Resonance Network Research Articles

Reconstruction of chaotic dynamics via a network of stochastic resonance neurons and its application to speech

Recently, a great deal of attention has been paid tostochastic resonance as a new framework to understand sensory mechanisms of biological systems. Stochastic resonance explains important properties of sensory neurons that accurately detect weak input stimuli by using a small amount of internal noise. In particular, Collins et al. reported that a network of stochastic resonance neurons gives rise to a robust sensory function for detecting a variety of complex input signals. In this study, we investigate effectiveness of such stochastic resonance neural networks to chaotic input signals. Using the Rossler equations, we analyze the network's capability to detect chaotic dynamics. We also apply the stochastic resonance network systems to speech signals, and examine a plausibility of the stochastic resonance neural network as a possible model for the human auditory system.

Predictive value of gadolinium-enhanced magnetic resonance imaging for relapse rate and changes in disability or impairment in multiple sclerosis: a meta-analysis

Reliable prognostic factors are lacking for multiple sclerosis (MS). Gadolinium enhancement in magnetic resonance imaging (MRI) of the brain detects with high sensitivity disturbance of the blood-brain barrier, an early event in the development of inflammatory lesions in MS. To investigate the prognostic value of gadolinium-enhanced MRI, we did a meta-analysis of longitudinal MRI studies. From the members of MAGNIMS (European Magnetic Resonance Network in Multiple Sclerosis) and additional centres in the USA, we collected data from five natural-course studies and four placebo groups of clinical trials completed between 1992 and 1995. We included a total of 307 patients, 237 with relapsing disease course and 70 with secondary progressive disease course. We investigated by regression analysis the relation between initial count of gadolinium-enhancing lesions and subsequent worsening of disability or impairment as measured by the expanded disability status scale (EDSS) and relapse rate. The relapse rate in the first year was predicted with moderate ability by the mean number of gadolinium-enhancing lesions in monthly scans during the first 6 months (relative risk per five lesions 1.13, p=0.023). The predictive value of the number of gadolinium-enhancing lesions in one baseline scan was less strong. The best predictor for relapse rate was the variation (SD) of lesion counts in the first six monthly scans which allowed an estimate of relapse in the first year (relative risk 1.2, p=0.020) and in the second year (risk ratio=1.59, p=0.010). Neither the initial scan nor monthly scans over six months were predictive of change in the EDSS in the subsequent 12 months or 24 months. The mean of gadolinium-enhancing-lesion counts in the first six monthly scans was weakly predictive of EDSS change after 1 year (odds ratio=1.34, p=0.082) and 2 years (odds ratio=1.65, p=0.049). Although disturbance of the blood-brain barrier as shown by gadolinium enhancement in MRI is a predictor of the occurrence of relapses, it is not a strong predictor of the development of cumulative impairment or disability. This discrepancy supports the idea that variant pathogenetic mechanisms are operative in the occurrence of relapses and in the development of long-term disability in MS.

An LDMOS VHF class-E power amplifier using a high-Q novel variable inductor

In this paper, an lateral diffused metal-oxide-semiconductor-based very high-frequency class-E power amplifier has been investigated theoretically and experimentally. Simulations were verified by amplifier measurements and a record-high class-E output power was obtained at 144 MHz, which is in excellent agreement with simulations. The key of the results is the use of efficient device models, simulation tools, and the invention of a novel high-Q inductor for the output series resonance network. The latter allows for low losses in the output network and, simultaneously, a wide tuning range for maximum output power or maximum efficiency optimization.

On-line monitoring of tool breakage in face milling using a self-organized neural network

This study introduces a new tool breakage monitoring methodology consisting of an unsupervised neural network combined with an adaptive time-series modeling algorithm. Cutting force signals are modeled by a discrete autoregressive model in which parameters are estimated recursively at each sampling instant using a parameter-adaptation algorithm based on a recursive least square. The experiment shows that monitoring the evolution of autoregressive parameters during milling is effective for detecting tool breakage. An adaptive resonance network based on Grossberg's adaptive resonance theory (ART 2) is employed for clustering tool states using model parameters, and this network has unsupervised learning capability. This system subsequently operates successfully with a fast monitoring time in a wide range of cutting conditions without a priori knowledge of the cutting process.

AN UNSUPERVISED NEURAL NETWORK APPROACH TO TOOL WEAR IDENTIFICATION

An unsupervised neural network approach is proposed for tool wear identification. Conventional pattern recognition approaches to automating the wear monitoring task are non-adaptive and require expensive or inaccessible information. Rangwala's application of the supervised backpropagation neural network to tool wear identification in a turning operation represented a pioneering effort to integrate sensor signals (cutting force and acoustic emission) and to employ a neural network in the classification of those signals. However, backpropagation also requires expensive training information and cannot remain adaptive after training. The unsupervised adaptive resonance network exhibited the ability to classify sensor signals into fresh and worn classes, to remain adaptive, and to utilize considerably less training information.

Chinese character classification using an adaptive resonance network

The ability to see through noise and distortion to a pattern is vital to the task of character recognition. Artificial neural networks exhibit such a capability as they are able to generalize automatically once they are trained. An application of an artificial neural network model, the Adaptive Resonance Theory (ART), to Chinese character classification is described. The ART classifier is used to classify 3755 Chinese characters. Our experimental results indicate that the classifier is able to achieve a high classification rate.

Invariant object recognition with the adaptive resonance (ART) network

The cloud-gap-filled (CGF) method developed by Hall et al. (2010) is a novel method for efficiently mitigating cloud obscuration using the most recently available cloud-free observations from prior days at each pixel. In this paper, we extend this method not only using prior observations but also using subsequent observations. Four CGF using 3 or 5 days prior or subsequent observations are applied to the standard MODerate resolution Imaging Spectroradiometer (MODIS) snow cover product and are evaluated against 3-years, in situ observations at 244 SNOwpack TELemetry (SNOTEL) stations. Results indicate that daily CGF snow cover maps using prior or subsequent observations are obviously different due to daily cloud shifting. However, the monthly and annual cloud reductions are similar. Although the overall accuracies under all-sky conditions of CGF using prior or subsequent observations are very similar, CGF snow cover maps using subsequent images have less underestimation errors, except during the snow melting period, and higher snow accuracies. When the observation interval is expanded from 3 to 5 days, this method is able to fill more cloud obscurations and increase the overall accuracies by ∼30% and ∼37%, respectively; however, they introduce slightly more uncertainty. Monthly and regional analyses indicate that this method is more efficient during the snow-covered period or in high-elevation zones.

Deuteron magnetic resonance and hydrogen bond network of ammonium trihydrogen selenite

The DMR spectra of single-crystal ND 4D 3(SeO 3) 2 have been studied. The principal values and the direction cosines of the field-gradient tensor of deuterons located on three nonequivalent O · · · O hydrogen bonds have been determined. The lengths of hydrogen bonds have been calculated from eQq h values; the deuterons have been located on hydrogen bonds. The comparison with the DMR data of isomorphous compound RbD 3(SeO 3) 2 was made; the influence of NH · · · O hydrogen bonds on the structural parameters O · · · O hydrogen bonds is discussed.

Deuteron magnetic resonance and hydrogen bond network of ammonium trihydrogen selenite: I. S. Vinogradova, L. V. Kirensky Institute of Physics, Academy of Sciences USSR, Siberian Branch, Krasnoyarsk, 660036, USSR

Deuteron magnetic resonance and hydrogen bond network of ammonium trihydrogen selenite: I. S. Vinogradova, L. V. Kirensky Institute of Physics, Academy of Sciences USSR, Siberian Branch, Krasnoyarsk, 660036, USSR