Search

It is prone to get stuck in a local minimum when solving the Traveling Salesman Problem (TSP) by the traditional Hopfield neural network (HNN) and hard to converge to an efficient solution, resulting from the defect of the penalty method used by the HNN. In order to mend this defect, an accelerated augmented Lagrangian Hopfield neural network (AALHNN) algorithm was proposed in this paper. This algorithm gets out of the dilemma of penalty method by Lagrangian multiplier method, ensuring that the solution to the TSP is undoubtedly efficient. The second order factor added in the algorithm stabilizes the neural network dynamic model of the problem, thus improving the efficiency of solution. In this paper, when solving the TSP by AALHNN, some changes were made to the TSP models of Hopfield and Tank. Say, constraints of TSP are multiplied by Lagrange multipliers and augmented Lagrange multipliers respectively, The augmented Lagrange function composed of path length function can ensure robust convergence and escape from the local minimum trap. The Lagrange multipliers are updated by using nesterov acceleration technique. In addition, it was theoretically proved that the extremum obtained by this improved algorithm is the optimal solution of the initial problem and the approximate optimal solution of the TSP was successfully obtained several times in the simulation experiment. Compared with the traditional HNN, this method can ensure that it is effective for TSP solution and the solution to the TSP obtained is better.

Auditory evoked fields (AEFs) are commonly studied, yet their underlying neural mechanisms remain poorly understood. Here, we used the biophysical modelling software Human Neocortical Neurosolver (HNN) whose foundation is a canonical neocortical circuit model to interpret the cell and network mechanisms contributing to macroscale AEFs elicited by a simple tone, measured with magnetoencephalography. We found that AEFs can be reproduced by activating the neocortical circuit through a layer specific sequence of feedforward and feedback excitatory synaptic drives, similar to prior simulation of somatosensory evoked responses, supporting the notion that basic structures and activation patterns are preserved across sensory regions. We also applied the modeling framework to develop and test predictions on neural mechanisms underlying AEF differences in the left and right hemispheres, as well as in hemispheres contralateral and ipsilateral to the presentation of the auditory stimulus. We found that increasing the strength of the excitatory synaptic cortical feedback inputs to supragranular layers simulates the commonly observed right hemisphere dominance, while decreasing the input latencies and simultaneously increasing the number of cells contributing to the signal accounted for the contralateral dominance. These results provide a direct link between human data and prior animal studies and lay the foundation for future translational research examining the mechanisms underlying alteration in this fundamental biomarker of auditory processing in healthy cognition and neuropathology. [ABSTRACT FROM AUTHOR]

Статья посвящена анализу проблемы определения автора исходного кода, которая представляет интерес для исследователей в области информационной безопасности, компьютерной криминалистики, оценки качества образовательного процесса, защиты интеллектуальной собственности. Представлен подробный анализ современных решений проблемы. Предлагаются две новые методики идентификации на основе алгоритмов машинного обучения: машины опорных векторов, фильтра быстрой корреляции и информативных признаков; гибридной сверточно-рекуррентной нейронной сети. Эксперименты проводились на базе исходных кодов, написанных на наиболее популярных языках программирования. В экспериментальную базу вошли экземпляры исходных кодов, написанных на Java, C++, Python, PHP, JavaScript, C, C# и Ruby. Данные были получены с веб-сервиса для хостинга IT-проектов Github. Общее количество исходных кодов превышает 150 тысяч образцов, средняя длина каждого из которых составляет 850 символов. Размер корпуса — 542 автора. С помощью перекрестной проверки по 10 блокам оценена точность разработанных методик для различного количества авторов. Для наиболее популярного языка программирования Java проведен дополнительный ряд экспериментов с количеством авторов от 2 до 50 и приведены графики зависимости точности идентификации от размера корпуса. Анализ результатов показал, что методика на основе гибридной нейронной сети способна достигать точности 97%, что является наилучшим результатом на сегодняшний день. Методика на основе машины опорных векторов позволила добиться точности 96%. Гибридная нейронная сеть оказалась точнее машины опорных векторов в среднем на 5%.

Satellite image segmentation has gotten bunches of consideration of late because of the accessibility of commented on high-goals image informational indexes caught by the last age of satellites. The issue of fragmenting a satellite image can be characterized as ordering (or marking) every pixel of the image as indicated by various classes, for example, structures, streets, water, etc. In this paper centered to build up a satellite image segmenting process by utilizing distinctive optimization methods. The work is prepared dependent on three stages that are RGB change, preprocessing, and division. At first the database images are assembled from the database at that point select the blue band images by performing RGB change. To improve the differentiation and furthermore decreasing the commotion of these chose blue band images, Hopfield neural network (HNN) is utilized. After image upgrade, the images are fragmented dependent on fuzzy C means (FCM) clustering method. The images are clustered and segmented in the way of optimizing the centroid in FCM utilizing oppositional crow search algorithm. The exhibition of the proposed framework is investigated dependent on the presentation measurements, for example, affectability, particularity and accuracy. From the outcomes, the proposed strategy diminished the computational time by expanding the accuracy of 98.3% with HNN system. [ABSTRACT FROM AUTHOR]

Purpose: Our clinical understanding of the relationship between 3D bone morphology and knee osteoarthritis, as well as our ability to investigate potential causative factors of osteoarthritis, has been hampered by the time‐intensive nature of manually segmenting bone from MR images. Thus, we aim to develop and validate a fully automated deep learning framework for segmenting the patella and distal femur cortex, in both adults and actively growing adolescents. Methods: Data from 93 subjects, obtained from on institutional review board–approved protocol, formed the study database. 3D sagittal gradient recalled echo and gradient recalled echo with fat saturation images and manual models of the outer cortex were available for 86 femurs and 90 patellae. A deep‐learning–based 2D holistically nested network (HNN) architecture was developed to automatically segment the patella and distal femur using both single (sagittal, uniplanar) and 3 cardinal plane (triplanar) methodologies. Errors in the surface‐to‐surface distances and the Dice coefficient were the primary measures used to quantitatively evaluate segmentation accuracy using a 9‐fold cross‐validation. Results: Average absolute errors for segmenting both the patella and femur were 0.33 mm. The Dice coefficients were 97% and 94% for the femur and patella. The uniplanar, relative to the triplanar, methodology produced slightly superior segmentation. Neither the presence of active growth plates nor pathology influenced segmentation accuracy. Conclusion: The proposed HNN with multi‐feature architecture provides a fully automatic technique capable of delineating the often indistinct interfaces between the bone and other joint structures with an accuracy better than nearly all other techniques presented previously, even when active growth plates are present. [ABSTRACT FROM AUTHOR]

The most popular applications of Hopfield neural network algorithm (HNN) arepattern recognition and classification. But the HNN has some limitation like the localminima (oscillation) problem. In this paper a novel method of combining an activecontour (snake) and an artificial neural network to behave together as pattern recognitionand classification is presented. The approach used the technique of the gradient vectorflow (GVF) that locate the boundary of target pattern (image) then pass it to a classifierbuilt by Hopfield algorithm to classify it according to one of the storage pattern. Thesnakes can find the boundaries of objects so it is very accurate to take the shape of theobject wanted, that will eliminate the noise from the original image and reduce the biterror rate of the Hopfield network to 0.215 and overcome the oscillation state inrecognition of the entered pattern. MATLAB 7 program have been used for thesimulation of the active contour and the pattern classification.

Wireless Sensor Network (WSN) consists of large number of sensor nodes which are limited in battery power and communication range and are having multi-model sensing capability. One of the most significant applications of wireless sensor network is environment monitoring. A wireless sensor network (WSN) entail of spatially distributed self-sufficient sensors to monitor physical or conservational requirements, such as pressure, sound, vibration, temperature, motion or pollutants and to collaboratively pass their data to a main location(base station or sink) through the network. Wireless sensor networks are equipped with only Omni-directional antennas, which can cause high collisions. It is shown that the per node throughput in such networks is decreased with the increased number of nodes. Thus, the transmission with multiple short-range hops is preferred to reduce the interference. Packet collision causes packet loss and wastes resources in wireless networks. It becomes even worse in dense WSNs, due to burst-traffic and congestion around sinks. In this paper, we propose a neural network scheme to recover collided packets. The proposed work avoids the collision using Hopfield neural network and increases the packet delivery ratio and throughput of the network and due to this end 2 end delay is decreased. [ABSTRACT FROM AUTHOR]

Abstract: We present here an improvisation of HNN (Panchal, Bhavesh et al., 2001) called RD 3D HNCAN for backbone (HN, CA and 15N) assignment in both folded and unfolded proteins. This is a reduced dimensionality experiment which employs CA chemical shifts to improve dispersion. Distinct positive and negative peak patterns of various triplet segments along the polypeptide chain observed in HNN are retained and these provide start and check points for the sequential walk. Because of co-incrementing of CA and 15N, peaks along one of the dimensions appear at sums and differences of the CA and 15N chemical shifts. This changes the backbone assignment protocol slightly and we present this in explicit detail. The performance of the experiment has been demonstrated using Ubiquitin and Plasmodium falciparum P2 proteins. The experiment is particularly valuable when two neighboring amino acid residues have nearly identical backbone 15N chemical shifts. [Copyright &y& Elsevier]

In this paper, a Hybrid Neural Network trained by Modified Particle Swarm Optimization (MPSO-HNN) is first proposed for predicting warpage. Then the minimum warpage problem is solved by MPSO-HNN coupled with MPSO algorithm based on process parameters in injection molding. A model based on Taguchi experiment is designed to find process parameters' optimum levels by Signal-to-Noise (SN) ratio. All results show that proposed algorithm needs less calculating amounts and spends less time and has better generalization ability than those in related literature. Meanwhile, the method of combining MPSO-HNN with MPSO is suitable to optimization warpage compared with the Taguchi method. [ABSTRACT FROM AUTHOR]

Abstract: HNN has proven to be an extremely valuable experiment for rapid and unambiguous backbone (HN, 15N) assignment in (13C, 15N) labeled proteins. However, low sensitivity of the experiment is often a limiting factor, especially when the transverse relaxation times (T 2) are short. We show here that BEST modification Schanda et al. (2006) increases the sensitivity per unit time by more than a factor of 2.0 and thus substantially increases the speed of data collection; good 3D data can be collected in 8–10h. Next, we present a simple method for amino-acid type identification based on simple 2D versions of the HNN experiment, labeled here as 2D-(HN)NH. Each of these experiments which produce anchor points for Gly, Ala, Ser/Thr residues, can be recorded in less than an hour. These enable rapid data acquisition, rapid analysis, and consequently rapid assignment of backbone (HN, 15N) resonances. The 2D-(HN)NH experiment does not involve aliphatic/aromatic protons and hence can be applied to deuterated protein samples as well, which is an additional advantage. The experiments have been demonstrated with human ubiquitin (76 aa) and acetic-acid denatured HIV-1 protease (99 aa), as representatives of folded and unfolded protein systems, respectively. [Copyright &y& Elsevier]

We describe here the tunability of the HNN experiment to obtain certain residue specific peak patterns in the spectra of (15N, 13C) labeled proteins. This is achieved by tuning a band-selective 180° pulse on the carbon channel in the pulse sequence, whereby one can tamper with the Cα–Cβ coupling evolutions for the different residues. Specifically, we generate distinctive peak patterns for serine and threonine and their neighbors in the different planes of the three dimensional spectrum. These provide useful anchor points during sequential assignment of backbone resonances. The performance of this experiment, referred to as HNN-ST here, is demonstrated using two proteins, one properly folded and the other completely denatured. With the availability of high field spectrometers, techniques such as TROSY, and ever increasing sensitivities in the probes, this experiment with its large number of check points has a great potential for rapid and unambiguous backbone resonance assignment in large proteins. [ABSTRACT FROM AUTHOR]

The refolding of barstar from its urea-unfolded state has been studied extensively using various spectroscopic probes and real-time NMR, which provide global and residue-specific information, respectively, about the folding process. Here, a preliminary structural characterization by NMR of barstar in 8 M urea has been carried out at pH 6.5 and 25°C. Complete backbone resonance assignments of the urea-unfolded protein were obtained using the recently developed three-dimensional NMR techniques of HNN and HN(C)N. The conformational propensities of the polypeptide backbone in the presence of 8 M urea have been estimated by examining deviations of secondary chemical shifts from random coil values. For some residues that belong to helices in native barstar, 13Cα and 13CO secondary shifts show positive deviations in the urea-unfolded state, indicating that these residues have propensities toward helical conformations. These residues are, however, juxtaposed by residues that display negative deviations indicative of propensities toward extended conformations. Thus, segments that are helical in native barstar are unlikely to preferentially populate the helical conformation in the unfolded state. Similarly, residues belonging to β-strands 1 and 2 of native barstar do not appear to show any conformational preferences in the unfolded state. On the other hand, residues belonging to the β-strand 3 segment show weak nonnative helical conformational preferences in the unfolded state, indicating that this segment may possess a weak preference for populating a helical conformation in the unfolded state. [ABSTRACT FROM AUTHOR]

Two triple resonance experiments, HNN and HN(C)N, are presented which correlate HN and 15N resonances sequentially along the polypeptide chain of a doubly (13C, 15N) labeled protein. These incorporate several improvements over the previously published sequences for a similar purpose and have several novel features. The spectral characteristics enable direct identification of certain triplets of residues, which provide many starting points for the sequential assignment procedure. The experiments are sensitive and their utility has been demonstrated with a 22 kDa protein under unfolding conditions where most of the standard triple resonance experiments such as HNCA, CBCANH etc. have limited success because of poor amide, Cα and Cβ chemical shift dispersions. [ABSTRACT FROM AUTHOR]

It is prone to get stuck in a local minimum when solving the Traveling Salesman Problem (TSP) by the traditional Hopfield neural network (HNN) and hard to converge to an efficient solution, resulting from the defect of the penalty method used by the HNN. In order to mend this defect, an accelerated augmented Lagrangian Hopfield neural network (AALHNN) algorithm was proposed in this paper. This algorithm gets out of the dilemma of penalty method by Lagrangian multiplier method, ensuring that the solution to the TSP is undoubtedly efficient. The second order factor added in the algorithm stabilizes the neural network dynamic model of the problem, thus improving the efficiency of solution. In this paper, when solving the TSP by AALHNN, some changes were made to the TSP models of Hopfield and Tank. Say, constraints of TSP are multiplied by Lagrange multipliers and augmented Lagrange multipliers respectively, The augmented Lagrange function composed of path length function can ensure robust convergence and escape from the local minimum trap. The Lagrange multipliers are updated by using nesterov acceleration technique. In addition, it was theoretically proved that the extremum obtained by this improved algorithm is the optimal solution of the initial problem and the approximate optimal solution of the TSP was successfully obtained several times in the simulation experiment. Compared with the traditional HNN, this method can ensure that it is effective for TSP solution and the solution to the TSP obtained is better.

In the present study a reliable and structural decision system for production sequence of polymeric products is developed. Minimization of the amount of off-specs is the main objective in the decision of production sequence to maximize profit. Off-specs are generated when the production sequence of polymeric products is changed. The amount of off-specs depends on changes of product grades. In the present study we applied the traveling salesman problem (TSP) to achieve optimal decision of production sequence. To solve the optimal decision problem formulated by TSP, we employed three different approaches such as Branch and Bound (B&B) method, Dynamic Programming (DP) method and Hopfield Neural Network (HNN) method. Production sequences computed based on the actual plant off-spec data were compared with the sequences employed in the actual plant operation. From the comparison the decision method proposed in the present study showed increased profits and reduced off-specs. [ABSTRACT FROM AUTHOR]

Mobile and wireless systems beyond 3G are being designed under the user-centric paradigm. Dynamic resource allocation (DRA) is a topic on intensive research to address efficiently such paradigm. Hopfield neural networks (HNN) have proved useful in the past to solve this kind of complex optimization problems. Recently, various approaches have been proposed to realize HNN-based user-centric DRA. However, the initial algorithms suffer from severe instability problems impacting the overall performance. This letter analyses the source of the existing limitations and proposes an enhanced formulation, ensuring maximum resource utilization while optimizing the convergence of the neural network. The letter highlights the improved performance in terms of optimum convergence and bandwidth utilization [ABSTRACT FROM PUBLISHER]

The GTPase ARL13B is localized to primary cilia; small cellular protrusions that act as antennae. Its defective ARL13B hennin (HNN) variant is linked causally with Joubert Syndrome, a developmental ciliopathy attributed to poor sensing of extracellular chemical gradients. We tested the hypothesis that impaired detection of extracellular voltage gradients also contributes to the HNN phenotype. In vitro, extracellular electric fields stimulated migration of wild type (WT) and HNN fibroblasts toward the cathode but the field only increased the migration speed of WT cells. Cilia on WT cells did not align to the field vector. HNN cells divided more slowly than WT cells, arresting at the G2/M phase. Mechanistically, HNN cells had reduced phospho-ERK1/2 signaling and elevated levels of Suppressor of Fused protein. These suggest that cells may not be able to read extracellular chemical cues appropriately, resulting in deficits in cell migration and proliferation. Finally, an increase in tubulin stabilization (more detyrosinated tubulin) confirmed the general stagnation of HNN cells, which may further contribute to slower migration and cell cycle progression. We conclude that Arl13b dysfunction resulted in HNN cell stagnation due to poor growth factor signaling and impaired detection of extracellular electrical gradients, and that the role of Arl13b in cell proliferation may be understated.

Background: Little is known about trends in morbidity and/or mortality due to asthma in Latin America., Objective: To examine trends in hospitalizations and mortality due to asthma from 1997-2000 to 2011 in Costa Rica., Methods: The rates of hospitalization due to asthma were calculated for each sex in 3 age groups from 1997 to 2011. The number of deaths due to asthma was first calculated for all groups and then for each sex in 3 age groups from 2000 to 2011. All analyses were conducted over the entire period and separately for the periods before and after a National Asthma Program (NAP) in 2003. Data also were available for prescriptions for beclomethasone since 2004. All analyses were conducted by using Epi Info., Results: Substantial reductions were found in hospitalizations and deaths due to asthma in Costa Ricans (eg, from 25 deaths in 2000 to 5 deaths in 2011). Although, the percentage decrement in the rates of hospitalization for asthma in subjects <20 years old was similar before and after the NAP, the reduction in both deaths due to asthma and rates of asthma hospitalizations in older subjects were more pronounced after the NAP, when prescriptions for beclomethasone were also increased by approximately 129%., Conclusion: In Costa Rica, there was a marked decrement in hospitalizations and mortality due to asthma from 1997-2000 to 2011. In younger subjects, this is likely due to guidelines that, since 1988, recommend inhaled corticosteroids for persistent asthma. In older adults, the NAP probably enhanced reductions in hospitalizations and deaths due to asthma through inhaled corticosteroid use., (Copyright © 2013 American Academy of Allergy, Asthma & Immunology. Published by Elsevier Inc. All rights reserved.)

A three-dimensional nuclear magnetic resonance (NMR) pulse sequence named as hNCOcanH has been described to aid rapid sequential assignment of backbone resonances in (15)N/(13)C-labeled proteins. The experiment has been derived by a simple modification of the previously described HN(C)N pulse sequence [Panchal et al., J. Biomol. NMR 20 (2001) 135-147]; t2 evolution is used to frequency label (13)C' rather than (15)N (similar trick has also been used in the design of hNCAnH pulse sequence from hNcaNH [Frueh et al., JACS, 131 (2009) 12880-12881]). The modification results in a spectrum equivalent to HNCO, but in addition to inter-residue correlation peaks (i.e. Hi , Ci-1), the spectrum also contains additional intra-residue correlation peaks (i.e. Hi-1 , Ci-1) in the direct proton dimension which has maximum resolution. This is the main strength of the experiment and thus, even a small difference in amide (1) H chemical shifts (5-6 Hz) can be used for establishing a sequential connectivity. This experiment in combination with the HNN experiment described previously [Panchal et al., J. Biomol. NMR 20 (2001) 135-147] leads to a more robust assignment protocol for backbone resonances ((1) H(N) , (15)N) than could be derived from the combination of HNN and HN(C)N experiments [Bhavesh et al., Biochemistry, 40 (2001) 14727-14735]. Further, this new protocol enables assignment of (13)C' resonances as well. We believe that the experiment and the protocol presented here will be of immense value for structural-and functional-proteomics research by NMR. Performance of this experiment has been demonstrated using (13)C/(15)N labeled ubiquitin., (Copyright © 2011 John Wiley & Sons, Ltd.)