Search NASA⌕ Search

SEARCH · Search NASA

Results for “randomized algorithm”

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 865 records · Page 48

Monte Carlo simulator of realistic x-ray beam for diagnostic applications

Purpose: Monte Carlo simulation is a very useful tool for radiotherapy and diagnostic radiology. Yet even with the latest PCs, simulation of photon spectra emitted by an x-ray tube is a time-consuming task, potentially reducing the possibility to obtain relevant data such as dose evaluations, simulation of geometric settings, or monitor detector efficiency. This study developed and validated a method to generate random numbers for realistic beams in terms of photon spectrum and intensity to simulate x-ray tubes via Monte Carlo algorithms. Methods: Starting from literature data, the most common semiempirical models of bremsstrahlung are analyzed and implemented, adjusting their formulation to describe a large irradiation area (i.e., large field of view) and to take account of the heel effect as in common practice during patient examinations. Results: Simulation results show that Birch and Marshall's model is the fastest and most accurate for the aims of this work. Correction of the geometric size of the beam and validation of the intensity variation (heel effect) yielded excellent results with differences between experimental and simulated data of less than 6%. Conclusions: The results of validation and execution time showed that the tube simulator calculates the x-ray photons quickly and efficiently and is perfectly capable of considering all the phenomena occurring in a real beam (total filtration, focal spot size, and heel effect), so it can be used in a wide range of applications such as industry, medical physics, or quality assurance.

60 APPLIED LIFE SCIENCES↗

Computationally Efficient Decompositions of Oblique Projection Matrices

Oblique projection matrices arise in problems in weighted least squares, signal processing, and optimization. While these matrices can be potentially very large, their low-rank structure can be exploited for efficient computation. Here, we propose fast and scalable algorithms for computing their eigendecomposition and singular value decomposition (SVD). Numerical experiments that compare our proposed approaches to existing methods, including randomized SVD, are presented. In addition, we test their accuracy on linear systems from equality constrained optimization problems.

97 MATHEMATICS AND COMPUTING↗

A robust prognostic signature for hormone-positive node-negative breast cancer

Background: Systemic chemotherapy in the adjuvant setting can cure breast cancer in some patients that would otherwise recur with incurable, metastatic disease. However, since only a fraction of patients would have recurrence after surgery alone, the challenge is to stratify high-risk patients (who stand to benefit from systemic chemotherapy) from low-risk patients (who can safely be spared treatment related toxicities and costs). Methods: We focus here on risk stratification in node-negative, ER-positive, HER2-negative breast cancer. We use a large database of publicly available microarray datasets to build a random forests classifier and develop a robust multi-gene mRNA transcription-based predictor of relapse free survival at 10 years, which we call the Random Forests Relapse Score (RFRS). Performance was assessed by internal cross-validation, multiple independent data sets, and comparison to existing algorithms using receiver-operating characteristic and Kaplan-Meier survival analysis. Internal redundancy of features was determined using k-means clustering to define optimal signatures with smaller numbers of primary genes, each with multiple alternates. Results: Internal OOB cross-validation for the initial (full-gene-set) model on training data reported an ROC AUC of 0.704, which was comparable to or better than those reported previously or obtained by applying existing methods to our dataset. Three risk groups with probability cutoffs for low, intermediate, and high-risk were defined. Survival analysis determined a highly significant difference in relapse rate between these risk groups. Validation of the models against independent test datasets showed highly similar results. Smaller 17-gene and 8-gene optimized models were also developed with minimal reduction in performance. Furthermore, the signature was shown to be almost equally effective on both hormone-treated and untreated patients. Conclusions: RFRS allows flexibility in both the number and identity of genes utilized from thousands to as few as 17 or eight genes, each with multiple alternatives. The RFRS reports a probability score strongly correlated with risk of relapse. This score could therefore be used to assign systemic chemotherapy specifically to those high-risk patients most likely to benefit from further treatment.

59 BASIC BIOLOGICAL SCIENCES↗

Using GANs with adaptive training data to search for new molecules

The process of drug discovery involves a search over the space of all possible chemical compounds. Generative Adversarial Networks (GANs) provide a valuable tool towards exploring chemical space and optimizing known compounds for a desired functionality. Standard approaches to training GANs, however, can result in mode collapse, in which the generator primarily produces samples closely related to a small subset of the training data. In contrast, the search for novel compounds necessitates exploration beyond the original data. In this work, we present an approach to training GANs that promotes incremental exploration and limits the impacts of mode collapse using concepts from Genetic Algorithms. In our approach, valid samples from the generator are used to replace samples from the training data. We consider both random and guided selection along with recombination during replacement. By tracking the number of novel compounds produced during training, we show that updates to the training data drastically outperform the traditional approach, increasing potential applications for GANs in drug discovery.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Capacity Optimization of a Centralized Charging Station in Joint Operation with a Wind Farm

In the context of large-scale wind power integration and rapid development of electric vehicles (EVs), a joint operation pattern was proposed to use a centralized charging station (CCS) to address high uncertainties incurred by wind power integration. Additionally, this would directly remove the significant indeterminacy of wind power. Because the CCS is adjacent to a wind power gathering station, it could work jointly with wind farms to operate in power system as an independent enterprise. By combining the actual operational characteristics of the wind farm and the CCS, a multidimensional operating index evaluation system was created for the joint system. Based on the wind farm’s known capacity, a multi-target capacity planning model was established for CCS to maximize the probability of realizing diverse indices of the system based on dependent-chance goal programming. In this model, planning and dispatching are combined to improve the feasibility of results in the operational stage. This approach takes the effects of randomness into account for wind power and battery swapping. The model was solved through combining Monte Carlo simulation and a genetic algorithm (GA) based on segmented encoding. Finally, as shown in the simulation results, the proposed model could comprehensively include factors such as initial investment, wind power price, battery life, etc., to optimize CCS capacity and to ultimately improve the operating indices of the joint system.

17 WIND ENERGY↗

Sampling Conditions for Conforming Voronoi Meshing by the VoroCrust Algorithm

We study the problem of decomposing a volume with a smooth boundary into a collection of Voronoi cells. Unlike the dual problem of conforming Delaunay meshing, a principled solution to this problem for generic smooth surfaces remained elusive. VoroCrust leverages ideas from weighted α-shapes and the power crust algorithm to produce unweighted Voronoi cells conforming to the surface, yielding the first provably-correct algorithm for this problem. Given a κ-sparse ε-sample, we work with the balls of radius δ times the local feature size centered at each sample. The corners of the union of these balls on both sides of the surface are the Voronoi sites and the interface of their cells is a watertight surface reconstruction embedded in the dual shape of the union of balls. With the surface protected, the enclosed volume can be further decomposed by generating more sites inside it. Compared to clipping-based algorithms, VoroCrust cells are full Voronoi cells, with convexity and fatness guarantees. Compared to the power crust algorithm, VoroCrust cells are not filtered, are unweighted, and offer greater flexibility in meshing the enclosed volume by either structured or randomly genenerated samples.

97 MATHEMATICS AND COMPUTING↗

Hyperspectral Image Classification Using Graph Clustering Methods

Hyperspectral imagery is a challenging modality due to the dimension of the pixels which can range from hundreds to over a thousand frequencies depending on the sensor. Most methods in the literature reduce the dimension of the data using a method such as principal component analysis, however this procedure can lose information. More recently methods have been developed to address classification of large datasets in high dimensions. This paper presents two classes of graph-based classification methods for hyperspectral imagery. Using the full dimensionality of the data, we consider a similarity graph based on pairwise comparisons of pixels. The graph is segmented using a pseudospectral algorithm for graph clustering that requires information about the eigenfunctions of the graph Laplacian but does not require computation of the full graph. We develop a parallel version of the Nyström extension method to randomly sample the graph to construct a low rank approximation of the graph Laplacian. With at most a few hundred eigenfunctions, we can implement the clustering method designed to solve a variational problem for a graph-cut-based semi-supervised or unsupervised classification problem. We implement OpenMP directive-based parallelism in our algorithms and show performance improvement and strong, almost ideal, scaling behavior. The method can handle very large datasets including a video sequence with over a million pixels, and the problem of segmenting a data set into a pre-determined number of classes.

47 OTHER INSTRUMENTATION↗

Worst configurations (instantons) for compressed sensing over reals: a channel coding approach

We consider Linear Programming (LP) solution of a Compressed Sensing (CS) problem over reals, also known as the Basis Pursuit (BasP) algorithm. The BasP allows interpretation as a channel-coding problem, and it guarantees the error-free reconstruction over reals for properly chosen measurement matrix and sufficiently sparse error vectors. In this manuscript, we examine how the BasP performs on a given measurement matrix and develop a technique to discover sparse vectors for which the BasP fails. The resulting algorithm is a generalization of our previous results on finding the most probable error-patterns, so called instantons, degrading performance of a finite size Low-Density Parity-Check (LDPC) code in the error-floor regime. The BasP fails when its output is different from the actual error-pattern. We design CS-Instanton Search Algorithm (ISA) generating a sparse vector, called CS-instanton, such that the BasP fails on the instanton, while its action on any modification of the CS-instanton decreasing a properly defined norm is successful. We also prove that, given a sufficiently dense random input for the error-vector, the CS-ISA converges to an instanton in a small finite number of steps. Performance of the CS-ISA is tested on example of a randomly generated 512 * 120 matrix, that outputs the shortest instanton (error vector) pattern of length 11.

99 GENERAL AND MISCELLANEOUS↗

MONACO: accurate biological network alignment through optimal neighborhood matching between focal nodes

Motivation: Alignment of protein–protein interaction networks can be used for the unsupervised prediction of functional modules, such as protein complexes and signaling pathways, that are conserved across different species. To date, various algorithms have been proposed for biological network alignment, many of which attempt to incorporate topological similarity between the networks into the alignment process with the goal of constructing accurate and biologically meaningful alignments. Especially, random walk models have been shown to be effective for quantifying the global topological relatedness between nodes that belong to different networks by diffusing node-level similarity along the interaction edges. However, these schemes are not ideal for capturing the local topological similarity between nodes. Results: Here, we propose MONACO, a novel and versatile network alignment algorithm that finds highly accurate pairwise and multiple network alignments through the iterative optimal matching of ‘local’ neighborhoods around focal nodes. Extensive performance assessment based on real networks as well as synthetic networks, for which the ground truth is known, demonstrates that MONACO clearly and consistently outperforms all other state-of-the-art network alignment algorithms that we have tested, in terms of accuracy, coherence and topological quality of the aligned network regions. Furthermore, despite the sharply enhanced alignment accuracy, MONACO remains computationally efficient and it scales well with increasing size and number of networks.

97 MATHEMATICS AND COMPUTING↗

Ternary alloy material prediction using genetic algorithm and cluster expansion

This thesis summarizes our study on the crystal structures prediction of Fe-V-Si system using genetic algorithm and cluster expansion. Our goal is to explore and look for new stable compounds. We started from the current ten known experimental phases, and calculated formation energies of those compounds using density functional theory (DFT) package, namely, VASP. The convex hull was generated based on the DFT calculations of the experimental known phases. Then we did random search on some metal rich (Fe and V) compositions and found that the lowest energy structures were body centered cube (bcc) underlying lattice, under which we did our computational systematic searches using genetic algorithm and cluster expansion. Among hundreds of the searched compositions, thirteen were selected and DFT formation energies were obtained by VASP. The stability checking of those thirteen compounds was done in reference to the experimental convex hull. We found that the composition, 24-8-16, i.e., Fe 3 VSi 2 is a new stable phase and it can be very inspiring to the future experiments.

36 MATERIALS SCIENCE↗

A mathematical assessment of the isolation random forest method for anomaly detection in big data

We present the mathematical analysis of the Isolation Random Forest Method (IRF Method) for anomaly detection, proposed by Liu F.T., Ting K.M. and Zhou Z. H. in their seminal work as a heuristic method for anomaly detection in Big Data. We prove that the IRF space can be endowed with a probability induced by the Isolation Tree algorithm (iTree). In this setting, the convergence of the IRF method is proved, using the Law of Large Numbers. Here, a couple of counterexamples are presented to show that the method is inconclusive and no certificate of quality can be given, when using it as a means to detect anomalies. Hence, an alternative version of the method is proposed whose mathematical foundation is fully justified. Furthermore, a criterion for choosing the number of sampled trees needed to guarantee confidence intervals of the numerical results is presented. Finally, numerical experiments are presented to compare the performance of the classic method with the proposed one.

97 MATHEMATICS AND COMPUTING↗

Reinforcement learning in discrete action space applied to inverse defect design

Abstract Reinforcement learning (RL) algorithms that include Monte Carlo Tree Search (MCTS) have found tremendous success in computer games such as Go, Shiga and Chess. Such learning algorithms have demonstrated super-human capabilities in navigating through an exhaustive discrete action search space. Motivated by their success in computer games, we demonstrate that RL can be applied to inverse materials design problems. We deploy RL for a representative case of the optimal atomic scale inverse design of extended defects via rearrangement of chalcogen (e.g. S) vacancies in 2D transition metal dichalcogenides (e.g. MoS 2 ). These defect rearrangements and their dynamics are important from the perspective of tunable phase transition in 2D materials i.e. 2H (semi-conducting) to 1T (metallic) in MoS 2 . We demonstrate the ability of MCTS interfaced with a reactive molecular dynamics simulator to efficiently sample the defect phase space and perform inverse design—starting from randomly distributed S vacancies, the optimal defect rearrangement of defects corresponds a line defect of S vacancies. We compare MCTS performance with evolutionary optimization i.e. genetic algorithms and show that MCTS converges to a better optimal solution (lower objective) and in fewer evaluations compared to GA. We also comprehensively evaluate and discuss the effect of MCTS hyperparameters on the convergence to solution. Overall, our study demonstrates the effectives of using RL approaches that operate in discrete action space for inverse defect design problems.

42 ENGINEERING↗

Quantum Graphical Models and Belief Propagation

Belief Propagation algorithms acting on Graphical Models of classical probability distributions, such as Markov Networks, Factor Graphs and Bayesian Networks, are amongst the most powerful known methods for deriving probabilistic inferences amongst large numbers of random variables. This paper presents a generalization of these concepts and methods to the quantum case, based on the idea that quantum theory can be thought of as a noncommutative, operator-valued, generalization of classical probability theory. Some novel characterizations of quantum conditional independence are derived, and definitions of Quantum n-Bifactor Networks, Markov Networks, Factor Graphs and Bayesian Networks are proposed. The structure of Quantum Markov Networks is investigated and some partial characterization results are obtained, along the lines of the Hammersley-Clifford theorem. A Quantum Belief Propagation algorithm is presented and is shown to converge on 1-Bifactor Networks and Markov Networks when the underlying graph is a tree. The use of Quantum Belief Propagation as a heuristic algorithm in cases where it is not known to converge is discussed. Applications to decoding quantum error correcting codes and to the simulation of many-body quantum systems are described.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

GRAVITATIONAL ENCOUNTERS AND THE EVOLUTION OF GALACTIC NUCLEI. I. METHOD

An algorithm is described for evolving the phase-space density of stars or compact objects around a massive black hole at the center of a galaxy. The technique is based on numerical integration of the Fokker–Planck equation in energy-angular momentum space, f(E, L, t), and includes, for the first time, diffusion coefficients that describe the effects of both random and correlated encounters (“resonant relaxation”), as well as energy loss due to emission of gravitational waves. Destruction or loss of stars into the black hole is treated by means of a detailed boundary-layer analysis. Performance of the algorithm is illustrated by calculating two-dimensional, time-dependent, and steady-state distribution functions and their corresponding loss rates.

79 ASTRONOMY AND ASTROPHYSICS↗

SU-E-J-29: Automatic Image Registration Performance of Three IGRT Systems for Prostate Radiotherapy

Purpose: To compare the performance of an automatic image registration algorithm on image sets collected on three commercial image guidance systems, and explore its relationship with imaging parameters such as dose and sharpness. Methods: Images of a CIRS Virtually Human Male Pelvis phantom (VHMP) were collected on the CBCT systems of Varian TrueBeam/OBI and Elekta Synergy/XVI linear accelerators, across a range of mAs settings; and MVCT on a Tomotherapy Hi-ART accelerator with a range of pitch. Using the 6D correlation ratio algorithm of XVI, each image was registered to a mask of the prostate volume with a 5 mm expansion. Registrations were repeated 100 times, with random initial offsets introduced to simulate daily matching. Residual registration errors were calculated by correcting for the initial phantom set-up error. Automatic registration was also repeated after reconstructing images with different sharpness filters. Results: All three systems showed good registration performance, with residual translations <0.5mm (1σ) for typical clinical dose and reconstruction settings. Residual rotational error had larger range, with 0.8°, 1.2° and 1.9° for 1σ in XVI, OBI and Tomotherapy respectively. The registration accuracy of XVI images showed a strong dependence on imaging dose, particularly below 4mGy. No evidence of reduced performance was observed at the lowest dose settings for OBI and Tomotherapy, but these were above 4mGy. Registration failures (maximum target registration error > 3.6 mm on the surface of a 30mm sphere) occurred in 5% to 10% of registrations. Changing the sharpness of image reconstruction had no significant effect on registration performance. Conclusions: Using the present automatic image registration algorithm, all IGRT systems tested provided satisfactory registrations for clinical use, within a normal range of acquisition settings.

60 APPLIED LIFE SCIENCES↗

Machine Learning for Improving Surface-Layer-Flux Estimates

Abstract Flows in the atmospheric boundary layer are turbulent, characterized by a large Reynolds number, the existence of a roughness sublayer and the absence of a well-defined viscous layer. Exchanges with the surface are therefore dominated by turbulent fluxes. In numerical models for atmospheric flows, turbulent fluxes must be specified at the surface; however, surface fluxes are not known a priori and therefore must be parametrized. Atmospheric flow models, including global circulation, limited area models, and large-eddy simulation, employ Monin–Obukhov similarity theory (MOST) to parametrize surface fluxes. The MOST approach is a semi-empirical formulation that accounts for atmospheric stability effects through universal stability functions. The stability functions are determined based on limited observations using simple regression as a function of the non-dimensional stability parameter representing a ratio of distance from the surface and the Obukhov length scale (Obukhov in Trudy Inst Theor Geofiz AN SSSR 1:95–115, 1946), $$z/L$$ z / L . However, simple regression cannot capture the relationship between governing parameters and surface-layer structure under the wide range of conditions to which MOST is commonly applied. We therefore develop, train, and test two machine-learning models, an artificial neural network (ANN) and random forest (RF), to estimate surface fluxes of momentum, sensible heat, and moisture based on surface and near-surface observations. To train and test these machine-learning algorithms, we use several years of observations from the Cabauw mast in the Netherlands and from the National Oceanic and Atmospheric Administration’s Field Research Division tower in Idaho. The RF and ANN models outperform MOST. Even when we train the RF and ANN on one set of data and apply them to the second set, they provide more accurate estimates of all of the fluxes compared to MOST. Estimates of sensible heat and moisture fluxes are significantly improved, and model interpretability techniques highlight the logical physical relationships we expect in surface-layer processes.

Meteorology & Atmospheric Sciences↗

A probabilistic graphical model approach to stochastic multiscale partial differential equations

We develop a probabilistic graphical model based methodology to efficiently perform uncertainty quantification in the presence of both stochastic input and multiple scales. Both the stochastic input and model responses are treated as random variables in this framework. Their relationships are modeled by graphical models which give explicit factorization of a high-dimensional joint probability distribution. The hyperparameters in the probabilistic model are learned using sequential Monte Carlo (SMC) method, which is superior to standard Markov chain Monte Carlo (MCMC) methods for multi-modal distributions. Finally, we make predictions from the probabilistic graphical model using the belief propagation algorithm. Numerical examples are presented to show the accuracy and efficiency of the predictive capability of the developed graphical model.

97 MATHEMATICS AND COMPUTING↗

WE-DE-201-04: Cross Validation of Knowledge-Based Treatment Planning for Prostate LDR Brachytherapy Using Principle Component Analysis

Purpose: To validate a knowledge-based algorithm for prostate LDR brachytherapy treatment planning. Methods: A dataset of 100 cases was compiled from an active prostate seed implant service. Cases were randomized into 10 subsets. For each subset, the 90 remaining library cases were registered to a common reference frame and then characterized on a point by point basis using principle component analysis (PCA). Each test case was converted to PCA vectors using the same process and compared with each library case using a Mahalanobis distance to evaluate similarity. Rank order PCA scores were used to select the best-matched library case. The seed arrangement was extracted from the best-matched case and used as a starting point for planning the test case. Any subsequent modifications were recorded that required input from a treatment planner to achieve V100>95%, V150<60%, V200<20%. To simulate operating-room planning constraints, seed activity was held constant, and the seed count could not increase. Results: The computational time required to register test-case contours and evaluate PCA similarity across the library was 10s. Preliminary analysis of 2 subsets shows that 9 of 20 test cases did not require any seed modifications to obtain an acceptable plan. Five test cases required fewer than 10 seed modifications or a grid shift. Another 5 test cases required approximately 20 seed modifications. An acceptable plan was not achieved for 1 outlier, which was substantially larger than its best match. Modifications took between 5s and 6min. Conclusion: A knowledge-based treatment planning algorithm for prostate LDR brachytherapy is being cross validated using 100 prior cases. Preliminary results suggest that for this size library, acceptable plans can be achieved without planner input in about half of the cases while varying amounts of planner input are needed in remaining cases. Computational time and planning time are compatible with clinical practice.

60 APPLIED LIFE SCIENCES↗