Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Parallel Algorithms”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources
PubMed1992

A parallel implementation of the backward error propagation neural network training algorithm: experiments in event identification.

An artificial neural-network-based (ANN) event detection and alarm generation system has been developed to aid clinicians in the identification of critical events commonly occurring in the anesthesia breathing circuit. To detect breathing circuit problems, the system monitored CO2 gas concentration, gas flow, and airway pressure. Various parameters were extracted from each of these input waveforms and fed into an artificial neural network. To develop truly robust ANNs, investigators are required to train their networks on large training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from 6.4 to 1.5 h. By reducing the total run time of the computation through parallelism, we were able to optimize many of the neural network's initial parameters. We conclude that use of the master-worker model of parallel computation is an excellent method for speeding up the backward error propagation neural network training algorithm.

Algorithms↗
PubMed1992

A parallel implementation of a multi-state Kalman filtering algorithm to detect ECG arrhythmias.

Detecting arrhythmias from the electrocardiogram (ECG) is of great importance for the continued development of intelligent cardiovascular monitors (ICM). An ICM's main goal is to present to the clinician a 'high-level' analysis of the patient's condition (e.g., the patient is slightly hypovolemic) based upon 'low-level' physiologic signals (e.g., blood pressure, heart rate, etc.). This paper reports on a parallel implementation of a multi-state Kalman filtering algorithm, within a prototype ICM, to help detect ECG arrhythmias. Preliminary test results show that the parallel, multi-state implementation performed exactly as the original sequential version. Several different rhythm disturbances were correctly identified after 3-5 beats. We conclude that our parallel implementation of the multi-state Kalman filter provides a faster and still reliable means of accurately detecting ECG arrhythmias in real-time.

Algorithms↗
PubMed2026

MethylModes: computationally efficient detection of multimodal distributions in DNA methylation data.

SUMMARY: MethylModes is an R package and Shiny application to identify multimodal distributions in human DNA methylation at individual CpG sites. Multimodal distributions, which can be the result of nearby genetic variation, environmental exposures, or assay artifacts, are susceptible to confounding and important to identify for methylation analysis. MethylModes is easily incorporated into existing quality control pipelines of array-based DNA methylation data. The underlying algorithm uses kernel smoothing of probe-level data to locate the number and location of peaks. The algorithm can be parallelized across probes for efficient implementation at genome-scale. We provide a case study implementation of MethylModes in the Health and Retirement Study as well as the Airwave Health Monitoring Study. AVAILABILITY AND IMPLEMENTATION: MethylModes is available on GitHub at https://github.com/lutiffan/methylModes as an R package wrapping an R Shiny application. We include a toy dataset to validate installation. The codebase is also published on Zenodo at https://doi.org/10.5281/zenodo.17448517.

DNA Methylation↗
PubMed1991

Recognition of the folding consensus in RNA secondary structures by the topological-filtering method.

Functionally homologous RNA sequences can substantially diverge in their primary sequences but it can be reasonably assumed that they are related in their higher-degree structures. The problem to find such structures and simultaneously satisfy as far as possible the free-energy-minimization criterion, is considered here in two aspects. Firstly a quantitative measure of the folding consensus among secondary structures is defined, translating each structure into a linear representation and using the correlation theorem to compare them. Secondly an algorithm for the parallel search for secondary structures according to the free-energy-minimization criterion, but with a filtering action on the basis of the folding consensus measure is presented. The method is tested on groups of RNA sequences different in origin and in functions, for which proposals of homologous secondary structures based on experimental data exist. A comparison of the results with a blank consisting of a search on the basis of the free energy minimization alone is always performed. In these tests the method shows its ability in obtaining, from different sequences, secondary structures characterized by a high-folding consensus measure also when lower free energy but not homologous structures are possible. Two applications are also shown. The first demonstrates the transfer of experimental data available for one sequence, to a functionally related and therefore homologous one. The second application is the possibility of using a topological probe in the search for precise structural motifs.

Algorithms↗
PubMed1991

Evaluation of a parallel implementation of the learning portion of the backward error propagation neural network: experiments in artifact identification.

Various methods have been proposed in an attempt to solve problems in artifact and/or alarm identification including expert systems, statistical signal processing techniques, and artificial neural networks (ANN). ANNs consist of a large number of simple processing units connected by weighted links. To develop truly robust ANNs, investigators are required to train their networks on huge training data sets, requiring enormous computing power. We implemented a parallel version of the backward error propagation neural network training algorithm in the widely portable parallel programming language C-Linda. A maximum speedup of 4.06 was obtained with six processors. This speedup represents a reduction in total run-time from approximately 6.4 hours to 1.5 hours. We conclude that use of the master-worker model of parallel computation is an excellent method for obtaining speedups in the backward error propagation neural network training algorithm.

Algorithms↗
PubMed1992

Molecular dynamics simulation on a network of workstations using a machine-independent parallel programming language.

Molecular dynamics simulations investigate local and global motion in molecules. Several parallel computing approaches have been taken to attack the most computationally expensive phase of molecular simulations, the evaluation of long range interactions. This paper reviews these approaches and develops a straightforward but effective algorithm using the machine-independent parallel programming language, Linda. The algorithm was run both on a shared memory parallel computer and on a network of high performance Unix workstations. Performance benchmarks were performed on both systems using two proteins. This algorithm offers a portable cost-effective alternative for molecular dynamics simulations. In view of the increasing numbers of networked workstations, this approach could help make molecular dynamics simulations more easily accessible to the research community.

Algorithms↗
PubMedDate not supplied

Structure analysis and classification of cervical cells using a processing system based on TV.

This paper presents preliminary results of a cell classification experiment using a new approach for feature extraction. The algorithm takes into account the special requirements of a fast parallel processing system (processor-oriented algorithms). A cell image is described by several hundred features derived from the nucleus only. The most significant features with respect to classification are determined by statistical analysis. Applying principal axis transform, a new feature set is computed, reduced considerably in dimensions. The data base (1,925 cell images of Papanicolaou-stained cervical specimens) was divided into a training set (963 images) and a test set (962 images). The classification results of the test set show that the recognition rate for the two-class problem (normal, suspicious) is better than 91%, using only ten morphologic features.

Cervix Mucus↗
PubMed1992

Preverbal and verbal counting and computation.

We describe the preverbal system of counting and arithmetic reasoning revealed by experiments on numerical representations in animals. In this system, numerosities are represented by magnitudes, which are rapidly but inaccurately generated by the Meck and Church (1983) preverbal counting mechanism. We suggest the following. (1) The preverbal counting mechanism is the source of the implicit principles that guide the acquisition of verbal counting. (2) The preverbal system of arithmetic computation provides the framework for the assimilation of the verbal system. (3) Learning to count involves, in part, learning a mapping from the preverbal numerical magnitudes to the verbal and written number symbols and the inverse mappings from these symbols to the preverbal magnitudes. (4) Subitizing is the use of the preverbal counting process and the mapping from the resulting magnitudes to number words in order to generate rapidly the number words for small numerosities. (5) The retrieval of the number facts, which plays a central role in verbal computation, is mediated via the inverse mappings from verbal and written numbers to the preverbal magnitudes and the use of these magnitudes to find the appropriate cells in tabular arrangements of the answers. (6) This model of the fact retrieval process accounts for the salient features of the reaction time differences and error patterns revealed by experiments on mental arithmetic. (7) The application of verbal and written computational algorithms goes on in parallel with, and is to some extent guided by, preverbal computations, both in the child and in the adult.

Animals↗
PubMed1991

A parallel computing approach to genetic sequence comparison: the master-worker paradigm with interworker communication.

We have implemented a parallel version of a dynamic programming biological sequence comparison algorithm to study the potential applicability of using parallel computers for genetic sequence comparisons. Our parallel program is built using C-Linda, a machine-independent parallel programming language, and was tested on both a 10 CPU Sequent Symmetry and a 64 CPU Intel Hypercube. C-Linda implements a shared associative memory model, "tuple space," through which multiple processes can communicate and coordinate control. In our master-worker (MW) parallel implementation, a master process creates several worker processes, extracts a test sequence and multiple library sequences from a database and stores them in tuple space. Each worker reads the test sequence and then repeatedly extracts library strings from tuple space, performs pairwise sequence comparison using a local comparison algorithm to generate a similarity score, and returns the similarity scores to tuple space. The master collects the scores from tuple space and identifies the best match over all library sequences. We also implemented a method of global interworker communication to reduce the total search time by stopping those string comparisons that had no chance of improving on the current best match. Comparisons of the total run time, speedup, and efficiency were made for parallel and sequential versions of a basic MW implementation as well as versions with the global abort threshold.

Algorithms↗
PubMed1979

Utilization of cross-plane rays for three-dimensional reconstruction by filtered back-projection.

Present popular computed tomography (CT) algorithms reconstruct an object from the ray measurements lying on a set of parallel planes. This paper presents an algorithm that can also utilize "cross-plane" rays (i.e., rays that cross through many planes) to reconstruct the object. In this reconstruction algorithm, the ray measurements are grouped into two-dimensional projections, filtered, and stored. The filtered projections can then be back-projected onto a three-dimensional matrix or any plane through the three-dimensional volume. General theoretical aspects are presented and then applied to the special case in which ray measurements have been made in all directions. The algorithm is tested using computer-generated data. Expressions for the noise power spectrum and the variance in the reconstruction are derived. It is shown that the noise-to-signal ratio per detected photon for this reconstruction method is close to a theoretical limit, as it also is for normal CT. The ability to use ray measurements that cross many planes is especially useful in emission CT, where order-of-magnitude improvements in image quality per unit dose can be achieved.

Computers↗
PubMed1991

Fast space-filling molecular graphics using dynamic partitioning among parallel processors.

We present a novel algorithm for the efficient generation of high-quality space-filling molecular graphics that is particularly appropriate for the creation of the large number of images needed in the animation of molecular dynamics. Each atom of the molecule is represented by a sphere of an appropriate radius, and the image of the sphere is constructed pixel-by-pixel using a generalization of the lighting model proposed by Porter (Comp. Graphics 1978, 12, 282). The edges of the spheres are antialiased, and intersections between spheres are handled through a simple blending algorithm that provides very smooth edges. We have implemented this algorithm on a multiprocessor computer using a procedure that dynamically repartitions the effort among the processors based on the CPU time used by each processor to create the previous image. This dynamic reallocation among processors automatically maximizes efficiency in the face of both the changing nature of the image from frame to frame and the shifting demands of the other programs running simultaneously on the same processors. We present data showing the efficiency of this multiprocessing algorithm as the number of processors is increased. The combination of the graphics and multiprocessor algorithms allows the fast generation of many high-quality images.

Algorithms↗
PubMed1991

Molecular dynamics simulation on a network of workstations using a machine-independent parallel programming language.

Molecular dynamics simulations investigate local and global motion in molecules. Several parallel computing approaches have been taken to attack the most computationally expensive phase of molecular simulations, the evaluation of long range interactions. This paper develops a straightforward but effective algorithm for molecular dynamics simulations using the machine-independent parallel programming language, Linda. The algorithm was run both on a shared memory parallel computer and on a network of high performance Unix workstations. Performance benchmarks were performed on both systems using two proteins. This algorithm offers a portable cost-effective alternative for molecular dynamics simulations. In view of the increasing numbers of networked workstations, this approach could help make molecular dynamics simulations more easily accessible to the research community.

Algorithms↗
PubMedDate not supplied

Projection domain compensation of missing angles for fan-beam CT reconstruction.

An improved method is proposed for fan-beam computed tomographic (CT) reconstruction from data with limited views. Compensation for the missing projections for fan-beam CT can be partially accomplished by using the coincident ray or by an interpolation technique using circular sample theory. In this article, the authors propose a more accurate compensation method for the missing projections whether the coincident ray pairs exist or not. The fan-beam reprojection algorithm, which is the inverse operator of the convolution filter, was extended from the projection space iteration reconstruction-reprojection (PSIRR) in parallel beam geometry. In addition, this algorithm was validated by applying the Shepp-Logan phantom for a computer simulation in the equi-angular fan-beam CT geometry.

Algorithms↗
PubMed1991

Computer systems for three-dimensional diagnostic imaging: an examination of the state of the art.

This survey reviews three-dimensional (3D) medical imaging machines and 3D medical imaging operations. The survey is designed to provide a snapshot overview of the present state of computer architectures for 3D medical imaging. The basic volume manipulation, object segmentation, and graphics operations required of a 3D medical imaging machine are described and sample algorithms are presented. The architecture and 3D imaging algorithms employed in 11 machines which render medical images are assessed. The performance of the machines is compared across several dimensions, including image resolution, elapsed time to form an image, imaging algorithms employed in the machine, and the degree of parallelism employed in the architecture. The innovation in each machine, whether architectural or algorithmic, is described in detail. General trends for future developments in this field are delineated and an extensive bibliography is provided.

Computer Systems↗
PubMed1991

Efficient detection of three-dimensional structural motifs in biological macromolecules by computer vision techniques.

Macromolecules carrying biological information often consist of independent modules containing recurring structural motifs. Detection of a specific structural motif within a protein (or DNA) aids in elucidating the role played by the protein (DNA element) and the mechanism of its operation. The number of crystallographically known structures at high resolution is increasing very rapidly. Yet, comparison of three-dimensional structures is a laborious time-consuming procedure that typically requires a manual phase. To date, there is no fast automated procedure for structural comparisons. We present an efficient O(n3) worst case time complexity algorithm for achieving such a goal (where n is the number of atoms in the examined structure). The method is truly three-dimensional, sequence-order-independent, and thus insensitive to gaps, insertions, or deletions. This algorithm is based on the geometric hashing paradigm, which was originally developed for object recognition problems in computer vision. It introduces an indexing approach based on transformation invariant representations and is especially geared toward efficient recognition of partial structures in rigid objects belonging to large data bases. This algorithm is suitable for quick scanning of structural data bases and will detect a recurring structural motif that is a priori unknown. The algorithm uses protein (or DNA) structures, atomic labels, and their three-dimensional coordinates. Additional information pertaining to the structure speeds the comparisons. The algorithm is straightforwardly parallelizable, and several versions of it for computer vision applications have been implemented on the massively parallel connection machine. A prototype version of the algorithm has been implemented and applied to the detection of substructures in proteins.

Algorithms↗
PubMed1991

Multiresolution, error-convergence halftone algorithm.

A new halftone algorithm is described. The algorithm is designed for implementation on a parallel architecture in order to provide fast, progressive coding of moderate-resolution images. The design is based on a multiresolution, hierarchical, pyramidal structure. At each pyramid level, the binarized image is compared with the original, gray-tone image over a successively larger window of pixels for calculation of a weighted averaged error. Within each level, selected binarized pixels are tested for possible changes in the binary assignment. The binary assignment is changed if the change results in a lower average error over the entire window. Varying the selection of test pixels can cause the same process to provide clustered-dot patterns and dithering. A comparison of performance with the best implementation of the error-propagation algorithm is presented visually. Quality is compared also in terms of isotropy of the texture and the appropriate blue-noise characteristics in areas of uniform gray tone. The benefits of this algorithm are realized with moderate-resolution display of the order of 512 dots X 512 dots. The processing can be carried out on smaller blocks since the results can be combined without any visible seams or edge effects.

Algorithms↗
PubMed1991

Scanning protein sequence databanks using a distributed processing workstation network.

The programme pscan has been developed to distribute protein databank scans over a network of computers that share a common file system. pscan may be used in conjunction with most conventional sequence comparison programmes with few modifications. In test runs using the Smith-Waterman dynamic programming algorithm, the time required to scan a 6858 sequence databank using a query sequence 740 residues long was reduced from approximately 50 min for a single processor, to approximately 11 minutes for five processors. Accordingly, pscan provides a low-cost, portable alternative to dedicated parallel processing computers.

Algorithms↗
PubMed1991

Bayesian image reconstruction for emission tomography incorporating Good's roughness prior on massively parallel processors.

Since the introduction by Shepp and Vardi [Shepp, L. A. & Vardi, Y. (1982) IEEE Trans. Med. Imaging 1, 113-121] of the expectation-maximization algorithm for the generation of maximum-likelihood images in emission tomography, a number of investigators have applied the maximum-likelihood method to imaging problems. Though this approach is promising, it is now well known that the unconstrained maximum-likelihood approach has two major drawbacks: (i) the algorithm is computationally demanding, resulting in reconstruction times that are not acceptable for routine clinical application, and (ii) the unconstrained maximum-likelihood estimator has a fundamental noise artifact that worsens as the iterative algorithm climbs the likelihood hill. In this paper the computation issue is addressed by proposing an implementation on the class of massively parallel single-instruction, multiple-data architectures. By restructuring the superposition integrals required for the expectation-maximization algorithm as the solutions of partial differential equations, the local data passage required for efficient computation on this class of machines is satisfied. For dealing with the "noise artifact" a Markov random field prior determined by Good's rotationally invariant roughness penalty is incorporated. These methods are demonstrated on the single-instruction multiple-data class of parallel processors, with the computation times compared with those on conventional and hypercube architectures.

Algorithms↗