Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Edge computing”

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 1,315 records · Page 73

Community detection robustness of graph neural networks

Graph neural networks (GNNs) are increasingly widely used for community detection in attributed networks. They combine structural topology with node attributes through message passing and pooling. However, their robustness or lack thereof with respect to different perturbations and targeted attacks in conjunction with community detection tasks is not well understood. To shed light on latent mechanisms behind GNN sensitivity on community detection tasks, we conduct a systematic computational evaluation of six widely adopted GNN architectures graph convolutional network, graph attention network, graph sample and aggregate (GraphSAGE), differentiable pooling (DiffPool), minimum cut pooling (MinCUT), and deep modularity networks (DMoN). The analysis covers three perturbation categories: node attribute manipulations, edge topology distortions, and adversarial attacks. We use element-centric similarity as the evaluation metric on synthetic benchmarks and real-world citation networks. Our findings indicate that supervised GNNs tend to achieve higher baseline accuracy, while unsupervised methods, particularly DMoN, maintain stronger resilience under targeted and adversarial perturbations. Furthermore, robustness appears to be strongly influenced by community strength, with well-defined communities reducing performance loss. Across all models, node attribute perturbations associated with targeted edge deletions and shifts in attribute distributions tend to cause the largest degradation in community recovery. These findings highlight important trade-offs between accuracy and robustness in GNN-based community detection and offer insights into selecting architectures resilient to noise and adversarial attacks.

Goel, Jaidev [Virginia Polytechnic Inst. and State↗

Geometric remapping of particle distributions in the Discrete Element Model for Sea Ice (DEMSI v0.0)

Abstract. A new sea ice dynamical core, the Discrete Element Model for Sea Ice (DEMSI), is under development for use in coupled Earth system models. DEMSI is based on the discrete element method, which models collections of ice floes as interacting Lagrangian particles. In basin-scale sea ice simulations the Lagrangian motion results in significant convergence and ridging, which requires periodic remapping of sea ice variables from a deformed particle configuration back to an undeformed initial distribution. At the resolution required for Earth system models we cannot resolve individual sea ice floes, so we adopt the sub-grid-scale thickness distribution used in continuum sea ice models. This choice leads to a series of hierarchical tracers depending on ice fractional area or concentration that must be remapped consistently. The circular discrete elements employed in DEMSI help improve the computational efficiency at the cost of increased complexity in the effective element area definitions for sea ice cover that are required for the accurate enforcement of conservation. An additional challenge is the accurate remapping of element values along the ice edge, the location of which varies due to the Lagrangian motion of the particles. In this paper we describe a particle-to-particle remapping approach based on well-established geometric remapping ideas that enforces conservation, bounds preservation, and compatibility between associated tracer quantities, while also robustly managing remapping at the ice edge. One element of the remapping algorithm is a novel optimization-based flux correction that enforces concentration bounds in the case of nonuniform motion. We demonstrate the accuracy and utility of the algorithm in a series of numerical test cases.

58 GEOSCIENCES↗

Unsupervised, supervised and reinforced learning via spiking computation

The present invention relates to unsupervised, supervised and reinforced learning via spiking computation. The neural network comprises a plurality of neural modules. Each neural module comprises multiple digital neurons such that each neuron in a neural module has a corresponding neuron in another neural module. An interconnection network comprising a plurality of edges interconnects the plurality of neural modules. Each edge interconnects a first neural module to a second neural module, and each edge comprises a weighted synaptic connection between every neuron in the first neural module and a corresponding neuron in the second neural module.

Modha, Dharmendra S.↗

TH-A-19A-07: The Effect of Particle Tracking Step Size Limit On Monte Carlo- Computed LET Spectrum of Therapeutic Proton Beams

Purpose: To investigate the effect of charged particle tracking step size limit in the determination of the LET spectrum of therapeutic proton beams using Monte Carlo simulations. Methods: The LET spectra at different depths in a water phantom from a 79.7 MeV spot-scanning proton beam were calculated using Geant4. Five different tracking step limits 0.5 mm, 0.1 mm, 0.05 mm, 0.01 mm and 1 μm were adopted. The field size was set to 10×10 cm{sup 2} on the isocenter plane. A 40×40×6 cm{sup 3} water phantom was modelled as the irradiation target. The voxel size was set to 1×1×0.5 mm{sup 3} to obtain high resolution results. The LET spectra were scored ranging from 0.01 keV/μm to 10{sup 4}keV/μm in the logarithm scale. In addition, the proton energy spectra at different depths were also scored. Results: The LET spectra calculated using different step size limits were compared at four depths along the Bragg curve. At any depths, the spread of the LET spectra increases with the decrease of step size limit. In the dose buildup region (z = 1.9 cm) and in the region proximal to the Bragg peak (z = 3.95 cm), the frequency mean LET does not vary with decreasing step size limit. At Bragg peak (z = 4.75 cm) and in the distal edge (z = 4.85 cm), frequency mean LET decreases with decreasing step size limit. The energy spectrum at any specified depths does not vary with the step size limit. Conclusion: The calculated LET has a spectral distribution rather than a single value at any depths along the Bragg curve and the spread of the computed spectrum depends on the tracking step limit. Incorporating the LET spectrum distribution into the robust IMPT optimization plan may provide more accurate biological dose distribution than using the dose- or fluence-averaged LET. NIH Program Project Grant P01CA021239.

07 ISOTOPE AND RADIATION SOURCES↗

One-loop integrals from volumes of orthoschemes

Recently in arXiv:2012.05599 Rudenko presented a formula for the volume of hyperbolic orthoschemes in terms of alternating polylogarithms. We use this result to provide an explicit analytic result for the one-loop scalar n -gon Feynman integral in n dimensions, for even n , with massless or massive internal and external edges. Furthermore, we evaluate the general six-dimensional hexagon integral in terms of classical polylogarithms.

97 MATHEMATICS AND COMPUTING↗

Comparing Exciton-Polariton Couplings in Two-Dimensional MoS2 Cavities

Exciton-polaritons (EPs) are an emerging approach to achieve strong light-matter coupling and robust quantum phenomena, with application to areas such as photonic neuromorphic computing for instance reservoir computing. 2D transition metal dichalcogenides (TMDs) are promising candidates to host EPs due to their potential for strong coupling as evidenced by large Rabi splitting energies, which in the case of few-layer MoS2 has been demonstrated to reach 293 meV when coupled to the C exciton, placing it in the ultrastrong coupling regime, at room temperature. Here we examine how the Rabi splitting of CVD-grown monolayer MoS2 in a cavity can be optimized by tuning two variables: cavity geometry (thickness of silver and dielectric spacers) and the targeted exciton for coupling (A at about 1.9 eV, B at about 2.06 eV, or C at about 2.9 eV). With the cavity geometry, there is a tradeoff between optical field confinement and transmission. With the targeted exciton, there may be a tradeoff between Rabi splitting and lifetime. Transfer matrix simulations are used to inform how thick the cavity layers need to be, and at what angle measurements should be performed, to achieve Rabi splitting at a certain exciton energy. Experimentally, the Rabi splitting is measured by placing the cavity at a specified angle and measuring the optical reflectance and transmittance. By fitting the data, we find Rabi splittings of 124 meV, 194 meV, and 288 meV for the A, B, and C excitons, respectively, which places them in the strong, and on the edge of the ultrastrong, coupling regime.

exciton-polaritons↗

SPH with dynamical smoothing length adjustment based on the local flow kinematics

Due to the Lagrangian nature of Smoothed Particle Hydrodynamics (SPH), the adaptive resolution remains a challenging task. In this work, we first analyse the influence of the simulation parameters and the smoothing length on solution accuracy, in particular in high strain regions. Based on this analysis we develop a novel approach to dynamically adjust the kernel range for each SPH particle separately, accounting for the local flow kinematics. We use the Okubo–Weiss parameter that distinguishes the strain and vorticity dominated regions in the flow domain. The proposed development is relatively simple and implies only a moderate computational overhead. We validate the modified SPH algorithm for a selection of two-dimensional test cases: the Taylor–Green flow, the vortex spin-down, the lid-driven cavity and the dam-break flow against a sharp-edged obstacle. The simulation results show good agreement with the reference data and improvement of the long-term accuracy for unsteady flows. For the lid-driven cavity case, the proposed dynamical adjustment remedies the problem of tensile instability (particle clustering).

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Radiocesium interaction with clay minerals: Theory and simulation advances Post–Fukushima

Insights at the microscopic level of the process of radiocesium adsorption and interaction with clay mineral particles have improved substantially over the past several years, triggered by pressing social issues such as management of huge amounts of waste soil accumulated after the Fukushima Dai-ichi nuclear power plant accident. In particular, computer-based molecular modeling supported by advanced hardware and algorithms has proven to be a powerful approach. Its application can now generally encompass the full complexity of clay particle adsorption sites from basal surfaces to interlayers with inserted water molecules, to edges including fresh and weathered frayed ones. On the other hand, its methodological schemes are now varied from traditional force-field molecular dynamics on large-scale realizations composed of many thousands of atoms including water molecules to first-principles methods on smaller models in rather exacting fashion. In this article, we overview new understanding enabled by simulations across methodological variations, focusing on recent insights that connect with experimental observations, namely: 1) the energy scale for cesium adsorption on the basal surface, 2) progress in understanding the structure of clay edges, which is difficult to probe experimentally, 3) cesium adsorption properties at hydrated interlayer sites, 4) the importance of the size relationship between the ionic radius of cesium and the interlayer distance at frayed edge sites, 5) the migration of cesium into deep interlayer sites, and 6) the effects of nuclear decay of radiocesium. Key experimental observations that motivate these simulation advances are also summarized. Furthermore, some directions toward future solutions of waste soil management are discussed based on the obtained microscopic insights.

54 ENVIRONMENTAL SCIENCES↗

Adaptive two-regime method: Application to front propagation

The Adaptive Two-Regime Method (ATRM) is developed for hybrid (multiscale) stochastic simulation of reaction-diffusion problems. It efficiently couples detailed Brownian dynamics simulations with coarser lattice-based models. The ATRM is a generalization of the previously developed Two-Regime Method [Flegg et al., J. R. Soc., Interface 9, 859 (2012)] to multiscale problems which require a dynamic selection of regions where detailed Brownian dynamics simulation is used. Typical applications include a front propagation or spatio-temporal oscillations. In this paper, the ATRM is used for an in-depth study of front propagation in a stochastic reaction-diffusion system which has its mean-field model given in terms of the Fisher equation [R. Fisher, Ann. Eugen. 7, 355 (1937)]. It exhibits a travelling reaction front which is sensitive to stochastic fluctuations at the leading edge of the wavefront. Previous studies into stochastic effects on the Fisher wave propagation speed have focused on lattice-based models, but there has been limited progress using off-lattice (Brownian dynamics) models, which suffer due to their high computational cost, particularly at the high molecular numbers that are necessary to approach the Fisher mean-field model. By modelling only the wavefront itself with the off-lattice model, it is shown that the ATRM leads to the same Fisher wave results as purely off-lattice models, but at a fraction of the computational cost. The error analysis of the ATRM is also presented for a morphogen gradient model.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Quantitative resonant soft x-ray reflectivity of ultrathin anisotropic organic layers: Simulation and experiment of PTCDA on Au

Resonant soft X-ray reflectivity at the carbon K edge, with linearly polarized light, was used to derive quantitative information of film morphology, molecular arrangement, and electronic orbital anisotropies of an ultrathin 3,4,9,10-perylene tetracarboxylic dianhydride (PTCDA) film on Au(111). The experimental spectra were simulated by computing the propagation of the electromagnetic field in a trilayer system (vacuum/PTCDA/Au), where the organic film was treated as an anisotropic medium. Optical constants were derived from the calculated (through density functional theory) absorption cross sections of the single molecule along the three principal molecular axes. These were used to construct the dielectric tensor of the film, assuming the molecules to be lying flat with respect to the substrate and with a herringbone arrangement parallel to the substrate plane. Resonant soft X-ray reflectivity proved to be extremely sensitive to film thickness, down to the single molecular layer. The best agreement between simulation and experiment was found for a film of 1.6 nm, with flat laying configuration of the molecules. The high sensitivity to experimental geometries in terms of beam incidence and light polarization was also clarified through simulations. The optical anisotropies of the organic film were experimentally determined and through the comparison with calculations, it was possible to relate them to the orbital symmetry of the empty electronic states.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

STRUCTURE AND EVOLUTION OF CIRCUMBINARY DISKS AROUND SUPERMASSIVE BLACK HOLE BINARIES

We explore properties of circumbinary disks around supermassive black hole (SMBH) binaries in centers of galaxies by reformulating standard viscous disk evolution in terms of the viscous angular momentum flux F{sub J}. If the binary stops gas inflow and opens a cavity in the disk, then the inner disk evolves toward a constant-F{sub J} (rather than a constant M-dot ) state. We compute disk properties in different physical regimes relevant for SMBH binaries, focusing on the gas-assisted evolution of systems starting at separations 10{sup -4} - 10{sup -2} pc, and find the following. (1) Mass pileup at the inner disk edge caused by the tidal barrier accelerates binary inspiral. (2) Binaries can be forced to merge even by a disk with a mass below that of the secondary. (3) Torque on the binary is set non-locally, at radii far larger than the binary semi-major axis; its magnitude does not reflect disk properties in the vicinity of the binary. (4) Binary inspiral exhibits hysteresis-it depends on the past evolution of the disk. (5) The Eddington limit can be important for circumbinary disks even if they accrete at sub-Eddington rates, but only at late stages of the inspiral. (6) Gas overflow across the orbit of the secondary can be important for low secondary mass, high- M-dot systems, but mainly during the inspiral phase dominated by the gravitational wave emission. (7) Circumbinary disks emit more power and have harder spectra than constant M-dot disks; their spectra are very sensitive to the amount of overflow across the secondary orbit.

79 ASTRONOMY AND ASTROPHYSICS↗

Topological Entanglement and Clustering of Jain Hierarchy States

We obtain several clustering properties of the Jain states at filling (k/2k+1): they are a product of a Vandermonde determinant and a bosonic polynomial at filling (k/k+1) which vanishes when k+1 particles cluster together. We show that all Jain states satisfy a 'squeezing rule' which severely reduces the dimension of the Hilbert space necessary to generate them. We compute the topological entanglement spectrum of the Jain nu=(2/5) state and compare it to both the Coulomb ground state and the nonunitary Gaffnian state. All three states have a very similar 'low-energy' structure. However, the Jain state entanglement 'edge' state counting matches both the Coulomb counting as well as two decoupled U(1) free bosons, whereas the Gaffnian edge counting misses some of the edge states of the Coulomb spectrum.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Material separation in x-ray CT with energy resolved photon-counting detectors

Purpose: The objective of the study was to demonstrate that, in x-ray computed tomography (CT), more than two types of materials can be effectively separated with the use of an energy resolved photon-counting detector and classification methodology. Specifically, this applies to the case when contrast agents that contain K-absorption edges in the energy range of interest are present in the object. This separation is enabled via the use of recently developed energy resolved photon-counting detectors with multiple thresholds, which allow simultaneous measurements of the x-ray attenuation at multiple energies. Methods: To demonstrate this capability, we performed simulations and physical experiments using a six-threshold energy resolved photon-counting detector. We imaged mouse-sized cylindrical phantoms filled with several soft-tissue-like and bone-like materials and with iodine-based and gadolinium-based contrast agents. The linear attenuation coefficients were reconstructed for each material in each energy window and were visualized as scatter plots between pairs of energy windows. For comparison, a dual-kVp CT was also simulated using the same phantom materials. In this case, the linear attenuation coefficients at the lower kVp were plotted against those at the higher kVp. Results: In both the simulations and the physical experiments, the contrast agents were easily separable from other soft-tissue-like and bone-like materials, thanks to the availability of the attenuation coefficient measurements at more than two energies provided by the energy resolved photon-counting detector. In the simulations, the amount of separation was observed to be proportional to the concentration of the contrast agents; however, this was not observed in the physical experiments due to limitations of the real detector system. We used the angle between pairs of attenuation coefficient vectors in either the 5-D space (for non-contrast-agent materials using energy resolved photon-counting acquisition) or a 2-D space (for contrast agents using energy resolved photon-counting acquisition and all materials using dual-kVp acquisition) as a measure of the degree of separation. Compared to dual-kVp techniques, an energy resolved detector provided a larger separation and the ability to separate different target materials using measurements acquired in different energy window pairs with a single x-ray exposure. Conclusions: We concluded that x-ray CT with an energy resolved photon-counting detector with more than two energy windows allows the separation of more than two types of materials, e.g., soft-tissue-like, bone-like, and one or more materials with K-edges in the energy range of interest. Separating material types using energy resolved photon-counting detectors has a number of advantages over dual-kVp CT in terms of the degree of separation and the number of materials that can be separated simultaneously.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Artificial Intelligence for a Resilient and Flexible Power Grid

Recent widespread and extreme natural disasters as well as the drive towards clean sustainable energy sources necessitate a transformational operational and technological approaches to improve power grid resilience. Convergence of artificial intelligence (AI), distributed computing, and connectivity technologies enables these new approaches to deliver a sustainable electric grid. In particular, AI-based techniques can enable proactive decision-making by processing massive amounts of data to deliver intelligence at various levels of the system, edge devices to control room. In this chapter, we will explore the potential applications AI-based technology will unlock and challenges to address to achieve a sustainable, flexible power grid.

Omitaomu, Femi↗

Lifeline-based Global Load Balancing

On shared-memory systems, Cilk-style work-stealing has been used to effectively parallelize irregular task-graph based applications such as Unbalanced Tree Search (UTS). There are two main difficulties in extending this approach to distributed memory. In the shared memory approach, thieves (nodes without work) constantly attempt to asynchronously steal work from randomly chosen victims until they find work. In distributed memory, thieves cannot autonomously steal work from a victim without disrupting its execution. When work is sparse, this results in performance degradation. In essence, a direct extension of traditional work-stealing to distributed memory violates the work-first principle underlying work-stealing. Further, thieves spend useless CPU cycles attacking victims that have no work, resulting in system inefficiencies in multi-programmed contexts. Second, it is non-trivial to detect active distributed termination (detect that programs at all nodes are looking for work, hence there is no work). This problem is well-studied and requires careful design for good performance. Unfortunately, in most existing languages/frameworks, application developers are forced to implement their own distributed termination detection. In this paper, we develop a simple set of ideas that allow work-stealing to be efficiently extended to distributed memory. First, we introduce lifeline graphs: low-degree, low diameter, fully connected directed graphs. Such graphs can be constructed from k-dimensional hypercubes. When a node is unable to find work after w unsuccessful steals, it quiesces after informing the outgoing edges in its lifeline graph. Quiescent nodes do not disturb other nodes. A quiesced node is reactivated when work arrives from a lifeline and itself shares this work with those of its incoming lifelines that are activated. Termination occurs precisely when computation at all nodes has quiesced. In a language such as X10, such passive distributed termination can be detected automatically using the finish construct -- no application code is necessary.

99 GENERAL AND MISCELLANEOUS↗

Graph Theory and IC Component Design Analysis

Graph analysis in large integrated circuit (IC) designs is an essential tool for verifying design logic and timing via dynamic timing analysis (DTA). IC designs resemble graphs with each logic gate as a vertex and the conductive connections between gates as edges. Using DTA digital statistical correlations, graph condensation, and graph partitioning, it is possible to identify high-entropy component centers and paths within an IC design. Identification of high-entropy component centers (HECC) enables focused DTA, effectively lowering the computational complexity of DTA on large integrated circuit graphs. In this paper, a devised methodology termed IC layout subgraph component center identification (CCI) is used to identify described. CCI lowers DTA computationally complexity by condensing IC graphs into reduced subgraphs in which dominant logic functions are verified.

42 ENGINEERING↗

Computationally predicted energies and properties of defects in GaN

Recent developments in theoretical techniques have significantly improved the predictive power of density-functional-based calculations. In this article, we discuss how such advancements have enabled improved understanding of native point defects in GaN. We review the methodologies for the calculation of point defects, and discuss how techniques for overcoming the band-gap problem of density functional theory affect native defect calculations. In particular, we examine to what extent calculations performed with semilocal functionals (such as the generalized gradient approximation), combined with correction schemes, can produce accurate results. The properties of vacancy, interstitial, and antisite defects in GaN are described, as well as their interaction with common impurities. We also connect the first-principles results to experimental observations, and discuss how native defects and their complexes impact the performance of nitride devices. Overall, we find that lower-cost functionals, such as the generalized gradient approximation, combined with band-edge correction schemes can produce results that are qualitatively correct. However, important physics may be missed in some important cases, particularly for optical transitions and when carrier localization occurs.

36 MATERIALS SCIENCE↗

High-performance finite-difference time-domain simulations of C-Mod and ITER RF antennas

Finite-difference time-domain methods have, in recent years, developed powerful capabilities for modeling realistic ICRF behavior in fusion plasmas [1, 2, 3, 4]. When coupled with the power of modern high-performance computing platforms, such techniques allow the behavior of antenna near and far fields, and the flow of RF power, to be studied in realistic experimental scenarios at previously inaccessible levels of resolution. In this talk, we present results and 3D animations from high-performance FDTD simulations on the Titan Cray XK7 supercomputer, modeling both Alcator C-Mod’s field-aligned ICRF antenna and the ITER antenna module. Much of this work focuses on scans over edge density, and tailored edge density profiles, to study dispersion and the physics of slow wave excitation in the immediate vicinity of the antenna hardware and SOL. An understanding of the role of the lower-hybrid resonance in low-density scenarios is emerging, and possible implications of this for the NSTX launcher and power balance are also discussed. In addition, we discuss ongoing work centered on using these simulations to estimate sputtering and impurity production, as driven by the self-consistent sheath potentials at antenna surfaces.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗