Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “VECTOR PROCESSING”

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 145 records · Page 8

PARA: A New Platform for the Rapid Assembly of gRNA Arrays for Multiplexed CRISPR Technologies

Multiplexed CRISPR technologies have great potential for pathway engineering and genome editing. However, their applications are constrained by complex, laborious and time-consuming cloning steps. In this research, we developed a novel method, PARA, which allows for the one-step assembly of multiple guide RNAs (gRNAs) into a CRISPR vector with up to 18 gRNAs. Here, we demonstrate that PARA is capable of the efficient assembly of transfer RNA/Csy4/ribozyme-based gRNA arrays. To aid in this process and to streamline vector construction, we developed a user-friendly PARAweb tool for designing PCR primers and component DNA parts and simulating assembled gRNA arrays and vector sequences.

59 BASIC BIOLOGICAL SCIENCES↗

Investigation of process history and underlying phenomena associated with the synthesis of plutonium oxides using Vector Quantizing Variational Autoencoder

Accurate, high throughput, and unbiased analysis of plutonium oxide particles is needed for analysis of the phenomenology associated with process parameters in their synthesis. Compared to qualitative and taxonomic descriptors, quantitative descriptors of particle morphology through scanning electron microscopy (SEM) have shown success in analyzing process parameters of uranium oxides. Among other candidates, a neural network called a Vector Quantizing Variational Autoencoder (VQ-VAE) has shown the ability to quantitatively describe particle morphology to attain >85% accuracy in identifying uranium oxide processing routes. We utilize a VQ-VAE to quantitatively describe plutonium dioxide (PuO 2 ) particles created in a designed experiment and investigate their phenomenology and prediction of their process parameters. PuO 2 was calcined from Pu(III) oxalates that were precipitated under varying synthetic conditions that related to concentrations, temperature, addition and digestion times, precipitant feed, and strike order; the surface morphology of the resulting PuO 2 powders were analyzed by SEM. A pipeline was developed to extract and quantify useful image representations for individual particles with the VQ-VAE, then further reduce the dimensionality of the feature space using a bottlenecking neural network fit to perform multiple classification tasks simultaneously. The reduced feature space could predict process parameters with greater than 80% accuracies for some parameters with a single particle. They also showed utility for grouping particles with similar surface morphology characteristics together. Both the clustering and classification results reveal valuable information regarding which chemical process parameters chiefly influence the PuO 2 particle morphologies: strike order and oxalic acid feedstock. Doing the same analysis with multiple particles was shown to improve the classification accuracy on each process parameter over the use of a single particle, with statistically significant results generally seen with as few as four particles in a sample.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

A New Perspective for the Calibration of Computational Predictor Models

This paper presents a framework for calibrating computational models using data from several and possibly dissimilar validation experiments. The offset between model predictions and observations, which might be caused by measurement noise, model-form uncertainty, and numerical error, drives the process by which uncertainty in the models parameters is characterized. The resulting description of uncertainty along with the computational model constitute a predictor model. Two types of predictor models are studied: Interval Predictor Models (IPMs) and Random Predictor Models (RPMs). IPMs use sets to characterize uncertainty, whereas RPMs use random vectors. The propagation of a set through a model makes the response an interval valued function of the state, whereas the propagation of a random vector yields a random process. Optimization-based strategies for calculating both types of predictor models are proposed. Whereas the formulations used to calculate IPMs target solutions leading to the interval value function of minimal spread containing all observations, those for RPMs seek to maximize the models’ ability to reproduce the distribution of observations. Regarding RPMs, we choose a structure for the random vector (i.e., the assignment of probability to points in the parameter space) solely dependent on the prediction error. As such, the probabilistic description of uncertainty is not a subjective assignment of belief, nor is it expected to asymptotically converge to a fixed value, but instead it is a description of the model’s ability to reproduce the experimental data. This framework enables evaluating the spread and distribution of the predicted response of target applications depending on the same parameters beyond the validation domain (i.e., roll-up and extrapolation).

42 ENGINEERING↗

A New Perspective for the Calibration of Computational Predictor Models.

This paper presents a framework for calibrating computational models using data from sev- eral and possibly dissimilar validation experiments. The offset between model predictions and observations, which might be caused by measurement noise, model-form uncertainty, and numerical error, drives the process by which uncertainty in the models parameters is characterized. The resulting description of uncertainty along with the computational model constitute a predictor model. Two types of predictor models are studied: Interval Predictor Models (IPMs) and Random Predictor Models (RPMs). IPMs use sets to characterize uncer- tainty, whereas RPMs use random vectors. The propagation of a set through a model makes the response an interval valued function of the state, whereas the propagation of a random vector yields a random process. Optimization-based strategies for calculating both types of predictor models are proposed. Whereas the formulations used to calculate IPMs target solutions leading to the interval value function of minimal spread containing all observations, those for RPMs seek to maximize the models' ability to reproduce the distribution of obser- vations. Regarding RPMs, we choose a structure for the random vector (i.e., the assignment of probability to points in the parameter space) solely dependent on the prediction error. As such, the probabilistic description of uncertainty is not a subjective assignment of belief, nor is it expected to asymptotically converge to a fixed value, but instead it is a description of the model's ability to reproduce the experimental data. This framework enables evaluating the spread and distribution of the predicted response of target applications depending on the same parameters beyond the validation domain (i.e., roll-up and extrapolation).

Crespo, Luis Guillermo↗

Gaussian processes with built-in dimensionality reduction: Applications to high-dimensional uncertainty propagation

Uncertainty quantification (UQ) tasks, such as model calibration, uncertainty propagation, and optimization under uncertainty, typically require several thousand evaluations of the underlying computer codes. To cope with the cost of simulations, one replaces the real response surface with a cheap surrogate based, e.g., on polynomial chaos expansions, neural networks, support vector machines, or Gaussian processes (GP). However, the number of simulations required to learn a generic multivariate response grows exponentially as the input dimension increases. This curse of dimensionality can only be addressed, if the response exhibits some special structure that can be discovered and exploited. A wide range of physical responses exhibit a special structure known as an active subspace (AS). An AS is a linear manifold of the stochastic space characterized by maximal response variation. The idea is that one should first identify this low dimensional manifold, project the high-dimensional input onto it, and then link the projection to the output. If the dimensionality of the AS is low enough, then learning the link function is a much easier problem than the original problem of learning a high-dimensional function. The classic approach to discovering the AS requires gradient information, a fact that severely limits its applicability. Furthermore, and partly because of its reliance to gradients, it is not able to handle noisy observations. The latter is an essential trait if one wants to be able to propagate uncertainty through stochastic simulators, e.g., through molecular dynamics codes. In this work, we develop a probabilistic version of AS which is gradient-free and robust to observational noise. Our approach relies on a novel Gaussian process regression with built-in dimensionality reduction. In particular, the AS is represented as an orthogonal projection matrix that serves as yet another covariance function hyper-parameter to be estimated from the data. To train the model, we design a two-step maximum likelihood optimization procedure that ensures the orthogonality of the projection matrix by exploiting recent results on the Stiefel manifold, i.e., the manifold of matrices with orthogonal columns. The additional benefit of our probabilistic formulation, is that it allows us to select the dimensionality of the AS via the Bayesian information criterion. We validate our approach by showing that it can discover the right AS in synthetic examples without gradient information using both noiseless and noisy observations. We demonstrate that our method is able to discover the same AS as the classical approach in a challenging one-hundred-dimensional problem involving an elliptic stochastic partial differential equation with random conductivity. Finally, we use our approach to study the effect of geometric and material uncertainties in the propagation of solitary waves in a one dimensional granular system.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

On polymorphism for discrete evolutionary dynamics driven either by selection or segregation distortion

We revisit some problems arising in the context of multiallelic discrete-time and deterministic evolutionary dynamics driven first by fitness differences and then by segregation distortion. In the model with fitness, we describe classes of fitness matrices exhibiting polymorphism. In the segregation case, still in search for conditions of polymorphism, we focus on a class of skew-symmetric matrices with a unique strictly positive kernel vector. Our main results reduce the study of these cases to the analysis of stochastic matrices.

97 MATHEMATICS AND COMPUTING↗

Method and system for determining precursors of health abnormalities from processing medical records

Medical reports are converted to document vectors in computing apparatus and sampled by applying a maximum variation sampling function including a fitness function to the document vectors to reduce a number of medical records being processed and to increase the diversity of the medical records being processed. Linguistic phrases are extracted from the medical records and converted to s-grams. A Haar wavelet function is applied to the s-grams over the preselected time interval; and the coefficient results of the Haar wavelet function are examined for patterns representing the likelihood of health abnormalities. This confirms certain s-grams as precursors of the health abnormality and a parameter can be calculated in relation to the occurrence of such a health abnormality.

96 KNOWLEDGE MANAGEMENT AND PRESERVATION↗

Frequency-domain computing using nonlinear acoustic-wave device on lithium niobate

Abstract Multiply-accumulation are crucial computing operations in signal processing, numerical simulations, and machine learning. In recent years, optical analog approaches have demonstrated higher computing performance and better power efficiency than their digital counterparts. However, analog computing chips usually need large areas and complex structures for parallel computing, as a single device element only executes one computing operation at a single time. Here, we demonstrate frequency-domain computing using the nonlinear acoustic-wave devices on lithium niobate, featuring a normalized external second-harmonic generation conversion efficiency of ~ 5.7 × 10-4 W-1. The second-order sum-frequency nonlinear process of lithium niobate enables multiplication of inputs encoded in the frequency domain. Compared to the analog schemes, our device features a notably simpler design, and nanofabrication requires only one lift-off. Using a single acoustic-wave device within an area of 0.03 mm2, we can simultaneously conduct over 130,000 multiply-accumulation operations. Our acoustic-wave device shows applications in real and complex vector convolutions and image processing. This demonstration sets the stage for experimental realizations into frequency-domain integrated nonlinear acoustic computing systems, potentially shaping future developments in acoustic neural networks and quantum computing.

chai, mingzhao (ORCID:0009000466226341)↗

Transition probability from matter-wave soliton to chaos

For a Bose-Einstein condensate loaded into a weak traveling optical superlattice, it is demonstrated that under a stochastic initial set and in a given parameter region, the solitonic chaos appears with a certain probability. Effects of the lattice depths and wave vectors on the chaos probability are investigated analytically and numerically and different chaotic regions associated with different chaos probabilities are found. The results suggest a method for weakening or strengthening chaos by modulating the moving superlattice.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Boson–Boson Interactions at the LHC

Vector boson scattering is a key production process to probe the electroweak symmetry breaking of the Standard Model and is one of the most important topics of the physics program for the HL-LHC since it involves both self-couplings of vector bosons and their coupling with the Higgs boson. If the Higgs mechanism is not the sole source of electroweak symmetry breaking, the scattering amplitude deviates from the Standard Model prediction at high scattering energy. Moreover, deviations may be detectable even if a New Physics scale is higher than the reach of direct searches. In this review, the most recent experimental measurements of the production cross sections of vector boson pairs in association with two jets in proton–proton collisions at $\sqrt{s}$ = 13 TeV at the LHC are reported, using data sets recorded by the ATLAS and CMS detectors. Applications to searches for New Physics, as well as prospects for measuring the electroweak vector boson scattering processes with larger data samples, are also summarized.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Field evaluation of remote wind sensing technologies: Shore-based and buoy mounted LIDAR systems

In developing a national energy strategy, the United States has a number of objectives, including increasing economic growth, improving environmental quality, and enhancing national energy security. Wind power contributes to these objectives through the deployment of clean, affordable and reliable domestic energy. To achieve U.S. wind generation objectives, the Wind and Water Power Program within the Department of Energy’s (DOE) Office of Energy Efficiency and Renewable Energy (EERE) instituted the U.S. Offshore Wind: Removing Market Barriers Program in FY 2011. Accurate and comprehensive information on offshore wind resource characteristics across a range of spatial and temporal scales is one market barrier that needs to be addressed through advanced research in remote sensing technologies. There is a pressing need for reliable offshore wind-speed measurements to assess the availability of the potential wind energy resource in terms of power production and to identify any frequently occurring spatial variability in the offshore wind resource that may impact the operational reliability and lifetime of wind turbines and their components and to provide a verification program to validate the “bankability” of the output of these alternative technologies for use by finance institutions for the financing of offshore wind farm construction. The application of emerging remote sensing technologies is viewed as a means to cost-effectively meet the data needs of the offshore wind industry. In particular, scanning and buoy mounted LIDAR have been proposed as a means to obtain accurate offshore wind data at multiple locations without the high cost and regulatory hurdles associated with the construction of offshore meteorological towers. However; before these remote sensing technologies can be accepted the validity of the measured data must be evaluated to ensure their accuracy. The proposed research will establish a unique coastal ocean test-bed in the Mid-Atlantic for the evaluation of LIDAR-based wind measurement systems to validate the accuracy of remotely measured wind data in marine applications. Specifically, the test-bed will be utilized to systematically evaluate the capability of emerging scanning LIDAR and buoy mounted vertically profiling LIDAR by: (1) Evaluating a fixed scanning LIDAR against land-based 50 and 60 meter high meteorological masts fitted with research quality cup-vane and/or sonic anemometers; (2) Evaluating a buoy mounted vertically profiling LIDAR fixed on land and floating in a sheltered bay against a co-located 60 meter high meteorological mast fitted with a research quality cup-vane and/or sonic anemometers and the fixed scanning LIDAR; and (3) Offshore field evaluation of both LIDAR platforms through a comparison of the fixed scanning LIDAR data and data obtained by the buoy mounted LIDAR located 10 miles offshore. The proposed research will systematically validate Light Detection and Ranging (LIDAR) based wind measurement systems and assess the temporal and spatial variability of the offshore wind resource in the Mid-Atlantic east of New Jersey. The goal of the proposed project is to address the technical and commercial challenges of the offshore wind energy industry by validating and assessing cost-effective, over ocean wind resource characterization technologies. The objective is to systematically evaluate the capability of both scanning and vertically profiling LIDARs to accurately measure 3D wind fields through comparison with fixed met masts and intercomparison among LIDAR platforms. Once validated, data collected by both buoy mounted vertically profiling LIDARs and shore-based, pulsed horizontally scanning LIDARs can be used to accurately assess offshore wind resources and to quantify the spatial and temporal variability in the offshore wind fields. One of the fundamental research questions to be addressed in phase 1 is the assessment of various measurement and data processing schemes to retrieve accurate wind vectors in the marine environment over large sampling ranges (10 to 12 km) and varying atmospheric aerosol levels. Atmospheric conditions and aerosol content within the coastal ocean region of the Mid-Atlantic seaboard of the US can vary significantly over short time periods in response to frontal passages and extratropical and tropical low pressure system passage offshore of the coast. Since aerosols provide the scattering medium for the determination of LIDAR Doppler shifts in the atmosphere the accuracy and range of LIDAR derived velocity measurements as a function of variation in aerosol content in the marine environment is a key research question to be addressed. In phase 1, it is desired to capture as much variation in atmospheric conditions and aerosol content as possible. To this end, collocated measurements of LIDAR and standard anemometer wind fields will be captured by the project PIs over all four seasons and during specific events (e.g., coastal low pressure system passage) in year 1. Additionally, since the meteorological masts are permanent structures, additional events can be captured over the three year duration of the field research project. All research instruments are owned by Fishermen’s Energy and made available to the PIs though a lease agreement as part of the DOE grant. Energy Fishermen’s Energy will be responsible for the operation and maintenance of the scanning LIDAR and met mast anemometers. On a daily basis, environmental data and systems performance indicators will be transmitted from each measurement station to the Fishermen’s project team consisting of both in-house personnel and equipment manufacturer engineers. Data sets include compiled LIDAR files as well as data sets from ancillary sensors. Diagnostic parameters to be monitored include standard deviations of measured values, battery levels and charging systems output, and the operational status. Once data have been confirmed as complete and reliable, files will be transferred to the Garrad Hassan (a subcontractor to Fishermen’s Energy) for incorporation in to the validation database, which is accessible to other scientific team members. Data collection times and durations will be determined by the PI and Co-PIs in consultation with instrument engineers to ensure the capture of data representative of the expected range of mid-Atlantic atmospheric conditions (e.g., temperature, moisture, coastal low pressure systems, tropical systems, rain, snow, fog). The collection and processing of the data is a function of site specific measurement requirements (Kelley et.al. 2007; Hannon et.al. 2008). To determine the optimal profiles of wind speed and direction from the LIDAR radial velocities as a function of azimuth angle, rigorous estimates of the bias and random error of each radial velocity estimate are required. Lockheed Martin Coherent Technologies, Inc., under contract with Fishermen’s Energy, will provide analyses of raw and processed data using various scan patterns to determine optimal performance settings for the pulsed scanning LIDAR. Once optimized, appropriate processing and analyses techniques will be evaluated by Garrad Hassan for use in validating the accuracy of the LIDAR wind field measurements against the standard anemometer measurements from the meteorological masts. The most attractive capability of the scanning LIDAR is the ability to provide high spatial resolution observations in a three-dimensional volume which provides superior statistical accuracy due to the large number of samples obtained. Each radial scan provides measurements in 100 range gates over a distance of 10 to 12 km at an update rate of 5 to 10 Hz and rotation of 2.5° per second. Each rotation at a fixed azimuth requires 2.4 minutes. Depending on the number of azimuths desired a complete scan can take up to 10 minutes or longer to complete. Once collected the radial velocities are processed to produce vector wind velocity estimates based on a set of data distributed in angle and range around points of interest, typically a standard grid within the radial wind map. To calculate wind vectors over a limited spatial area of interest for the comparison of data with other measurement platforms a localized least-squares approach has been applied by Hannon et.al. (2008) and a Variation Assimilation (VAR) processing technique has been applied by Chan and al Assimilation (VAR) processing technique has been applied by Chan and Shao (2006). Additionally, Kelley et.al. (2007) applied a “stare” technique that fixed a scanning LIDAR in both azimuth and elevation angles to measure over collocated sampling volumes of the LIDAR and a 3D sonic anemometer mounted to a fixed mast. Although the stare technique is limited to wind directions aligned with the sampling radial of the LIDAR, it does provide a direct comparison of sampling volumes. Each of the processing techniques described above (and possibly others) will be evaluated to determine the validity of the LIDAR derived wind fields in the marine environment. Numerical methods such as linear regression and comparison of probability density functions of wind fields measured by each instrument platform will be used to assess the processing techniques. Linear regression has the advantage of directly evaluating corresponding pairs of wind data measured by each instrument and can lend insight into deviations and bias between instruments as a function of wind speed. Assessment of the coherency between the probability density function of the wind measured by each instrument provides insight into processes that may not be accurately resolved by each instrument at specific frequencies. Once processed and assessed the most appropriate technique will be utilized to provide valid wind measurements from the pulsed scanning LIDAR. A detailed analysis of the measurement data from the LIDAR and the three meteorological towers and comparison of the coastal wind characteristics from the different systems will be performed by the CO_PI at NREL. This analysis will evaluate how the WT LIDAR performance and measurement of the wind characteristics vary with distance from the LIDAR and by atmospheric conditions; using the tower measurements at different distances (4.8, 9.6, and 19 km) from the LIDAR as a reference. The comparative analysis will include, to the extent possible, evaluation of parameters such as wind speed and direction distributions, wind shear, turbulence intensity and their variations by atmospheric conditions, month or season, and time of day.

17 WIND ENERGY↗

Vectorization of CMSSW offline software

The CMS experiment has been utilizing vectorization, or SIMD, in parts of its data processing applications for over a decade. On x86 platforms the vectorization level is still SSE3. In the past attempts to use wider vector instruction sets such as AVX or AVX-512 have, in practice, not resulted in improvements in the overall event processing throughput, because the CPUs scale down their frequency when processing AVX instructions. In addition, a notable part of the global pool of CMS’ resources has been old systems either not supporting AVX, or where the CPU frequency downscaling impacts all cores of the CPU. CMS has nevertheless continued to vectorize more of its application code, and in this work we review profiling methods we have found effective to find out pieces of code that would benefit from vectorization, and techniques to transform those codes such that the GCC compiler is able to auto-vectorize those codes. The build system used for CMSSW, Scram, has also been enha nced to be able to build code for multiple CPU microarchitectures such that the shared libraries of desired microarchitecture level can be loaded based on the CPU of the system. This multi-microarchitecture setup is invisible to the workflow management system, which makes its deployment straightforward. We describe in detail how this multi-microarchitecture build is set up, and measure the impact of using wider vector units than SSE3 on the event processing throughput of CMS applications such as simulation and reconstruction on recent x86 CPUs.

Gartung, Patrick↗

Multi-objective optimization of peel and shear strengths in ultrasonic metal welding using machine learning-based response surface methodology

Ultrasonic metal welding (UMW) is a solid-state joining technique with varied industrial applications. Despite of its numerous advantages, UMW has a relative narrow operating window and is sensitive to variations in process conditions. As such, it is imperative to quantitatively characterize the influence of welding parameters on the resulting joint quality. The quantification model can be subsequently used to optimize the parameters. Conventional response surface methodology (RSM) usually employs linear or polynomial models, which may not be able to capture the intricate, nonlinear input-output relationships in UMW. Furthermore, some UMW applications call for simultaneous optimization of multiple quality indices such as peel strength, shear strength, electrical conductivity, and thermal conductivity. To address these challenges, this paper develops a machine learning (ML)-based RSM to model the input-output relationships in UMW and jointly optimize two quality indices, namely, peel and shear strengths. The performance of various ML methods including spline regression, Gaussian process regression (GPR), support vector regression (SVR), and conventional polynomial regression models with different orders is compared. A case study using experimental data shows that GPR with radial basis function (RBF) kernel and SVR with RBF kernel achieve the best prediction accuracy. The obtained response surface models are then used to optimize a compound joint strength indicator that is defined as the average of normalized shear and peel strengths. In addition, the case study reveals different patterns in the response surfaces of shear and peel strengths, which has not been systematically studied in the literature. While developed for the UMW application, the method can be extended to other manufacturing processes.

42 ENGINEERING↗

Reprogramming of human fibroblasts to pluripotent stem cells using mRNA of four transcription factors

Reprogramming of differentiated cells into induced pluripotent cells (iPS) was accomplished in 2006 by expressing four, or less, embryonic stem cell (ESC)-specific transcription factors. Due to the possible danger of DNA damage and the potential tumorigenicity associated with such DNA damage, attempts were made to minimize DNA integration by the vectors involved in this process without complete success. Here we present a method of using RNA transfection as a tool for reprogramming human fibroblasts to iPS. We used RNA synthesized in vitro from cDNA of the same reprogramming four transcription factors. After transfection of the RNA, we show intracellular expression and nuclear localization of the respective proteins in at least 70% of the cells. We used five consecutive transfections to support continuous protein expression resulting in the formation of iPS colonies that express alkaline phosphatase and several ESC markers and that can be expanded. This method completely avoids DNA integration and may be developed to replace the use of DNA vectors in the formation of iPS.

60 APPLIED LIFE SCIENCES↗

Iterated Gauss-Seidel GMRES

The GMRES algorithm of Saad and Schultz [SIAM J. Sci. Stat. Comput., 7 (1986), pp. 856-869] is an iterative method for approximately solving linear systems Ax = b, with initial guess x0 and residual r0 = b Ax0. The algorithm employs the Arnoldi process to generate the Krylov basis vectors (the columns of Vk ). It is well known that this process can be viewed as a QR factorization of the matrix Bk = [r0, AVk] at each iteration. Despite an O (..epsilon..)..kappa.. (Bk ) loss of orthogonality, for unit roundoff ..epsilon..and condition number ..kappa.. , the modified Gram-Schmidt formulation was shown to be backward stable in the seminal paper by Paige et al. [SIAM J. Matrix Anal.Appl., 28 (2006), pp. 264-284]. We present an iterated Gauss-Seidel formulation of the GMRES algorithm (IGS-GMRES) based on the ideas of Ruhe [Linear Algebra Appl., 52 (1983), pp. 591-601] and Swirydowicz et al. [Numer. Linear Algebra Appl., 28 (2020), pp. 1-20]. IGS-GMRES maintains orthogonality to the level O (..epsilon..)..kappa.. (Bk ) or O (..epsilon..), depending on the choice of one or two iterations; for two Gauss-Seidel iterations, the computed Krylov basis vectors remain orthogonal to working accuracy and the smallest singular value of Vk remains close to one. The resulting GMRES method is thus backward stable. We show that IGS-GMRES can be implemented with only a single synchronization point per iteration, making it relevant to large-scale parallel computing environments. We also demonstrate that, unlike MGS-GMRES, in IGS-GMRES the relative Arnoldi residual corresponding to the computed approximate solution no longer stagnates above machine precision even for highly nonnormal systems.

Arnoldi-QR↗

Strong decays of radially excited mesons in a chiral approach

We study radial excitations of pseudoscalar and vector qq mesons within a chiral approach. We derive a general form for a chiral Lagrangian describing processes involving excited pseudoscalar and vector mesons. The parameters of the chiral Lagrangian are fitted using data and previous calculations in the framework of the {sup 3}P{sub 0} model. Finite-width effects are examined and predictions for mesons previously not discussed are given. Available experimental data is analyzed whenever possible. Possible hints for exotic mesons and open interpretation issues are discussed.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Applying Machine‐Learning Methods to Laser Acceleration of Protons: Lessons Learned From Synthetic Data

ABSTRACT In this study, we consider three different machine‐learning methods—a three‐hidden‐layer neural network, support vector regression, and Gaussian process regression—and compare how well they can learn from a synthetic data set for proton acceleration in the Target Normal Sheath Acceleration regime. The synthetic data set was generated from a previously published theoretical model by Fuchs et al. 2005 that we modified. Once trained, these machine‐learning methods can assist with efforts to maximize the peak proton energy, or with the more general problem of configuring the laser system to produce a proton energy spectrum with desired characteristics. In our study, we focus on both the accuracy of the machine‐learning methods and the performance on one GPU including memory consumption. Although it is arguably the least sophisticated machine‐learning model we considered, support vector regression performed very well in our tests.

Desai, Ronak↗

A communication-avoiding, hybrid-parallel, rank-revealing orthogonalization method.

Orthogonalization consumes much of the run time of many iterative methods for solving sparse linear systems and eigenvalue problems. Commonly used algorithms, such as variants of Gram-Schmidt or Householder QR, have performance dominated by communication. Here, 'communication' includes both data movement between the CPU and memory, and messages between processors in parallel. Our Tall Skinny QR (TSQR) family of algorithms requires asymptotically fewer messages between processors and data movement between CPU and memory than typical orthogonalization methods, yet achieves the same accuracy as Householder QR factorization. Furthermore, in block orthogonalizations, TSQR is faster and more accurate than existing approaches for orthogonalizing the vectors within each block ('normalization'). TSQR's rank-revealing capability also makes it useful for detecting deflation in block iterative methods, for which existing approaches sacrifice performance, accuracy, or both. We have implemented a version of TSQR that exploits both distributed-memory and shared-memory parallelism, and supports real and complex arithmetic. Our implementation is optimized for the case of orthogonalizing a small number (5-20) of very long vectors. The shared-memory parallel component uses Intel's Threading Building Blocks, though its modular design supports other shared-memory programming models as well, including computation on the GPU. Our implementation achieves speedups of 2 times or more over competing orthogonalizations. It is available now in the development branch of the Trilinos software package, and will be included in the 10.8 release.

99 GENERAL AND MISCELLANEOUS↗