Search NASA⌕ Search

SEARCH · Search NASA

Results for “Variational 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.

At least 505 records · Page 28

Modeling Radiation Detection Using a Mixed-Effects Statistical Model for Background Variation

In this paper, we consider radiation source localization in an urban environment as an inverse problem and use Bayesian parameter estimation techniques to infer the source location and intensity. Since parameter estimation generally requires a large number of model evaluations, we employ a simplified photon transport model, as in a 2017 paper by Stefanescu, et al., to reduce computational requirements. However, the model in the 2017 paper does not account for the physical reality that background radiation varies with position within an urban locale. We incorporate statistical mixed-effects modeling into the simplified transport equation from the Stefanescu, et al. paper to reflect individual variation among the background radiation measurements at multiple detector locations. With our new model, we use a Bayesian DRAM-within-Gibbs technique, which incorporates the Delayed Rejection Adaptive Metropolis (DRAM) algorithm, to estimate the mixed-effects parameters. We note that without adequate prior information, we were unable to simultaneously estimate the set of parameters corresponding to the source intensity, source location, and individual background radiation counts. By calibrating the background in the absence of a source, we obtained narrow prior distributions, which permitted us to jointly infer the source intensity and location and the background radiation distribution.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Global Statistics of Microphysical Properties of Cloud-Top Ice Crystals

Ice crystals in clouds are highly complex. Their sizes, macroscale shape (i.e., habit), mesoscale shape (i.e., aspect ratio of components) and microscale shape (i.e., surface roughness) determine optical properties and affect physical properties such as fall speeds, growth rates and aggregation efficiency. Our current understanding on the formation and evolution of ice crystals under various conditions can be considered poor. Commonly, ice crystal size and shape are related to ambient temperature and humidity, but global observational statistics on the variation of ice crystal size and particularly shape have not been available. Here we show results of a project aiming to infer ice crystal size, shape and scattering properties from a combination of MODIS measurements and POLDER-PARASOL multi-angle polarimetry. The shape retrieval procedure infers the mean aspect ratios of components of ice crystals and the mean microscale surface roughness levels, which are quantifiable parameters that mostly affect the scattering properties, in contrast to a habit. We present global statistics on the variation of ice effective radius, component aspect ratio, microscale surface roughness and scattering asymmetry parameter as a function of cloud top temperature, latitude, location, cloud type, season, etc. Generally, with increasing height, sizes decrease, roughness increases, asymmetry parameters decrease and aspect ratios increase towards unity. Some systematic differences are observed for clouds warmer and colder than the homogeneous freezing level. Uncertainties in the retrievals will be discussed. These statistics can be used as observational targets for modeling efforts and to better constrain other satellite remote sensing applications and their uncertainties.

Van Diedenhoven, Bastiaan↗

Lithospheric velocity structure of the Anatolian plateau-Caucasus-Caspian region

The Anatolian plateau-Caucasus-Caspian region is an area of complex lithospheric structure accompanied by large variations in seismic wave velocities. Despite the complexity of the region, little is known about the detailed lithospheric structure. Using data from 31 new, permanent broadband seismic stations along with results from a previous 29 temporary seismic stations and 3 existing global seismic stations in the region, a 3-D velocity model is developed using joint inversion of teleseismic receiver functions and surface waves. Both group and phase dispersion curves (Love and Rayleigh) were derived from regional and teleseismic events. Additional Rayleigh wave group dispersion curves were determined using ambient noise correlation. Receiver functions were calculated using P arrivals from 789 teleseismic (30°–90°) earthquakes. The stacked receiver functions and surface wave dispersion curves were jointly inverted to yield the absolute shear wave velocity to a depth of 100 km at each station. The depths of major discontinuities (sediment-basement, crust-mantle, and lithosphere-asthenosphere) were inferred from the velocity-depth profiles at the location of each station. Distinct spatial variations in crustal and upper mantle shear velocities were observed. The Kura basin showed slow (~2.7–2.9 km/s) upper crustal (0–11 km) velocities but elevated (~3.8–3.9 km/s) velocities in the lower crust. The Anatolian plateau varied from ~3.1–3.2 in the upper crust to ~3.5–3.7 in the lower crust, while velocities in the Arabian plate (south of the Bitlis suture) were slightly faster (upper crust between 3.3 and 3.4 km/s and lower crust between 3.8 and 3.9 km/s). The depth of the Moho, which was estimated from the shear velocity profiles, was 35 km in the Arabian plate and increased northward to 54 km at the southern edge of the Greater Caucasus. Moho depths in the Kura and at the edge of the Caspian showed more spatial variability but ranged between 35 and 45 km. Upper mantle velocities were slow under the Anatolian plateau but increased to the south under the Arabian plate and to the east (4.3–4.4 km/s) under the Kura basin and Greater Caucasus. The areas of slow mantle coincided with the locations of Holocene volcanoes. Differences between Rayleigh and Love dispersions at long wavelengths reveal a pronounced variation in anisotropy between the Anatolian plateau and the Kura basin.

58 GEOSCIENCES↗

Theory of an Automatic Seepage Meter and Ramifications for Applications

A new approach for measuring fluxes across surface water—groundwater interfaces was recently proposed. The Automatic Seepage Meter (ASM) is equipped with a precise water level sensor and digital memory that analyzes water level time series in a vertical tube inserted into a streambed. The ability to infer flux values with high temporal resolution relies on an accurate interpretation of water level dynamics inside the tube. Here, we reduce the three-dimensional hydrodynamic problem that describes the ASM water level in a variety of field conditions to a single ordinary differential equation. This novel general analytical solution for estimating ASM responses is more comprehensive and flexible than previous approaches and is applicable to the entire range of field conditions, including steady or transient stream stages, evaporation, rainfall, and noise. For example, our analysis determines the timing of the nonmonotonic ASM response to a monotonic linear stream stage variation and explains previously used empirical parabolic approximation for estimating fluxes. We present algorithms for simultaneous inference of vertical interface flux and hydraulic conductivity values together with an example code. We quantify how the accuracy of parameter estimation depends on test duration and noise amplitude and propose how our analysis can be used to optimize field test protocols. On this basis, changing the ASM geometry by increasing the radius and decreasing tube insertion depth may enable ASM field test protocols that estimate interface flux and hydraulic conductivity faster while maintaining desired accuracy. Potential applications of joint parameter estimation are suggested.

47 OTHER INSTRUMENTATION↗

A Hardware and Software Co-design Framework for Energy Efficient Neuromorphic Systems

Neuromorphic systems can be realized by a variety of algorithms and architectures. A common understanding is that spiking neuromorphic designs, which encode information into spatio-temporal spiking events, are both a biologically-accurate and efficient way of processing information. However, representing the information through timing relationships induces sophisticated circuit designs in traditional CMOS-based implementations. In recent years, high-capacity resistive memory (RRAM, aka, memristor) has demonstrated great potential in mimicking synaptic behaviors. Several RRAM-based spiking neuromorphic designs exist, most of which focus on rate coding schemes. These designs simplify circuit implementations of neuron models and explore challenges such as unsatisfactory speed, resolution, and performance. As an alternative, we will explore temporal coding spiking neuromorphic systems that encode information as the relative timing of neuron activations (spikes), which have been proven to be more adaptive and energy-efficient. Developing a neuromorphic system for spiking neural network (SNN) inference and online training, however, faces some major technical challenges: (1) It lacks circuit implementation support for temporal-coding SNN to achieve satisfying power efficiency and accuracy; (2) Although existing research works have investigated memristive synapse and neuron designs for spike-timing-dependent plasticity, the non-ideal conditions in implementation, such as device variations and signal degradation, degrade online learning accuracy of large scale systems; and (3) Non-optimized, inter-layer data traffic in SNNs, leads to unnecessary data communication costs. In this project, we plan to address these challenges by a hardware and software co-design framework that incorporates solutions at the circuit, architecture, and algorithm levels. At the circuit-level, we will elaborate on the in-situ SNN processing element designs for supporting both inference and online training modes. Variation-aware schemes will be studied to improve reliability. At the architecture level, we propose a pipelined, asynchronous architecture to retain the timing resolution of spikes. At the algorithm level, we will investigate an innovative SNN training algorithm for enabling activation sparsification and reducing unnecessary data communication costs. This neuromorphic system will provide an effective solution to real-life energy-constrained applications and significantly contribute to the exploration of next-generation high-performance computing systems under the DOE context.

97 MATHEMATICS AND COMPUTING↗

Period variations in pulsating X-ray sources. I - Accretion flow parameters and neutron star structure from timing observations

Torque fluctuations which can lead to variations in the periods of pulsating X-ray sources are examined. A description of torque variations in terms of noise processes is developed, and the resulting noise models are applied to observations of several pulsating X-ray sources. It is shown that fluctuations in accretion torque could account for the observed period variations and spindown episodes in Her X-1 and Cen X-3. The values of the torque noise strengths inferred from either a nonresonant response or, in the case of Her X-1, a Tkachenko-mode interpretation of the data are found to be consistent with those expected from processes at the magnetospheric boundary of an accreting neutron star. Ways to distinguish among the various interpretations of the period variations are considered. It is noted that fluctuating mass-flow rates may be responsible for other phenomena observed in compact X-ray sources, such as wobble with zero initial amplitude and binary period variations in close binary systems experiencing mass transfer.

Lamb, F. K.↗

Machine Learning Models for Network Traffic Classification in Programmable Logic

Network traffic classification via machine learning on network packet payloads has emerged as an active area of research for network security due to the high accuracy machine learning models have achieved in classifying payloads. For effective deployment as part of network security, these machine learning models must not only classify malicious packet payloads accurately, they must also identify anomalous payloads and perform inference at speeds generally faster than 10,000 packets per second to be effective. This work explores the in- ference speeds and accuracy of several neural network models implemented in programmable logic on various field programmable gate arrays (FPGA) including the Xilinx VC1902 and Xilinx Zynq Ultrascale+. This work also presents the design and performance of both an autoencoder and variational autoencoder programmed on the FPGA for identifying anomalous packet payloads. The performance benefits of the FPGA implementation for this type of packet payload inspection driven by machine learning are compared against graphics processing unit (GPU) inference implementations run on two state-of-the-art datacenter GPU devices, the NVIDIA V100 and A100. The model accuracy difference between the FPGA and GPU implementations was found to be 4% or less while the Xilinx VC1902 outperformed both the NVIDIA V100 and A100 for inference speeds on all the models explored except the variational autoencoder.

97 MATHEMATICS AND COMPUTING↗

Achieving zT > 1 in Inexpensive Zintl Phase Ca 9 Zn 4+ x Sb 9 by Phase Boundary Mapping

Complex multinary compounds (ternary, quaternary, and higher) offer countless opportunities for discovering new semiconductors for applications such as photovoltaics and thermoelectrics. However, controlling doping has been a major challenge in complex semiconductors as there are many possibilities for charged intrinsic defects (e.g., vacancies, interstitials, antisite defects) whose energy depends on competing impurity phases. Even with compounds with no apparent deviation from a stoichiometric nominal composition, such defects commonly lead to free carrier concentrations in excess of 1020 cm -3 . Nevertheless, by slightly altering the nominal composition, these defect concentrations can be tuned with small variation of the chemical potentials (composition) of each element. While the variation of chemical composition is undetectable, it is shown that the changes can be inferred by mapping (in nominal composition space) the boundaries where different competing impurity phases form. In the inexpensive Zintl compound Ca 9 Zn 4+x Sb 9 , the carrier concentrations can be finely tuned within three different three-phase regions by altering the nominal composition (x = 0.2–0.8), enabling the doubling of thermoelectric performance (zT). Because of the low thermal conductivity, the zT can reach as high as 1.1 at 875 K, which is one of the highest among the earth abundant p-type thermoelectrics with no ion conducting.

36 MATERIALS SCIENCE↗

Watersheds and stream networks viewed longitudinally: Example insights from novel spatial portrayals of watershed characteristics

Abstract Longitudinal depictions of watershed structure and characteristics, including topography, stream networks, wetlands, ground water levels, and land use, can provide watershed knowledge and understanding unavailable from standard plan view maps. Three case studies provide examples of knowledge gained by applying longitudinal views of stream networks, watershed hydrologic behavior, and land use distributions. Longitudinal views of mountain stream networks show extreme variability in the slope‐area relationships of low Strahler order streams, large discontinuities in drainage area (large parts of drainage area space are absent in networks), and large variations in network curvature. Longitudinal views of a groundwater‐dominated headwater watershed increase the inference available from limited groundwater observations and clearly reveal how groundwater connections affect the permanence of surface water features and the distribution of vadose zone storage in the landscape. Plotting land uses longitudinally illuminates and allows a quantitative analysis of how land uses are distributed relative to topographic position. Viewing watersheds and stream networks longitudinally can provide new insights into watershed forms and processes and motivate new questions and research.

60 APPLIED LIFE SCIENCES↗

The scalar chemical potential in cosmological collider physics

Non-analyticity in co-moving momenta within the non-Gaussian bispectrum is a distinctive sign of on-shell particle production during inflation, presenting a unique opportunity for the “direct detection” of particles with masses as large as the inflationary Hubble scale ( H ). However, the strength of such non-analyticity ordinarily drops exponentially by a Boltzmann-like factor as masses exceed H . In this paper, we study an exception provided by a dimension-5 derivative coupling of the inflaton to heavy-particle currents, applying it specifically to the case of two real scalars. The operator has a “chemical potential” form, which harnesses the large kinetic energy scale of the inflaton, \( {\overset{\cdot }{\phi}}_0^{1/2}\approx 60H \) , to act as an efficient source of scalar particle production. Derivative couplings of inflaton ensure radiative stability of the slow-roll potential, which in turn maintains (approximate) scale-invariance of the inflationary correlations. We show that a signal not suffering Boltzmann suppression can be obtained in the bispectrum with strength f NL ~ \( \mathcal{O} \) (0 . 01–10) for an extended range of scalar masses \( \lesssim {\overset{\cdot }{\phi}}_0^{1/2} \) , potentially as high as 10 15 GeV, within the sensitivity of upcoming LSS and more futuristic 21-cm experiments. The mechanism does not invoke any particular fine-tuning of parameters or breakdown of perturbation-theoretic control. The leading contribution appears at tree-level , which makes the calculation analytically tractable and removes the loop-suppression as compared to earlier chemical potential studies of non-zero spins. The steady particle production allows us to infer the effective mass of the heavy particles and the chemical potential from the variation in bispectrum oscillations as a function of co-moving momenta. Our analysis sets the stage for generalization to heavy bosons with non-zero spin.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

VpROM: a novel variational autoencoder-boosted reduced order model for the treatment of parametric dependencies in nonlinear systems

Reduced Order Models (ROMs) are of considerable importance in many areas of engineering in which computational time presents difficulties. Established approaches employ projection-based reduction, such as Proper Orthogonal Decomposition. The limitation of the linear nature of such operators is typically tackled via a library of local reduction subspaces, which requires the assembly of numerous local ROMs to address parametric dependencies. Our work attempts to define a more generalisable mapping between parametric inputs and reduced bases for the purpose of generative modeling. We propose the use of Variational Autoencoders (VAEs) in place of the typically utilised clustering or interpolation operations, for inferring the fundamental vectors, termed as modes, which approximate the manifold of the model response for any and each parametric input state. The derived ROM still relies on projection bases, built on the basis of full-order model simulations, thus retaining the imprinted physical connotation. However, it additionally exploits a matrix of coefficients that relates each local sample response and dynamics to the global phenomena across the parametric input domain. The VAE scheme is utilised for approximating these coefficients for any input state. This coupling leads to a high-precision low-order representation, which is particularly suited for problems where model dependencies or excitation traits cause the dynamic behavior to span multiple response regimes. Moreover, the probabilistic treatment of the VAE representation allows for uncertainty quantification on the reduction bases, which may then be propagated to the ROM response. The performance of the proposed approach is validated on an open-source simulation benchmark featuring hysteresis and multi-parametric dependencies, and on a large-scale wind turbine tower characterised by nonlinear material behavior and model uncertainty.

Conditional VAEs↗

ON THE FORM OF THE SPITZER LEAVITT LAW AND ITS DEPENDENCE ON METALLICITY

The form and metallicity dependence of Spitzer mid-infrared Cepheid relations are a source of debate. Consequently, Spitzer 3.6 and 4.5 {mu}m period-magnitude and period-color diagrams were re-examined via robust routines, thus providing an alternative interpretation to consider. The relations (nearly mean-magnitude) appear non-linear over an extensive baseline (0.45 < log P{sub 0} < 2.0), particularly the period-color trend, which to first order follows constant (3.6-4.5) color for shorter-period Cepheids and may transition into a bluer convex trough at longer periods. The period-magnitude functions can be described by polynomials (e.g., [3.6 {mu}m] = K{sub 0} - (3.071 {+-} 0.059)log P{sub 0} - (0.120 {+-} 0.032)log P{sub 0} {sup 2}), and Cepheid distances computed using 3.6 and 4.5 {mu}m relations agree with each other and the latter provides a first-order consistency check (CO sampled at 4.5 {mu}m does not seriously compromise those distances). The period-magnitude relations appear relatively insensitive to metallicity variations ([Fe/H] {approx} 0 to -0.75 |{gamma}| < 0.1 mag dex{sup -1}), a conclusion inferred partly from comparing galaxy distances established from those relations and NED-D (n > 700), yet a solid conclusion awaits comprehensive mid-infrared observations for metal-poor Cepheids in IC 1613 ([Fe/H] {approx} -1). The Cepheid-based distances were corrected for dust obscuration using a new ratio (i.e., A{sub 3.6}/E{sub B-V} = 0.18 {+-} 0.06) deduced from GLIMPSE (Spitzer) data.

79 ASTRONOMY AND ASTROPHYSICS↗

The architecture of the hierarchical triple star KOI 928 from eclipse timing variations seen in Kepler photometry

We present a hierarchical triple star system (KIC 9140402) where a low mass eclipsing binary orbits a more massive third star. The orbital period of the binary (4.98829 Days) is determined by the eclipse times seen in photometry from NASA's Kepler spacecraft. The periodically changing tidal field, due to the eccentric orbit of the binary about the tertiary, causes a change in the orbital period of the binary. The resulting eclipse timing variations provide insight into the dynamics and architecture of this system and allow the inference of the total mass of the binary (0.424±0.017M circle-dot ) and the orbital parameters of the binary about the central star.

79 ASTRONOMY AND ASTROPHYSICS↗

Supernova-driven Turbulent Metal Mixing in High-redshift Galactic Disks: Metallicity Fluctuations in the Interstellar Medium and its Imprints on Metal-poor Stars in the Milky Way

The extent to which turbulence mixes gas in the face of recurrent infusions of fresh metals by supernovae (SN) could help provide important constraints on the local star formation conditions. This includes predictions of the metallicity dispersion among metal-poor stars, which suggests that the interstellar medium was not very well mixed at these early times. The purpose of this Letter is to help isolate, via a series of numerical experiments, some of the key processes that regulate turbulent mixing of SN elements in galactic disks. We study the gas interactions in small simulated patches of a galaxy disk with the goal of resolving the small-scale mixing effects of metals at parsec scales, which enables us to measure the turbulent diffusion coefficient in various galaxy environments. By investigating the statistics of variations of α elements in these simulations, we are able to derive constraints not only on the allowed range of intrinsic yield variations in SN explosions but also on the star formation history of the Milky Way. We argue that the observed dispersion of [Mg/Fe] in metal-poor halo stars is compatible with the star-forming conditions expected in dwarf satellites or in an early low-star-forming Milky Way progenitor. In particular, metal variations in stars that have not been phase-mixed can be used to infer the star-forming conditions of disrupted dwarf satellites.

79 ASTRONOMY AND ASTROPHYSICS↗

The X-ray pulsars 4U1145-61 and 1E1145.1-6141

Observations of the X-ray sources near 4U1145-61 are presented. The presence of two periodicities at 291 and 297 seconds is confirmed. The location of the shorter period is isolated to within 3 arc minutes of 4U1145-61. The spectrum of 4U1145-61 is a power law out to at least 60 keV with a 500 eV EW iron line at 6.7 keV. The inferred spectrum of the other is absorbed. No evidence is found for a monotonic variation in either pulse period.

White, N. E.↗

Saturn's magnetic tail - Structure and dynamics

Voyager 1 magnetic field observations have provided evidence of a Saturnian magnetic tail. Tail current system distributions are inferred through comparison of the observations with a realistic magnetotail current system model. Temporal variations observed in the tail were probably produced by solar wind variations.

Behannon, K. W.↗

Electromagnetic inductive heating of the asteroids and moon as evidence bearing on the primordial solar wind

Thermal evolutionary models of the asteroids and the moon based on the hypothesis of electrical inductive heating by a dense primordial solar wind are compared with observation. Inferences (derived from spectroscopy) of asteroidal mineralogy when compared with electrical heating models, the variation of cooling rates recorded in iron meteorites, and the occurrence of the lunar magma ocean event lend support to the hypothesis of inductive heating. A high primordial solar rotation rate and magnetic field together with the loss due to plasma outflow of a substantial fraction of the initial solar mass would follow.

Herbert, F.↗