Efficient Bayesian Inference Approaches for Matrix and Tensor Learning: Applications to Surrogate Modeling of Complex Systems .
Abstract not provided.
De, Saibal↗
SEARCH · Engineering Papers
Results for “model inference”
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.
Simple and Sophisticated Models of Taylor's Cylinder Impact
We present a study based on the simplest possible models for Taylor’s cylinder impact problem, in addition to examination of using convolutional neural networks (CNNs) to map cylinder profiles to strength calibrations. We find that the approximate treatments of Taylor and Hawkyard compare well with hydrodynamic simulations using an equivalent assumption of constant flow stress. The CNN models prove to be well suited to successfully infer parameterizations of the Preston-Tonks-Wallace model of plastic deformation based on the deformed profile of an impacted cylinder
36 MATERIALS SCIENCE↗
Data reduction through optimized scalar quantization for more compact neural networks
Raw data generation for several existing and planned large physics experiments now exceeds TB/s rates, generating untenable data sets in very little time. Those data often demonstrate high dimensionality while containing limited information. Meanwhile, Machine Learning algorithms are now becoming an essential part of data processing and data analysis. Those algorithms can be used offline for post processing and post data analysis, or they can be used online for real time processing providing ultra low latency experiment monitoring. Both use cases would benefit from data throughput reduction while preserving relevant information: one by reducing the offline storage requirements by several orders of magnitude and the other by allowing ultra fast online inferencing with low complexity Machine Learning models. Moreover, reducing the data source throughput also reduces material cost, power and data management requirements. In this work we demonstrate optimized nonuniform scalar quantization for data source reduction. This data reduction allows lower dimensional representations while preserving the relevant information of the data, thus enabling high accuracy Tiny Machine Learning classifier models for online fast inferences. We demonstrate this approach with an initial proof of concept targeting the CookieBox, an array of electron spectrometers used for angular streaking, that was developed for LCLS-II as an online beam diagnostic tool. We used the Lloyd-Max algorithm with the CookieBox dataset to design an optimized nonuniform scalar quantizer. Optimized quantization lets us reduce input data volume by 69% with no significant impact on inference accuracy. When we tolerate a 2% loss on inference accuracy, we achieved 81% of input data reduction. Finally, the change from a 7-bit to a 3-bit input data quantization reduces our neural network size by 38%.
97 MATHEMATICS AND COMPUTING↗
A massive molecular gas reservoir in the z = 2.221 type-2 quasar host galaxy SMM J0939+8315 lensed by the radio galaxy 3C220.3
We report the detection of CO(J = 3 → 2) line emission in the strongly lensed submillimeter galaxy (SMG) SMM J0939+8315 at z = 2.221, using the Combined Array for Research in Millimeter-wave Astronomy. SMM J0939+8315 hosts a type-2 quasar, and is gravitationally lensed by the radio galaxy 3C220.3 and its companion galaxy at z = 0.685. The 104 GHz continuum emission underlying the CO line is detected toward 3C220.3 with an integrated flux density of S{sub cont} = 7.4 ± 1.4 mJy. Using the CO(J = 3 → 2) line intensity of I{sub CO(3-2)} = (12.6 ± 2.0) Jy km s{sup −1}, we derive a lensing- and excitation-corrected CO line luminosity of L{sub CO(1−0)}{sup ′} = (3.4 ± 0.7) × 10{sup 10} (10.1/μ{sub L}) K km s{sup −1} pc{sup 2} for the SMG, where μ{sub L} is the lensing magnification factor inferred from our lens modeling. This translates to a molecular gas mass of M{sub gas} = (2.7 ± 0.6) × 10{sup 10} (10.1/μ{sub L}) M{sub ⊙}. Fitting spectral energy distribution models to the (sub)-millimeter data of this SMG yields a dust temperature of T = 63.1{sub −1.3}{sup +1.1} K, a dust mass of M{sub dust} = (5.2 ± 2.1) × 10{sup 8} (10.1/μ{sub L}) M{sub ⊙}, and a total infrared luminosity of L{sub IR} = (9.1 ± 1.2) ×10{sup 12} (10.1/μ{sub L}) L{sub ⊙}. We find that the properties of the interstellar medium of SMM J0939+8315 overlap with both SMGs and type-2 quasars. Hence, SMM J0939+8315 may be transitioning from a starbursting phase to an unobscured quasar phase as described by the “evolutionary link” model, according to which this system may represent an intermediate stage in the evolution of present-day galaxies at an earlier epoch.
79 ASTRONOMY AND ASTROPHYSICS↗
Data from: Nitrification is a minor source of nitrous oxide (N2O) in an agricultural landscape and declines with increasing management intensity
The long-term contribution of nitrification to nitrous oxide (N2O) emissions from terrestrial ecosystems is poorly known and thus poorly constrained in biogeochemical models. Here, using Bayesian inference to couple 25 years of in situ N2O flux measurements with site-specific Michaelis-Menten kinetics of nitrification-derived N2O, we test the relative importance of nitrification-derived N2O across six cropped and unmanaged ecosystems along a management intensity gradient in the U.S. Midwest. We found that the maximum potential contribution from nitrification to in situ N2O fluxes was 13-17% in a conventionally fertilized annual cropping system, 27-42% in a low-input cover-cropped annual cropping system, and 52-63% in perennial systems including a late successional deciduous forest. Actual values are likely to be less than 10% of these values because of low N2O yields in cultured nitrifiers (typically 0.04 to 8% of NH3 oxidized) and competing sinks for available NH4+ in situ. Most nitrification-derived N2O was produced by ammonia oxidizing bacteria (AOB) rather than archaea (AOA), who appeared responsible for no more than 30% of nitrification-derived N2O production in all but one ecosystem. Although the proportion of nitrification-derived N2O production was lowest in annual cropping systems, these ecosystems nevertheless produced more nitrification-derived N2O (higher Vmax) than perennial and successional ecosystems. We conclude that nitrification is minor relative to other sources of N2O in all ecosystems examined.
54 ENVIRONMENTAL SCIENCES↗
Arctic Mixed-Phase Cloud Base Ice Precipitation Properties During the M-PACE Field Campaign
Cloud-climate feedbacks are still the greatest source of uncertainty in current climate projections. Arctic clouds, which are predominantly stratiform and supercooled, often long-lived, and nearly-continuously precipitate ice particles, contribute roughly 10% of the uncertainty attributed to the global cloud feedback. This Arctic cloud uncertainty is driven by incomplete observational and theoretical knowledge required to estimate and explain the state and active processes occurring in those clouds. A focus on ice precipitation properties at Arctic cloud base rather than the surface deconfounds the product of cloud condensate sink processes from the influence of the atmospheric thermodynamic state below cloud base, rendering cloud-base properties a more appealing target for inference and evaluation of model simulations. This dataset provides a set of 25 samples from the M-PACE field campaign, all of which were retrieved using the synthesis of ARM radar and lidar measurements. The retrieved ice precipitation variables in this dataset include, among others, the ice number concentration, water content, PSD parameters, precipitation rate, mass-weighted fall velocity, vertical air motion, and effective radius, all of which are highly valuable for model evaluation and a general understanding of polar cloud sink processes. Each variable sample includes its mean value and associated uncertainty. Additional variables based on ARM measurements (liquid layer statistics, etc.) are included in this dataset as well. The retrieval algorithm and analysis of this dataset are described in Silber (JGR, 2023, https://doi.org/10.1029/2022JD038202).
54 ENVIRONMENTAL SCIENCES↗
Arctic Mixed-Phase Cloud Base Ice Precipitation Properties Over the NSA Site
Cloud-climate feedbacks are still the greatest source of uncertainty in current climate projections. Arctic clouds, which are predominantly stratiform and supercooled, often long-lived, and nearly continuously precipitate ice particles, contribute roughly 10% of the uncertainty attributed to the global cloud feedback. This arctic cloud uncertainty is driven by incomplete observational and theoretical knowledge required to estimate and explain the state and active processes occurring in those clouds. A focus on ice precipitation properties at arctic cloud base rather than the surface deconfounds the product of cloud condensate sink processes from the influence of the atmospheric thermodynamic state below cloud base, rendering cloud-base properties a more appealing target for inference and evaluation of model simulations. This data set provides more than 1800 samples of cloud-base ice precipitation properties over Utqiagvik, North Slope of Alaska, all of which were retrieved using the synthesis of ARM radar and lidar measurements. The retrieved ice precipitation variables in this data set include, among others, the ice number concentration, water content, PSD parameters, precipitation rate, mass-weighted fall velocity, vertical air motion, and effective radius, all of which are highly valuable for model evaluation and a general understanding of polar cloud sink processes. Each variable sample includes its mean value and associated uncertainty. Additional variables based on ARM measurements (liquid layer statistics, etc.) are included in this data set. The retrieval algorithm and analysis of this data set are described in Silber (JGR, 2023, https://doi.org/10.1029/2022JD038202).
54 ENVIRONMENTAL SCIENCES↗
On fracture phenomena in advanced fiber composite materials.
The extension of linear elastic fracture mechanics (LEFM) from metallic alloys to advanced fiber composite laminates is considered. LEFM is shown to be valid for both isotropic and anisotropic homogeneous continua; the applicability of LEFM to advanced fiber composites is thus dependent on the validity of a homogeneous model of such materials. An experimental program to determine the validity of such a model for graphite/epoxy laminates is reviewed. Such laminates are found to have an apparent fracture toughness, from which it is inferred that a homogeneous material model is valid for the particular specimen geometry and composite laminates considered. Strain energy release rates are calculated from the experimentally determined fracture toughness of the various laminates. These strain energy release rates are found to lie in one of two groups, depending upon whether crack extension required fiber failure or matrix failure. The latter case is further investigated. It is concluded that matrix failure is governed by the tensile stress normal to the crack path.
Konish, H. J., Jr.↗
The relation between the polarity of the interplanetary magnetic field and the polar geomagnetic field
The relation between the azimuthal component of the interplanetary magnetic field and the polar cap geomagnetic field is discussed. The geomagnetic effects can be described as produced by an ionospheric current system encircling the magnetic pole. The sense of the current is clockwise during toward-sectors and reversed during away-sectors. The importance of this very direct solar-terrestrial relation is stressed. A recent magnetic sunspot cycle model is discussed as inferred from this relationship, the basic feature being that the sun reproduces the same sector pattern during every sunspot cycle.
Svalgaard, L.↗
External field effects on solidification - Macroscopic and microscopic models
Comparison of theoretical effects of magnetic and gravitational fields on melts of metals and semiconductors provides two kinds of insight into space processing materials problems. First, the task of comparing effects of the two fields yields mathematical models which may be used to estimate material properties and responses to processing under various field conditions. Then, by using these models, microgravity effects can be inferred from magnetic field data already obtained in laboratories on earth. Magnetohydrodynamics (static and oscillating fields) and the Free Volume Model are used to study macroscopic effects, and Sekerka's interface stability theory is used to show how external fields can affect microsegregation. It is demonstrated that the mathematical formalism is essentially the same for describing macroscopic effects of both microgravity and magnetic fields.
Miller, R. I.↗
Round jet in a cross flow - Influence of injection angle on vortex properties
A model is developed to infer the properties of a pair of diffuse contrarotating vortices for perpendicular jet injection into the cross flow. Attention is given to pressure distribution on the surface where the jet exhausts, especially in terms of V/STOL applications. A V/STOL wind tunnel experiment is described, whereby a round jet of air (10.16 cm dia) was discharged through a horizontal flat plate into the cross flow of the wind tunnel test section at angles of 45, 60, 75, and 90 deg. Results indicate that the effective vortex strength for both velocity ratios is highest near the jet orifice, and decreases as the vortices are swept downstream, at which point the effective vortex spacing increases. In addition, the core radius of the vortex exhibits properties similar to those of the effective vortex spacing
Krausche, D.↗
X-ray imaging - Supernova remnants
Consideration is given to imaging observations of supernova remnants (SNRs) obtained during the first year of the Einstein Observatory's operation. Inferences are drawn regarding models for stellar explosions, remnant evolution, neutron star formation and the interstellar medium. Because the X-ray emission traces the expanding shock boundary and dominates the radiative energy losses of an SNR over much of its lifetime, it can provide data on the possible collapsed remnants of the explosion, such as neutron stars and/or black holes. X-ray emission also allows a supernova shock to be used as a probe of interstellar medium structure. The imaging instrument aboard the Einstein satellite has been used to observe over 30 known Galactic remnants, and a similar number of objects in other galaxies, in the 0.15-4.5 keV band.
Helfand, D. J.↗
The photospheric magnetic field of the dM3.5e flare star AD Leonis
A high-resolution infrared spectrum of the dM3.5e flare star AD Leo, obtained with the Kitt Peak 4 m Fourier Transform Spectrometer, clearly shows the presence of strong magnetic fields. Five absorption lines in the 4400-4600 per cm region have been modeled, and it is inferred that 73 percent + or - 6 percent of the surface of AD Leo is covered by active regions outside of dark spots containing a mean field strength of 3800 + or - 260 G. If these active regions are brighter than the quiet photosphere, the surface filling factor will be somewhat smaller. Since simultaneous H-alpha observations exhibited no evidence of flares, the observations probably represent the quiescent magnetic flux level. The inferred field strength is consistent with equipartition of magnetic and thermal pressures in the photosphere and is similar to values derived using the scaling laws of Golub. The large observed filling factor is consistent with efficient dynamo generation of magnetic flux in this rapidly rotating star.
Saar, S. H.↗
Stabilization functions of unforced cumulus clouds - Their nature and components
The nature and the spatial distribution of cumuli within an unforced cumulus cloud field are investigated. The thermodynamic effects of convection are quantified as functions of changes of convective available potential energy (CAPE) induced by the convective overturning, and the time rate of change of CAPE is parameterized in terms of a kernel of influence or stabilization function. A three-dimensional cloud model is used to infer and quantify stabilization functions by performing single-cloud experiments. On the basis of the results obtained, a new hypothesis with respect to the spatial distribution of cumuli is postulated, which states that, under completely homogeneous external conditions and assuming a spatially random distribution of cloud-triggering mechanisms, the spatial distribution of cumuli in the resulting cloud field must be regular, as opposed to either random or clustered, because cumulus clouds tend to reduce the available energy for convection, thereby inhibiting further convection nearby.
Ramirez, Jorge A.↗
Optical, near, infrared and ultraviolet monitoring of the Seyfert 1 galaxy Markarian 335
Preliminary results of a multifrequency monitoring campaign for the bright, Seyfert 1 galactic nuclei Mkn335 are presented. Nearly uniform sampling at 3 day intervals is achieved quasi simultaneously at each wavelength band. Wavelength dependent variability is seen at the 20 to 30 percent level. Interpretation of variability in terms of geometrically thin, optically thick accretion disk models is discussed. The inferred blackhole masses and accretion rates are discussed. Possible correlation between continuum and emission line variations is discussed.
Shrader, Chris R.↗
Results of a study on polarization mix selection for the NSCAT scatterometer
The NASA scatterometer (NSCAT) is an instrument designed to measure the radar backscatter of the ocean's surface for estimating the near-surface wind velocity. A given resolution element is observed from several different azimuth angles. From these measurements the near-surface vector wind over the ocean may be inferred using a geophysical model function relating the normalized radar backscatter coefficient (sigma0) to the near-surface wind. The results of a study to select a polarization mix for NSCAT using an end-to-end simulation of the NSCAT scatterometer and ground processing of the sigma0 measurements into unambiguous wind fields using a median-filter-based ambiguity-removal algorithm are presented. The system simulation was used to compare the wind measurement accuracy and ambiguity removal skill over a set of realistic mesoscale wind fields for various polarization mixes. Considerations in the analysis and simulation are discussed, and a recommended polarization mix is given.
Long, David G.↗
Observations of the late-summer to fall sea ice transition with the 14.6 GHz Seasat scatterometer
Seasat 14.6-GHz scatterometer data are used to investigate variations in the large area-averaged backscatter signature during the late summer to fall transition. The temporal evolution in the backscatter coefficient indicates general changes in the surface properties of the sea ice which are attributed to the seasonal cycle. Notable events in this typical cycle are the meltpond maximum, rapid drainage or disappearance of the meltponds, fall cooling and freeze-up, and ultimately snowfall. All are manifested in the signature record, and a simple empirical scattering model is used to infer the mean snow and ice properties and their temporal variation.
Drinkwater, Mark R.↗
Laboratory measurements of the microwave and millimeter-wave opacity of gaseous sulfur dioxide (SO2) under simulated conditions for the Venus atmosphere
Laboratory measurements have been conducted of the opacity of gaseous SO2 in a CO2 atmosphere at 12.3 cm, 1.32 cm, and 0.32 cm, with a view to the effects of this gas on the mm-wave emission of the Venus atmosphere. Close agreement is noted between the results obtained and the absorptivity predicted from a Van Vleck-Weisskopf formalism at the two shortest wavelengths, but not at the longest. These results have been incorporated into a radiative transfer model in order to infer an abundance profile for gaseous SO2 in Venus' middle atmosphere, and are also used to ascertain the effects of a SO2/CO2 gaseous mixture on the mm-wavelength spectrum of Venus.
Fahd, Antoine K.↗