covariance functions parametrized by a tunable length field, and on an application of this theory to reproduce the Quasi-Biennial Oscillation (QBO) in the Goddard Earth Observing System, Version 4 (GEOS-4) data assimilation system. These Covariance models are referred to as multi-level or nonseparable, to associate them with the application where a multi-level covariance with a large troposphere to stratosphere length field gradient is used to reproduce the QBO from sparse radiosonde observations in the tropical lower stratosphere. The multi-level covariance functions extend well-known single level covariance functions depending only on a length scale. Generalizations of the first- and third-order autoregressive covariances in three dimensions are given, providing multi-level covariances with zero and three derivatives at zero separation, respectively. Multi-level piecewise rational covariances with two continuous derivatives at zero separation are also provided. Multi-level powerlaw covariances are constructed with continuous derivatives of all orders. Additional multi-level covariance functions are constructed using the Schur product of single and multi-level covariance functions. A multi-level powerlaw covariance used to reproduce the QBO in GEOS-4 is described along with details of the assimilation experiments. The new covariance model is shown to represent the vertical wind shear associated with the QBO much more effectively than in the baseline GEOS-4 system.

Gaspari, Gregory↗
NASA NTRS2005

Biomimetic Models for An Ecological Approach to Massively-Deployed Sensor Networks

Promises of ubiquitous control of the physical environment by massively-deployed wireless sensor networks open avenues for new applications that will redefine the way we live and work. Due to small size and low cost of sensor devices, visionaries promise systems enabled by deployment of massive numbers of sensors ubiquitous throughout our environment working in concert. Recent research has concentrated on developing techniques for performing relatively simple tasks with minimal energy expense, assuming some form of centralized control. Unfortunately, centralized control is not conducive to parallel activities and does not scale to massive size networks. Execution of simple tasks in sparse networks will not lead to the sophisticated applications predicted. We propose a new way of looking at massively-deployed sensor networks, motivated by lessons learned from the way biological ecosystems are organized. We demonstrate that in such a model, fully distributed data aggregation can be performed in a scalable fashion in massively deployed sensor networks, where motes operate on local information, making local decisions that are aggregated across the network to achieve globally-meaningful effects. We show that such architectures may be used to facilitate communication and synchronization in a fault-tolerant manner, while balancing workload and required energy expenditure throughout the network.

Jones, Kennie H.↗
NASA NTRSDate not supplied

Development of a Global Reference Surface Reflectance and BRDF Datasets from Geostationary Satellite Observations and AERONET Measurements

Surface reflectances and their dependency on illumination-view geometries (i.e., BRDF) are the foundation of many high-level satellite products for land and water monitoring. Yet it is difficult to evaluate the quality of satellite-based surface reflectances with ground-based measurements due to the spatial scale differences. In order to fill the gap, here we develop a reference dataset of surface reflectance and BRDF at the global AERONET sites with data streams from operational geostationary sensors including Himawari 8/9 AHI, GK-2A AMI, and GOES 16/17 ABI. Taking the top-of-atmosphere (TOA) reflectance and the site measured atmospheric aerosol optical depth (AOD) as the main inputs, we apply the GeoNEX-AC algorithm to performance accurate atmospheric correction and derive 10-minute surface reflectance and daily Ross-Thick-Li-Sparse (RTLS) BRDF parameters at AERONET sites where coincident AOD measurements and TOA observations are available from 2016 (for Himawari) or 2018 (for GOES) onwards. The algorithm ensures that the retrieved surface BRDF parameters, along with the site-measured AOD, allow the atmospheric radiative transfer model, SHARM, accurately simulate the observed TOA reflectance at diurnal and longer time scales. They are our best estimates of the surface optical properties and thus can serve as the “reference” to evaluate the performance of operational atmospheric correction algorithms (where AOD is assumed unknown and needs to be retrieved). The reference BRDF also allow us to evaluate the spectral band ratios between the SWIR (e.g., 2200 nm) and the visible (e.g., 650 nm) regions, which are commonly used in operational atmospheric correction algorithms. Finally, we demonstrate that the reference dataset can be used to develop potential data synergies between different GEO satellites as well as GEO-LEO sensors.

Weile Wang↗
DOE OSTI2026

ChatMPI: LLM-Driven MPI Code Generation for HPC Workloads

The Message Passing Interface (MPI) standard plays a crucial role in enabling scientific applications for parallel computing and is an essential component in high-performance computing (HPC). However, implementing MPI code manually—especially applying a proper domain decomposition and communication pattern—is a challenging and error-prone task. We present ChatMPI, an AI assistant for MPI parallelization of sequential C codes. In our analysis, we focus on testing six essential HPC workloads, which are based on Basic Linear Algebra Subprograms levels 1, 2, and 3 as well as sparse, stencil, and iterative operations. We analyze the process of creating ChatMPI by using the ChatHPC library. This lightweight large language model (LLM)–based infrastructure enables HPC experts to efficiently create and supervise trustworthy AI capabilities for critical HPC software tasks. We study the data required for training (fine-tuning) ChatMPI to generate parallel codes that not only use MPI syntax correctly but also apply HPC techniques to reduce memory communication and maximize performance by using proper work decomposition. With a relatively small training dataset composed of a few dozen prompts and fewer than 15 minutes of fine-tuning on one node equipped with two NVIDIA H100 GPUs, ChatMPI elevates trustworthiness for MPI code generation of current LLMs (e.g., Code Llama, ChatGPT-4o and ChatGPT 5). Additionally, we evaluate the performance of the MPI codes generated by ChatMPI in comparison with the ones generated by ChatGPT-4o and ChatGPT-5. The codes generated by ChatMPI provide up to a 4 × boost in performance by using better problem decomposition, communication patterns, and HPC techniques (e.g., communication avoiding).

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗
NASA NTRS2004

Precision X-ray Timing of RX J0806.3+1527 with CHANDRA: Evidence for Gravitational Radiation from an Ultracompact Binary

RX J0806.3+1527 is a candidate double degenerate binary with possibly the shortest known orbital period. The source shows an approximately equal to 100% X-ray intensity modulation at the putative orbital frequency of 3.11 mHz (321.5 s). If the system is a detached, ultracompact binary gravitational radiation should drive spin-up with a magnitude of nu(sup dot) approximately 10(exp -16) Hz per second. Efforts to constrain the X-ray frequency evolution to date have met with mixed success, principally due to the sparseness of earlier observations. Here we describe the results of the first phase coherent X-ray monitoring campaign on RX J0806.3+1527 with Chandra. We obtained a total of 70 ksec of exposure in 6 epochs logarithmically spaced over 320 days. With these data we conclusively show that the X-ray frequency is increasing at a rate of 3.77 plus or minus 0.8 x 10(exp -16) Hz per second. Using the ephemeris derived from the new data we are able to phase up all the earlier Chandra and ROSAT data and show they are consistent with a constant nu(sup dot) = 3.63 plus or minus 0.06 x 10(exp -16) Hz per second over the past decade. This value appears consistent with that recently derived by Israel et al. largely from monitoring of the optical modulation, and is in rough agreement with the solutions reported initially by Hakala et al., based on ground-based optical observations. The large and stable nu(sup dot) over a decade is consistent with gravitational radiation losses driving the evolution. An intermediate polar (IP) scenario where the observed X-ray period is the spin period of an accreting white dwarf appears less tenable because the observed nu(sup dot) requires an m(sup dot) approximately equal to 4 x 10 (exp -8) solar mass yr(sup -l), that is much larger than that inferred from the observed X-ray luminosity (although this depends on the uncertain distance and bolometric corrections), and it is difficult to drive such a high m(sup dot) in a binary system with parameters consistent with all the multiwavelength data. If the ultracompact scenario is correct, then the X-ray flux cannot be powered by stable accretion which would drive the components apart, suggesting a new type of energy source (perhaps electromagnetic) may power the X-ray flux.

Strohymayer, Tod E.↗
NASA NTRS2009

Large Area Lunar Dust Flux Measurement Instrument

The instrument under development is designed to characterize the flux and size distribution of the lunar micrometeoroid and secondary ejecta environment. When deployed on the lunar surface, the data collected will benefit fundamental lunar science as well as enabling more reliable impact risk assessments for human lunar exploration activities. To perform this task, the instrument requirements are demanding. It must have as large a surface area as possible to sample the very sparse population of the larger potentially damage-inducing micrometeorites. It must also have very high sensitivity to enable it to measure the flux of small (<10 micron) micrometeorite and secondary ejecta dust particles. To be delivered to the lunar surface, it must also be very low mass, rugged and stow compactly. The instrument designed to meet these requirements is called FOMIS. It is a large-area thin film under tension (i.e. a drum) with multiple fiber optic displacement (FOD) sensors to monitor displacements of the film. This sensor was chosen since it can measure displacements over a wide dynamic range: 1 cm to sub-Angstrom. A prototype system was successfully demonstrated using the hypervelocity impact test facility at the University of Kent (Canterbury, UK). Based on these results, the prototype system can detect hypervelocity (approx.5 km/s) impacts by particles as small as 2 microns diameter. Additional tests using slow speeds find that it can detect secondary ejecta particles (which do not penetrate the film) with momentums as small as 15 pico-gram 100m/s, or nominally 5 microns diameter at 100 m/s.

Corsaro, R.↗
NASA NTRS2023

Smc-Last Point Source Catalog Spitzer'S Last Survey of the Small Magellanic Cloud

We present a processing for the point source catalog derived from Spitzer's final two surveys of the Small Magellanic Cloud (SMC). These 2017-2018 maps duplicate the footprint from the SAGE-SMC program from 2008 and were obtained at 3.6 and 4.5 μm with the Infrared Array Camera (IRAC). They provide a nearly 10-year baseline of all observations of the SMC made by Spitzer. We extracted point sources from both the basic calibrated data (BCD) as well as the final processed mosaics. We did not use any prior catalogs as inputs in order to search for moving objects (e.g., foreground brown dwarfs) and other transient phenomena. The extractions from two test fields are shown here: a field sparsely populated with stars and a second containing more complicated structures. Comparing our extractions between both of our epochs and with the SAGE epochs shows the completeness and range of our survey. The mosaics are currently available on the Infrared Science Archive (IRSA) and the Harvard Dataverse. The final point source catalog will be available at both archives, as well. This work is supported by NASA Award 80NSSC19K0585.

Thomas Kuchar↗
NASA NTRS2005

Precision X-ray Timing of RX J0806.3+1527 with CHANDRA: Evidence for Gravitational Radiation from an Ultracompact Binary

RX J0806.3+1527 is a candidate double degenerate binary with possibly the shortest known orbital period. The source shows an $\approx 100 \%$ X-ray intensity modulation at the putative orbital frequency of 3.11 mHz (321.5 s). If the system is a detached, ultracompact binary gravitational radiation should drive spin-up with a magnitude of $\dot\nu \sim 10(exp -16)$ Hi s$(exp -l}$. Efforts to constrain the X-ray frequency evolution to date have met with mixed success, principally due to the sparseness of earlier observations. Here we describe the results of the first phase coherent X-ray monitoring campaign on RX J0806.3+1527 with {\it Chandra) . We obtained a total of 70 ksec of exposure in 6 epochs logarithmically spaced over 320 days. With these data we conclusively show that the X-ray frequency is increasing at a rate of $3.77 \pm\ 0.8 \times l0(exp -16)$ Hz s$(exp -l)$. Using the ephemeris derived from the new data we are able to phase up all the earlier {\it Chandra) and ROSAT data and show they are consistent with a constant $\dot\nu = 3.63 \pm 0.06 \times 10(exp -16)$ Hz s$(exp -1)s over the past decade. This value appears consistent with that recently derived by Israel et al. largely from monitoring of the optical modulation, and is in rough agreement with the solutions reported initially by Hakala et al., based on ground-based optical observations. The large and stable $\dot\nu$ over a decade is consistent with gravitational radiation losses driving the evolution. An intermediate polar (IP) scenario where the observed X-ray period is the spin period of an accreting white dwarf appears less tenable because the observed $\dot\nu$ requires an $\dot m \approx 2 \times 10(exp -8)$ $M-{\odot)$ yr$(exp -l)$, that is much larger than that inferred from the observed X-ray luminosity (although this depends on the uncertain distance and bolometric corrections), and it is difficult to drive such a high $\dot m$ in a binary system with parameters consistent with all the multiwavelength data. If the ultracompact scenario is

Strohmayer, Tod↗
NASA NTRS2020

Lunar Magnetic Field Models From Lunar Prospector and SELENE/Kaguya Along-Track Magnetic Field Gradients

We use L1-norm model regularization of |Br| component at the surface on magnetic monopoles bases and along-track magnetic field differences alone (without vector observations) to derive high quality global magnetic field models at the surface of the Moon. The practical advantages to this strategy are the following: monopoles are more stable at closer spacing in comparison to dipoles, improving spatial resolution; L1-norm model regularization leads to sparse models which may be appropriate for the Moon which has regions of localized magnetic field features; and along-track differences reduce the need for ad-hoc external field noise reduction strategies. We examine also the use of Lunar Prospector and SELENE/Kaguya magnetometer data, combined and separately, and find that the Lunar Prospector along-track vector field differences lead to surface field models that require weaker regularization and, hence, result in higher spatial resolution. Significantly higher spatial resolution (wavelengths of roughly 25–30 km) and higher amplitude surface magnetic fields can be derived over localized regions of high amplitude anomalies (due to their higher signal-to-noise ratio). These high-resolution field models are also compared with the results of Surface Vector Mapping approach of Tsunakawa et al. (2015, https://doi.org/10.1002/2014JE004785). Finally, the monopoles- as well as dipoles-based patterns of the Serenitatis high amplitude magnetic feature have characteristic textbook patterns of Br and Bθ component fields from a nearly vertically downwardly magnetized source region and it implies that the principal source of the anomaly was formed when the region was much closer to the north magnetic pole of the Moon.

D. Ravat↗
NASA NTRS2022

Midwest Water Resources: Developing an Evapotranspiration Climatology to Analyze Spatiotemporal Water Budget Patterns for Agriculture and Natural Resources Managers in the Midwest

Evapotranspiration (ET) is a climatic variable critical to the hydrologic cycle and is used to evaluate spatiotemporal trends in drought conditions. Although in-situ observations provide accurate ET information, these records are spatially sparse. The United States Department of Agriculture (USDA) Midwest Climate Hub, National Integrated Drought Information System, Minnesota Department of Agriculture, Michigan State University, and the USDA Foreign Agricultural Service have partnered with DEVELOP to gain new insights on spatiotemporal patterns of ET with NASA satellite data. This project evaluates the feasibility of using remotely sensed ET data products to understand trends from 2001 through 2020. Actual ET (aET) data were sourced from NASA’s Terra Moderate Resolution Imaging Spectroradiometer (MODIS), and reference ET (refET) and precipitation data were sourced from the Gridded Surface Meteorological Dataset (gridMET). Spatial and temporal differences in the MODIS and gridMET data sets were resolved such that gridded data could be compared. These data sets were then used to produce monthly Normals maps of mm/8-day of precipitation, aET, and refET. Using precipitation and ET Normals maps, hydrologic state maps were produced by subtracting either refET or aET from precipitation. These hydrologic state maps provide a proxy water balance by summarizing the difference between water entering and leaving the ground surface. Additionally, the ET products were compared in timeseries plots. The ET products from this project provided partners with comparable datasets to assess potential drought and flooding conditions to support Midwest agricultural and natural resource managers in decision-making.

Emma Myrick↗
NASA NTRS1993

The North Atlantic circulation in the early 1980s - An estimate from inversion of a finite-difference model

A finite-difference model of the North Atlantic is constructed for the purpose of making an estimate of the circulation through an inverse calculation. The data base is eclectic, and includes hydrography, oxygen, nutrients, current meter and float records, atmospheric momentum, heat and water vapor transfers, as well as estimates of certain integral fluxes. Owing to the available hydrographic database, the model resolution is restricted to 1 deg at best, and is much coarser in many aspects. This limited resolution is a major obstacle to accurate estimates of climatological fluxes. In its final form, there are about 9000 constraints in 29,000 formal unknowns plus 9000 noise unknowns. The system is solved as a tapered least-squares system by a sparse conjugate gradient algorithm. With the exception of a few float velocities, all constraints are found to be consistent within error estimates. The model produces estimates of large-scale fluxes and flux divergences for all conventional properties including heat and nutrients as well as carbon dioxide and alkalinity. Meridional fluxes of carbon are found to be indistinguishable from zero, whereas the North Atlantic tends to export nutrients to the south, but carry heat to the north. Traditional oceanographic depictions of the circulation through combination of nonsynoptic data into steady models may have reached their useful limit in the present calculation, as the conflicts between the data and physical requirements become quantitatively apparent.

Martel, Francoise↗
NASA NTRS2004

Investigating the Impact of UV Radiation on High-Altitude Shallow Lake Habitats, Life Diversity, and Life Survival Strategies: Clues for Mars' Past Habitability Potential?

We present data and results from an ongoing project of astrobiological high-altitude expeditions investigating the highest and least explored perennial lakes on Earth in the Bolivian and Chilean Andes, including several volcanic crater lakes nearing and beyond 6,000 m in elevation. In the next five years, they will provide the first integrated long-term astrobiological characterization and monitoring of lacustrine environments and their biology for such altitude. These extreme lakes are natural laboratories. They provide the field data missing beyond 4,000 m to complete our understanding of terrestrial lakes and biota. Research on the effects of UV has been performed in lower altitude lakes and models of UV flux over time are being developed. Lakes showing a high content of dissolved organic material (DOM) shield organisms from UV. DOM acts as a natural sunscreen as it influences the water transparency, therefore is a determinant of photic zone depth. In sparsely vegetated alpine areas, lakes are clearer and offer less protection from UV to organisms living in the water. Transparent water and high UV irradiance may maximize the penetration and effect of UV radiation. Shallow-water communities in these lakes are particularly sensitive to UV radiation. The periphyton can live on various susbtrates. While on rocks, it includes immobile species that cannot seek low UV refuges unlike sediment-dwelling periphyton or alpine phytoflagellates which undergo vertical migration. Inhibition of algal photosynthesis by UV radiation has been documented in laboratory and showed that phytoplankton production is reduced by formation of nucleic acid lesions or production of peroxides and free oxygen radicals. of peroxides and free oxygen radicals. Our project is providing the field data that is missing from natural laboratories beyond 4,000 m and will complement the vision of the effects of UV on life and its adaptation modes (or lack thereof).

Cabrol, A.↗
DOE OSTI2024

DS-GL: Advancing Graph Learning via Harnessing the Power of Nature within Dynamic Systems

With the rapid digitization of the world, an increasing number of real-world applications are turning to nonEuclidean data, modeled as graphs. Due to their intrinsic high complexity and irregularity, learning from graph data demands tremendous computational power. Recently, CMOS-compatible Ising machines, i.e., dynamic systems composed of CMOS components, have emerged as a new approach that harnesses the inherent power of natural annealing within dynamic systems to efficiently resolve binary optimization problems and have been adopted for traditional graph computation, such as max-cut. However, when performing complex Graph Learning (GL) tasks, Ising machines face significant hurdles: (i) they are inherently binary and thus ill-suited for real-valued problems; (ii) their expensive all-to-all coupling network that guarantees effective natural annealing poses daunting scalability concerns. To address these challenges, this paper proposes a nature-powered graph learning framework dubbed DS-GL, which is the first effort to transform the process of solving graph learning problems into the natural annealing process within a parameterized dynamic system embodied as a CMOS chip. To tackle the two major hurdles, DS-GL first augments the Ising machine architecture to modify the self-reaction term of its Hamiltonian function from linear to quadratic, effectively serving as an energy regulator. This adjustment maintains the system’s original physical interpretation while enabling it to process continuous, real-valued data. Second, to address the scaling issue, DS-GL further upgrades the real-valued dense Ising machine by decomposing it into a mesh-based multi-PE dynamic system that supports efficient distributed spatial-temporal co-annealing across different PEs through sparse interconnects. By exploiting the inherent sparsity and component structures in real-world graphs, DS-GL is able to map complex graph learning tasks onto the scalable dynamic system while maintaining high accuracy. Evaluations with three diverse GL applications across six real-world datasets, including traffic flow and COVID-19 prediction, show that DS-GL can deliver from 102× to 106× speedups and 500× energy reduction over Graph Neural Networks on GPUs, with 5% - 20% accuracy enhancement.

Song, Ruibing↗
NASA NTRS1985

Modeling the radiant transfers of sparse vegetation canopies

The scattering dynamics of sparse vegetation canopies were studied within the framework of the three-dimensional radiative transfer model of Kimes (1984). The model was upgraded by including an algorithm for the anisotropic scattering of a soil boundary. Validation of the model was carried out using measured directional reflectance data for two canopies exhibiting typical scattering behavior with low and intermediate vegetation density. The canopies were: an orchard grass canopy; and a hard wheat canopy. A number of factors were found contributing to the final reflectance distribution of the canopies, including: (1) the strong anisotropic scattering properties of the soil; (2) the geometric effect of the vegetation probability gap function on the soil anisotropy and solar irradiance; and (3) the anisotropic scattering of vegetation which is controlled by the phase function and the layering of leaves. The application of the theoretical results to the development of earth-observing sensor systems is discussed.

Kimes, D. S.↗
DOE OSTI2024

Extreme sparsification of physics-augmented neural networks for interpretable model discovery in mechanics

Data-driven constitutive modeling with neural networks has received increased interest in recent years due to its ability to easily incorporate physical and mechanistic constraints and to overcome the challenging and time-consuming task of formulating phenomenological constitutive laws that can accurately capture the observed material response. However, even though neural network-based constitutive laws have been shown to generalize proficiently, the generated representations are not easily interpretable due to their high number of trainable parameters. Sparse regression approaches exist that allow for obtaining interpretable expressions, but the user is tasked with creating a library of model forms which by construction limits their expressiveness to the functional forms provided in the libraries. Here, in this work, we propose to train regularized physics-augmented neural network-based constitutive models utilizing a smoothed version of $L^0$-regularization. This aims to maintain the trustworthiness inherited by the physical constraints, but also enables interpretability which has not been possible thus far on any type of machine learning-based constitutive model where model forms were not assumed a priori but were actually discovered. During the training process, the network simultaneously fits the training data and penalizes the number of active parameters, while also ensuring constitutive constraints such as thermodynamic consistency. We show that the method can reliably obtain interpretable and trustworthy constitutive models for compressible and incompressible hyperelasticity, yield functions, and hardening models for elastoplasticity, using synthetic and experimental data. This work aims to set a new paradigm for interpretable machine learning models in the broad area of solid mechanics where low and limited data is available along with prior knowledge of physical constraints that the learned maps need to obey. This paradigm can potentially be extended to a broader spectrum of scientific exploration.

Data-driven constitutive models↗
NASA NTRS1980

Investigation of the application of HCMM thermal data to snow hydrology

The author has identified the following significant results. Comparison of the thermal IR band temperatures over the rapidly melting snowcover in a number of locations in the Salt-Verde Arizona watershed by the U-2 high altitude multispectral scanner, with temperatures for these same locations measured by the corresponding HCMM pass, indicate that the U-2 temperatures are typically 5 C higher than the values analyzed from the HCMM infrared digital printout. Results tend to substantiate findings of other investigators that with the offset that has been applied to all HCMM data, the HCMM temperature values may, in fact, actually be 5 C too low. Analysis of differences between the HCMM day and night temperatures for selected snowcovered areas in the Sierra Nevada Mountains, derived from the daytime and nighttime printouts, show greater delta tau values for the sparsely vegetated higher elevations than for the lower elevation, more densely forested terrain. The differences for the 12 hour sequence are also greater in most instances than the 36 hour difference.

Barnes, J. C.↗