PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Neural Networks, Computer”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

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

At least 1,189 records · Page 66Linked to original sources
PubMed1995

alpha-Helix region prediction with stochastic rule learning.

We propose a new method, based on the theory of stochastic rule learning, for predicting alpha-helix regions in a given protein sequence. Our method (hereafter referred to as the SR method) produces stochastic rules, each of which assigns, to any region in an amino acid sequence, the probability that it is an alpha-helix region. When learning a stochastic rule from a particular alpha-helix region, our method makes use of positive training examples obtained from a number of regions that are homologous to that region. Each stochastic rule is optimized using the minimum description length (MDL) principle, and such optimized stochastic rules are used to predict alpha-helix regions of any given protein sequence. In our experiments, using 25 proteins selected from the HSSP database as training examples, we applied the SR method to the problem of predicting alpha-helix regions in test examples, which consisted of > 5000 residues with 38% alpha-helix content. Each of these test examples possesses < 25% homology to any proteins in the training and other test examples. Our method achieved 81% average prediction accuracy for the test examples; this compares favorably to Qian and Sejnowski's method, which attains no more than 75% average accuracy, and further which compares to Rost and Sander's method which has proven to be one of the best secondary structure prediction methods.

Amino Acid Sequence↗
PubMed2000

Assessing the accuracy of prediction algorithms for classification: an overview.

We provide a unified overview of methods that currently are widely used to assess the accuracy of prediction algorithms, from raw percentages, quadratic error measures and other distances, and correlation coefficients, and to information theoretic measures such as relative entropy and mutual information. We briefly discuss the advantages and disadvantages of each approach. For classification tasks, we derive new learning algorithms for the design of prediction systems by directly optimising the correlation coefficient. We observe and prove several results relating sensitivity and specificity of optimal systems. While the principles are general, we illustrate the applicability on specific problems such as protein secondary structure and signal peptide prediction.

Algorithms↗
PubMed2003

Evolving spike-timing-dependent plasticity for single-trial learning in robots.

Single-trial learning is studied in an evolved robot model of synaptic spike-timing-dependent plasticity (STDP). Robots must perform positive phototaxis but must learn to perform negative phototaxis in the presence of a short-lived aversive sound stimulus. STDP acts at the millisecond range and depends asymmetrically on the relative timing of pre- and post-synaptic spikes. Although it has been involved in learning models of input prediction, these models require the iterated presentation of the input pattern, and it is hard to see how this mechanism could sustain single-trial learning over a time-scale of tens of seconds. An incremental evolutionary approach is used to answer this question. The evolved robots succeed in learning the appropriate behaviour, but learning does not depend on achieving the right synaptic configuration but rather the right pattern of neural activity. Robot performance during positive phototaxis is quite robust to loss of spike-timing information, but in contrast, this loss is catastrophic for learning negative phototaxis where entrained firing is common. Tests show that the final weight configuration carries no information about whether a robot is performing one behaviour or the other. Fixing weights, however, has the effect of degrading performance, thus demonstrating that plasticity is used to sustain the neural activity corresponding both to the normal phototaxis condition and to the learned behaviour. The implications and limitations of this result are discussed.

Action Potentials↗
PubMed2003

Isotropic-sequence-order learning in a closed-loop behavioural system.

The simplest form of sensor-motor control is obtained with a reflex. In this case the reflex can be interpreted as part of a closed-loop control paradigm which measures a sensor input and generates a motor reaction as soon as the sensor signal deviates from its desired (resting) state. This is a typical case of feedback control. However, reflex reactions are tardy, because they occur always only after a (for example, unpleasant) reflex-eliciting sensor event. This defines an objective problem for an organism which can only be avoided if the corresponding motor reaction is generated earlier. The goal of this study is to design a closed-loop control situation where temporal-sequence learning supersedes a tardy reflex reaction with a proactive anticipatory action. We achieve this by employing a second, earlier-occurring and causally coupled sensor event. An appropriate motor reaction to this early event prevents triggering of the original, primary reflex. Such causally coupled sensor events are common for animals, for example when smell predicts taste or when heat radiation precedes pain. We show that trying to achieve anticipatory control is a fundamentally different goal from trying to model a classical conditioning paradigm, which is an open-loop condition. To this end, we use a novel learning rule for temporal-sequence learning called isotropic-sequence-order (ISO) learning, which performs a confounded correlation between the primary sensor signal associated to the reflex and a predictive, earlier-occurring sensor input: this way the system learns the relation between the primary reflex and the earlier sensor input in order to create an earlier-occurring motor reaction. As a consequence of learning, the primary reflex will not be triggered any more, thereby permanently remaining in its desired resting state. In a robot application, we demonstrate that ISO learning can successfully solve the classical obstacle-avoidance task by learning to correlate a built-in reflex behaviour (retraction after touching) with earlier arising signals from range finders (before touching). Finally, we show that avoidance and attraction tasks can be combined in the same agent.

Adaptation, Physiological↗
PubMed2003

Self-similar community structure in a network of human interactions.

We propose a procedure for analyzing and characterizing complex networks. We apply this to the social network as constructed from email communications within a medium sized university with about 1700 employees. Email networks provide an accurate and nonintrusive description of the flow of information within human organizations. Our results reveal the self-organization of the network into a state where the distribution of community sizes is self-similar. This suggests that a universal mechanism, responsible for emergence of scaling in other self-organized complex systems, as, for instance, river networks, could also be the underlying driving force in the formation and evolution of social networks.

Algorithms↗
PubMed2000

Hadamard-based image decomposition and compression.

In this paper, we develop a general algorithm for decomposition and compression of grayscale images. The decomposition can be expressed as a functional relation between the original image and the Hadamard waveforms. The dynamic adaptive clustering procedure incorporates potential functions as a similarity measure for clustering as well as a reclustering phase. The latter is a multi-iteration, convergent procedure which divides the inputs into nonoverlapping clusters. These two techniques allow us to efficiently store and transmit a class of half-tone medical images such as magnetic resonance imaging (MRI) of the human brain. Due to the redundant image structure of MRI, obtained after the decomposition and clustering, almost half of the image can be omitted all together. Naturally, the compression rates for this specific type of grayscale image are increased greatly. A run-length coding is performed in order to compress further the retained information from the first two steps. Although all the techniques applied are simple, they represent an efficient way to compress grayscale images. The algorithm exhibits a performance which is competitive and often outperforming some of the methods reported in the literature.

Algorithms↗
PubMed2005

Toward the neurocomputer: image processing and pattern recognition with neuronal cultures.

Information processing in the nervous system is based on parallel computation, adaptation and learning. These features cannot be easily implemented on conventional silicon devices. In order to obtain a better insight of how neurons process information, we have explored the possibility of using biological neurons as parallel and adaptable computing elements for image processing and pattern recognition. Commercially available multielectrode arrays (MEAs) were used to record and stimulate the electrical activity from neuronal cultures. By mapping digital images, i.e., arrays of pixels, into the stimulation of neuronal cultures, a low and bandpass filtering of images could be quickly and easily obtained. Responses to specific spatial patterns of stimulation were potentiated by an appropriate training (tetanization). Learning allowed pattern recognition and extraction of spatial features in processed images. Therefore, neurocomputers, (i.e., hybrid devices containing man-made elements and natural neurons) seem feasible and may become a new generation of computing devices, to be developed by a synergy of Neuroscience and Material Science.

Animals↗
PubMed2005

Uncertainty of data, fuzzy membership functions, and multilayer perceptrons.

Probability that a crisp logical rule applied to imprecise input data is true may be computed using fuzzy membership function (MF). All reasonable assumptions about input uncertainty distributions lead to MFs of sigmoidal shape. Convolution of several inputs with uniform uncertainty leads to bell-shaped Gaussian-like uncertainty functions. Relations between input uncertainties and fuzzy rules are systematically explored and several new types of MFs discovered. Multilayered perceptron (MLP) networks are shown to be a particular implementation of hierarchical sets of fuzzy threshold logic rules based on sigmoidal MFs. They are equivalent to crisp logical networks applied to input data with uncertainty. Leaving fuzziness on the input side makes the networks or the rule systems easier to understand. Practical applications of these ideas are presented for analysis of questionnaire data and gene expression data.

Algorithms↗
PubMed2005

A performance analysis of two approximate adaptive designs.

The performance of function approximator based adaptive control designs may scale badly with approximator dimension. For a simple system class, both projection based designs and multiresolution approximation based designs have been shown to have good scaling properties with respect to to linear quadratic (LQ) costs. Here we show that by considering a cost functional with penalties on the control rate, the multiresolution approximator based design can outperform the projection based design. Generalizations are briefly discussed.

Algorithms↗
PubMed2005

Designing asymmetric Hopfield-type associative memory with higher order hamming stability.

The problem of optimal asymmetric Hopfield-type associative memory (HAM) design based on perceptron-type learning algorithms is considered. It is found that most of the existing methods considered the design problem as either 1) finding optimal hyperplanes according to normal distance from the prototype vectors to the hyperplane surface or 2) obtaining weight matrix W = [w(ij)] by solving a constraint optimization problem. In this paper, we show that since the state space of the HAM consists of only bipolar patterns, i.e., V = (v1, v2, . . ., vN)T E {-1, +1}N, the basins of attraction around each prototype (training) vector should be expanded by using Hamming distance measure. For this reason, in this paper, the design problem is considered from a different point of view. Our idea is to systematically increase the size of the training set according to the desired basin of attraction around each prototype vector. We name this concept the higher order Hamming stability and show that conventional minimum-overlap algorithm can be modified to incorporate this concept. Experimental results show that the recall capability as well as the number of spurious memories are all improved by using the proposed method. Moreover, it is well known that setting all self-connections wiiVi to zero has the effect of reducing the number of spurious memories in state space. From the experimental results, we find that the basin width around each prototype vector can be enlarged by allowing nonzero diagonal elements on learning of the weight matrix W. If the magnitude of w(ii) is small for all i, then the condition w(ii) = OVi can be relaxed without seriously affecting the number of spurious memories in the state space. Therefore, the method proposed in this paper can be used to increase the basin width around each prototype vector with the cost of slightly increasing the number of spurious memories in the state space.

Algorithms↗
PubMed2004

An adaptive high-order neural tree for pattern recognition.

A new neural tree model, called adaptive high-order neural tree (AHNT), is proposed for classifying large sets of multidimensional patterns. The AHNT is built by recursively dividing the training set into subsets and by assigning each subset to a different child node. Each node is composed of a high-order perceptron (HOP) whose order is automatically tuned taking into account the complexity of the pattern set reaching that node. First-order nodes divide the input space with hyperplanes, while HOPs divide the input space arbitrarily, but at the expense of increased complexity. Experimental results demonstrate that the AHNT generalizes better than trees with homogeneous nodes, produces small trees and avoids the use of complex comparative statistical tests and/or a priori selection of large parameter sets.

Algorithms↗
PubMedDate not supplied

Evaluation of video gray-scale display.

Setting up and maintaining video display monitors properly will help to reduce display variation and improve overall presentation of the radiological image. Display monitor gray-scale characteristics were examined using the SMPTE test pattern. This test pattern may be used as a standard for adjusting brightness and contrast. The controls should be adjusted to display the full dynamic range so that the 5% and 95% signal levels in the pattern are visible. Measured luminance on a laboratory workstation used for radiological perceptual experiments, and on the Siemens CT gray-scale monitor was determined to range from 0.17 to 76.0 nit, and 0.17 to 24.66 nit, respectively. These were compared with the range of approximately 17 to 514 nit for a typical film-viewbox combination. Characteristic curves were determined for both monitors, and CRT gammas were 3.34 and 2.48 for the perceptual workstation and CT console, respectively. The display gamma was determined from fitting luminance data to a log-log plot of luminance versus input gray level. The usefulness of the SMPTE test pattern for visual presentation as well as photometric measurement is demonstrated.

Computer Terminals↗
PubMed1995

A computer model of dorsal cochlear nucleus pyramidal cells: intrinsic membrane properties.

Manis [P. B. Manis, J. Neurosci. 10, 2338-2351 (1990)] studied "simple spiking," pyramidal cells of the dorsal cochlear nucleus (DCN) maintained in vitro. Response profiles to hyperpolarizing and depolarizing current pulses were generated. Hyperpolarization of the cell membrane followed by depolarization produced markedly different response profiles from those generated when no prehyperpolarization was imposed. By manipulating the magnitude of the hyperpolarizing and depolarizing pulses, "chopper," "pauser" and "build-up" response patterns, similar to those in vivo, could be generated by individual cells. Manis concluded that the different response profiles resulted from the modulation of intrinsic membrane conductances by the prehyperpolarizing pulses. Here a computer model is used to show that (a) steady-state hyperpolarization can influence cell responding to subsequent depolarization in a manner consistent with the data reported by Manis; and (b) the effects reported can be generated by the addition of a modeled transient potassium conductance to the standard Hodgkin-Huxley model of spike generation [A. L. Hodgkin and A. F. Huxley, J. Physiol. 117, 500-544 (1952)]. The model will be of use to those who wish to consider the role of various excitatory and inhibitory inputs to pyramidal cells and to establish their functional role within the DCN.

Cochlear Nucleus↗
PubMed2004

Nonlinear system modelling via optimal design of neural trees.

This paper introduces a flexible neural tree model. The model is computed as a flexible multi-layer feed-forward neural network. A hybrid learning/evolutionary approach to automatically optimize the neural tree model is also proposed. The approach includes a modified probabilistic incremental program evolution algorithm (MPIPE) to evolve and determine a optimal structure of the neural tree and a parameter learning algorithm to optimize the free parameters embedded in the neural tree. The performance and effectiveness of the proposed method are evaluated using function approximation, time series prediction and system identification problems and compared with the related methods.

Algorithms↗