Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “algorithm selection”

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 1,135 records · Page 63

A scalable algorithm for the optimization of neural network architectures

In this work, we propose a new scalable method to optimize the architecture of an artificial neural network. The proposed algorithm, called Greedy Search for Neural Network Architecture, aims to determine a neural network with minimal number of layers that is at least as performant as neural networks of the same structure identified by other hyperparameter search algorithms in terms of accuracy and computational cost. Numerical results performed on benchmark datasets show that, for these datasets, our method outperforms state-of-the-art hyperparameter optimization algorithms in terms of attainable predictive performance by the selected neural network architecture, and time-to-solution for the hyperparameter optimization to complete.

97 MATHEMATICS AND COMPUTING↗

Leptons: $\mathbf{e}$, $\boldsymbol{\mu}$, $\boldsymbol{\tau}$ + systematic uncertainties

An overview of the charged lepton performance in the ATLAS and CMS experiments at the LHC and in the CDF and D0 experiments at the Tevatron is given, in the specific context of physics analyses requiring the selection of leptons from top-quark decays. The trigger and reconstruction algorithms are presented, as well as the identification and isolation criteria, pointing out the specificities of top-quark analysis selections. The impact of the lepton-related experimental systematic uncertainties are discussed.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Conceptual design of a hybrid neutron-gamma detector for study of β-delayed neutrons at the RIB facility of RIKEN

Here, the conceptual design of the BRIKEN neutron detector at the radioactive ion beam factory (RIBF) of the RIKEN Nishina Center is reported. The BRIKEN setup is a complex system aimed at detecting heavy-ion implants, β particles, γ rays and β-delayed neutrons. The whole setup includes the Advanced Implantation Detection Array (AIDA), two HPGe Clover detectors and up to 166 3 He-filled counters embedded in a high-density polyethylene moderator. The design is quite complex due to the large number and different types of 3 He-tubes involved and the additional constraints introduced by the ancillary detectors for charged particles and γ rays. This article reports on a novel methodology developed for the conceptual design and optimisation of the 3 He-counter array, aiming for the best possible performance in terms of neutron detection. The algorithm is based on a geometric representation of two selected detector parameters of merit, namely, the average neutron detection efficiency and the efficiency flatness as a function of a reduced number of geometric variables. The response of the neutron detector is obtained from a systematic Monte Carlo simulation implemented in GEANT4. The robustness of the algorithm allowed us to design a versatile detection system, which operated in hybrid mode includes the full neutron counter and two clover detectors for high-precision gamma spectroscopy. In addition, the system can be reconfigured into a compact mode by removing the clover detectors and re-arranging the 3 He tubes in order to maximize the neutron detection performance. Both operation modes shows a rather flat and high average efficiency. In summary, we have designed a system which shows an average efficiency for hybrid mode ( 3 He tubes + clovers) of 68.6% and 64% for neutron energies up to 1 and 5 MeV, respectively. For compact mode (only 3 He tubes), the average efficiency is 75.7% and 71% for neutron energies up to 1 and 5 MeV, respectively. The performance of the BRIKEN detection system has been also quantified by means of Monte Carlo simulations with different neutron energy distributions.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Automated Algorithms for Band Gap Analysis from Optical Absorption Spectra

As high-throughput combinatorial experimental methods become more common, the technical challenge is shifting from producing materials to dealing with increasingly large datasets. One of the most important metrics to determine suitability of semiconductor materials for various applications is the band gap. This paper discusses automated algorithms for determining band gaps from optical absorption spectra. The algorithms are applied to a database of 34,313 optical absorption spectra, and selected results are compared to published theoretical and experimental band gap data from 16 materials sets. The best algorithm determines the band gaps with an accuracy of 0.37 eV for direct- and 0.93 eV for indirect band gaps for >20,000 spectra.

97 MATHEMATICS AND COMPUTING↗

SRBench++: Principled Benchmarking of Symbolic Regression With Domain-Expert Interpretation

Symbolic regression searches for analytic expressions that accurately describe studied phenomena. The main promise of this approach is that it may return an interpretable model that can be insightful to users, while maintaining high accuracy. The current standard for benchmarking these algorithms is SRBench, which evaluates methods on hundreds of datasets that are a mix of real-world and simulated processes spanning multiple domains. At present, the ability of SRBench to evaluate interpretability is limited to measuring the size of expressions on real-world data, and the exactness of model forms on synthetic data. In practice, model size is only one of many factors used by subject experts to determine how interpretable a model truly is. Furthermore, SRBench does not characterize algorithm performance on specific, challenging sub-tasks of regression such as feature selection and evasion of local minima. In this work, we propose and evaluate an approach to benchmarking SR algorithms that addresses these limitations of SRBench by 1) incorporating expert evaluations of interpretability on a domain-specific task, and 2) evaluating algorithms over distinct properties of data science tasks. We evaluate 12 modern symbolic regression algorithms on these benchmarks and present an in-depth analysis of the results, discuss current challenges of symbolic regression algorithms and highlight possible improvements for the benchmark itself.

97 MATHEMATICS AND COMPUTING↗

A variable high-order shock-capturing finite difference method with GP-WENO

Here, we present a new finite difference shock-capturing scheme for hyperbolic equations on static uniform grids. The method provides selectable high-order accuracy by employing a kernel-based Gaussian Process (GP) data prediction method which is an extension of the GP high-order method originally introduced in a finite volume framework by the same authors. The method interpolates Riemann states to high order, replacing the conventional polynomial interpolations with polynomial-free GP-based interpolations. For shocks and discontinuities, this GP interpolation scheme uses a nonlinear shock handling strategy similar to Weighted Essentially Non-Oscillatory (WENO), with a novelty consisting in the fact that nonlinear smoothness indicators are formulated in terms of the Gaussian likelihood of the local stencil data, replacing the conventional L 2 -type smoothness indicators of the original WENO method. We demonstrate that these GP-based smoothness indicators play a key role in the new algorithm, providing significant improvements in delivering high – and selectable – order accuracy in smooth flows, while successfully delivering non-oscillatory solution behavior in discontinuous flows.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Active learning for SNAP interatomic potentials via Bayesian predictive uncertainty

Bayesian inference with a simple Gaussian error model is used to efficiently compute prediction variances for energies, forces, and stresses in the linear SNAP interatomic potential. Here, the prediction variance is shown to have a strong correlation with the absolute error over approximately 24 orders of magnitude. Using this prediction variance, an active learning algorithm is constructed to iteratively train a potential by selecting the structures with the most uncertain properties from a pool of candidate structures. The relative importance of the energy, force, and stress errors in the objective function is shown to have a strong impact upon the trajectory of their respective net error metrics when running the active learning algorithm. Batched training of different batch sizes is also tested against singular structure updates, and it is found that batches can be used to significantly reduce the number of retraining steps required with only minor impact on the active learning trajectory.

97 MATHEMATICS AND COMPUTING↗

Fouling modeling and prediction approach for heat exchangers using deep learning

In this article, we develop a generalized and scalable statistical model for accurate prediction of fouling resistance using commonly measured parameters of industrial heat exchangers. This prediction model is based on deep learning where a scalable algorithmic architecture learns non-linear functional relationships between a set of target and predictor variables from large number of training samples. Here, the efficacy of this modeling approach is demonstrated for predicting fouling in an analytically modeled cross-flow heat exchanger, designed for waste heat recovery from flue-gas using room temperature water. The performance results of the trained models demonstrate that the mean absolute prediction errors are under 10 –4 KW –1 for flue-gas side, water side and overall fouling resistances. The coefficients of determination (R 2 ), which characterize the goodness of fit between the predictions and observed data, are over 99%. Even under varying levels of measurement noise in the inputs, we demonstrate that predictions over an ensemble of multiple neural networks achieves better accuracy and robustness to noise. We find that the proposed deep-learning fouling prediction framework learns to follow heat exchanger flow and heat transfer physics, which we confirm using locally interpretable model agnostic explanations around randomly selected operating points. Overall, we provide a robust algorithmic framework for fouling prediction that can be generalized and scaled to various types of industrial heat exchangers.

42 ENGINEERING↗

Identifying Optimal Measurement Subspace for the Ensemble Kalman Filter

To reduce the computational load of the ensemble Kalman filter while maintaining its efficacy, an optimization algorithm based on the generalized eigenvalue decomposition method is proposed for identifying the most informative measurement subspace. When the number of measurements is large, the proposed algorithm can be used to make an effective tradeoff between computational complexity and estimation accuracy. This algorithm also can be extended to other Kalman filters for measurement subspace selection.

99 GENERAL AND MISCELLANEOUS↗

Feature Selection Techniques for a Machine Learning Model to Detect Autonomic Dysreflexia

Feature selection plays a crucial role in the development of machine learning algorithms. Understanding the impact of the features on a model, and their physiological relevance can improve the performance. This is particularly helpful in the healthcare domain wherein disease states need to be identified with relatively small quantities of data. Autonomic Dysreflexia (AD) is one such example, wherein mismanagement of this neurological condition could lead to severe consequences for individuals with spinal cord injuries. We explore different methods of feature selection needed to improve the performance of a machine learning model in the detection of the onset of AD. We present different techniques used as well as the ideal metrics using a dataset of thirty-six features extracted from electrocardiograms, skin nerve activity, blood pressure and temperature. The best performing algorithm was a 5-layer neural network with five relevant features, which resulted in 93.4% accuracy in the detection of AD. The techniques in this paper can be applied to a myriad of healthcare datasets allowing forays into deeper exploration and improved machine learning model development. Through critical feature selection, it is possible to design better machine learning algorithms for detection of niche disease states using smaller datasets.

electrocardiography↗

Study of guided wave transmission through complex junction in sodium cooled reactor

Ultrasonic guided wave techniques are seen as suitable candidates for the inspection of welded structures within sodium cooled fast reactors (SFR), as the long range propagation of guided waves without amplitude attenuation can overcome the accessibility problem due to the liquid sodium. In the context of the development of the Advanced Sodium Test Reactor for Industrial Demonstration (ASTRID), the French Atomic Commission (CEA) investigates non-destructive testing techniques based on guided wave propagation. In this work, guided wave NDT methods are applied to control the integrity of welds located in a junction-type structure welded to the main vessel. The method presented in this paper is based on the analysis of scattering matrices peculiar to each expected defect, and takes advantage of the multi-modal and dispersive characteristics of guided wave generation. In a simulation study, an algorithm developed using the CIVA software is presented. It permits selecting appropriate incident modes to optimize detection and identification of expected flawed configurations. In the second part of this paper, experimental results corresponding to a first validation step of the simulation results are presented. The goal of the experiments is to estimate the effectiveness of the incident mode selection in plates. The results show good agreement between experience and simulation. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Field-Programmable Gate Array Computer in Structural Analysis: An Initial Exploration

This paper reports on an initial assessment of using a Field-Programmable Gate Array (FPGA) computational device as a new tool for solving structural mechanics problems. A FPGA is an assemblage of binary gates arranged in logical blocks that are interconnected via software in a manner dependent on the algorithm being implemented and can be reprogrammed thousands of times per second. In effect, this creates a computer specialized for the problem that automatically exploits all the potential for parallel computing intrinsic in an algorithm. This inherent parallelism is the most important feature of the FPGA computational environment. It is therefore important that if a problem offers a choice of different solution algorithms, an algorithm of a higher degree of inherent parallelism should be selected. It is found that in structural analysis, an 'analog computer' style of programming, which solves problems by direct simulation of the terms in the governing differential equations, yields a more favorable solution algorithm than current solution methods. This style of programming is facilitated by a 'drag-and-drop' graphic programming language that is supplied with the particular type of FPGA computer reported in this paper. Simple examples in structural dynamics and statics illustrate the solution approach used. The FPGA system also allows linear scalability in computing capability. As the problem grows, the number of FPGA chips can be increased with no loss of computing efficiency due to data flow or algorithmic latency that occurs when a single problem is distributed among many conventional processors that operate in parallel. This initial assessment finds the FPGA hardware and software to be in their infancy in regard to the user conveniences; however, they have enormous potential for shrinking the elapsed time of structural analysis solutions if programmed with algorithms that exhibit inherent parallelism and linear scalability. This potential warrants further development of FPGA-tailored algorithms for structural analysis.

Singleterry, Robert C., Jr.↗

Supercomputing on massively parallel bit-serial architectures

Research on the Goodyear Massively Parallel Processor (MPP) suggests that high-level parallel languages are practical and can be designed with powerful new semantics that allow algorithms to be efficiently mapped to the real machines. For the MPP these semantics include parallel/associative array selection for both dense and sparse matrices, variable precision arithmetic to trade accuracy for speed, micro-pipelined train broadcast, and conditional branching at the processing element (PE) control unit level. The preliminary design of a FORTRAN-like parallel language for the MPP has been completed and is being used to write programs to perform sparse matrix array selection, min/max search, matrix multiplication, Gaussian elimination on single bit arrays and other generic algorithms. A description is given of the MPP design. Features of the system and its operation are illustrated in the form of charts and diagrams.

Iobst, Ken↗

THE HUNT FOR EXOMOONS WITH KEPLER (HEK). I. DESCRIPTION OF A NEW OBSERVATIONAL PROJECT

Two decades ago, empirical evidence concerning the existence and frequency of planets around stars, other than our own, was absent. Since that time, the detection of extrasolar planets from Jupiter-sized to, most recently, Earth-sized worlds has blossomed and we are finally able to shed light on the plurality of Earth-like, habitable planets in the cosmos. Extrasolar moons may also be frequently habitable worlds, but their detection or even systematic pursuit remains lacking in the current literature. Here, we present a description of the first systematic search for extrasolar moons as part of a new observational project called 'The Hunt for Exomoons with Kepler' (HEK). The HEK project distills the entire list of known transiting planet candidates found by Kepler (2326 at the time of writing) down to the most promising candidates for hosting a moon. Selected targets are fitted using a multimodal nested sampling algorithm coupled with a planet-with-moon light curve modeling routine. By comparing the Bayesian evidence of a planet-only model to that of a planet-with-moon, the detection process is handled in a Bayesian framework. In the case of null detections, upper limits derived from posteriors marginalized over the entire prior volume will be provided to inform the frequency of large moons around viable planetary hosts, {eta} leftmoon. After discussing our methodologies for target selection, modeling, fitting, and vetting, we provide two example analyses.

79 ASTRONOMY AND ASTROPHYSICS↗

Design tool for multiprocessor scheduling and evaluation of iterative dataflow algorithms

A graph-theoretic design process and software tool is defined for selecting a multiprocessing scheduling solution for a class of computational problems. The problems of interest are those that can be described with a dataflow graph and are intended to be executed repetitively on a set of identical processors. Typical applications include signal processing and control law problems. Graph-search algorithms and analysis techniques are introduced and shown to effectively determine performance bounds, scheduling constraints, and resource requirements. The software tool applies the design process to a given problem and includes performance optimization through the inclusion of additional precedence constraints among the schedulable tasks.

Jones, Robert L., III↗

The Precipitation Rate Retrieval Algorithms for the GPM Dual-frequency Precipitation Radar

In this paper, precipitation rate retrieval algorithms for the Global Precipitation Measurement mission's Dual-frequency Precipitation Radar (DPR) are developed. The DPR consists of a Ku-band radar (KuPR; 13.6 GHz) and a Ka-band radar (KaPR; 35.5 GHz). For the KuPR, an algorithm similar to the Tropical Rainfall Measuring Mission's Precipitation Radar algorithm is developed, with the relation between precipitation rate R and massweighted mean diameter D m (R−D m relation) replacing the relation between the specific attenuation k and effective radar reflectivity factor Z e . The R−D m relation can also be applied to the KaPR and dual-frequency algorithms. In both the single-frequency and dual-frequency algorithms, the forward retrieval method is applied with an assumed adjustment factor for the R−D m relation (ε) and the results are evaluated to select the best value of ε. The advantages of the dual-frequency algorithm are the availability of the dual-frequency surface reference technique and the ZfKa method, which is a method to use the attenuation-corrected radar reflectivity factor Z f of KaPR, to select ε as well as the possibility to selectively use measurements from KuPR or KaPR. This paper also describes the derivation of the scattering table and the R−D m relation as well as the procedure for non-uniform beam filling correction in detail. The outputs are then statistically analyzed to demonstrate algorithm performance.

precipitation radar↗