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 811 records · Page 45Linked to original sources
PubMed2001

Learning from noisy information in FasArt and FasBack neuro-fuzzy systems.

Neuro-fuzzy systems have been in the focus of recent research as a solution to jointly exploit the main features of fuzzy logic systems and neural networks. Within the application literature, neuro-fuzzy systems can be found as methods for function identification. This approach is supported by theorems that guarantee the possibility of representing arbitrary functions by fuzzy systems. However, due to the fact that real data are often noisy, generation of accurate identifiers is presented as an important problem. Within the Adaptive Resonance Theory (ART), PROBART architecture has been proposed as a solution to this problem. After a detailed comparison of these architectures based on their design principles, the FasArt and FasBack models are proposed. They are neuro-fuzzy identifiers that offer a dual interpretation, as fuzzy logic systems or neural networks. FasArt and FasBack can be trained on noisy data without need of change in their structure or data preprocessing. In the simulation work, a comparative study is carried out on the performances of Fuzzy ARTMAP, PROBART, FasArt and FasBack, focusing on prediction error and network complexity. Results show that FasArt and FasBack clearly enhance the performance of other models in this important problem.

Artifacts↗
PubMed1997

Artificial neural networks applied to outcome prediction for colorectal cancer patients in separate institutions.

BACKGROUND: Artificial neural networks are computer programs that can be used to discover complex relations within data sets. They permit the recognition of patterns in complex biological data sets that cannot be detected with conventional linear statistical analysis. One such complex problem is the prediction of outcome for individual patients treated for colorectal cancer. Predictions of outcome in such patients have traditionally been based on population statistics. However, these predictions have little meaning for the individual patient. We report the training of neural networks to predict outcome for individual patients from one institution and their predictive performance on data from a different institution in another region. METHODS: 5-year follow-up data from 334 patients treated for colorectal cancer were used to train and validate six neural networks designed for the prediction of death within 9, 12, 15, 18, 21, and 24 months. The previously trained 12-month neural network was then applied to 2-year follow-up data from patients from a second institution; outcome was concealed. No further training of the neural network was undertaken. The network's predictions were compared with those of two consultant colorectal surgeons supplied with the same data. FINDINGS: All six neural networks were able to achieve overall accuracy greater than 80% for the prediction of death for individual patients at institution 1 within 9, 12, 15, 18, 21, and 24 months. The mean sensitivity and specificity were 60% and 88%. When the neural network trained to predict death within 12 months was applied to data from the second institution, overall accuracy of 90% (95% CI 84-96) was achieved, compared with the overall accuracy of the colorectal surgeons of 79% (71-87) and 75% (66-84). INTERPRETATION: The neural networks were able to predict outcome for individual patients with colorectal cancer much more accurately than the currently available clinicopathological methods. Once trained on data from one institution, the neural networks were able to predict outcome for patients from an unrelated institution.

Bias↗
PubMedDate not supplied

On the use of neural network ensembles in QSAR and QSPR.

Despite their growing popularity among neural network practitioners, ensemble methods have not been widely adopted in structure-activity and structure-property correlation. Neural networks are inherently unstable, in that small changes in the training set and/or training parameters can lead to large changes in their generalization performance. Recent research has shown that by capitalizing on the diversity of the individual models, ensemble techniques can minimize uncertainty and produce more stable and accurate predictors. In this work, we present a critical assessment of the most common ensemble technique known as bootstrap aggregation, or bagging, as applied to QSAR and QSPR. Although aggregation does offer definitive advantages, we demonstrate that bagging may not be the best possible choice and that simpler techniques such as retraining with the full sample can often produce superior results. These findings are rationalized using Krogh and Vedelsby's decomposition of the generalization error into a term that measures the average generalization performance of the individual networks and a term that measures the diversity among them. For networks that are designed to resist over-fitting, the benefits of aggregation are clear but not overwhelming.

Computer Simulation↗
PubMed1996

Empirical error-confidence curves for neural network and Gaussian classifiers.

"Error-Confidence" measures the probability that the proportion of errors made by a classifier will be within epsilon of EB, the optimal (Bayes) error. Probably Almost Bayes (PAB) theory attempts to quantify how this confidence increases with the number of training samples. We investigate the relationship empirically by comparing average error versus number of training patterns (m) for linear and neural network classifiers. On Gaussian problems, the resulting EC curves demonstrate that the PAB bounds are extremely conservative. Asymptotic statistics predicts a linear relationship between the logarithms of the average error and the number of training patterns. For low Bayes error rates we found excellent agreement between the prediction and the linear discriminant performance. At higher Bayes error rates we still found a linear relationship, but with a shallower slope than the predicted-1. When the underlying true model is a three-layer network, the EC curves show a greater dependence on classifier capacity, and the linear predictions no longer seem to hold.

Bayes Theorem↗
PubMed1994

The influence of limited presynaptic growth and synapse removal on adaptive synaptogenesis.

This report continues our research into the effectiveness of adaptive synaptogenesis in constructing feed-forward networks which perform good transformations on their inputs. Good transformations are characterized by the maintenance of input information and the removal of statistical dependence. Adaptive synaptogenesis stochastically builds and sculpts a synaptic connectivity in initially unconnected networks using two mechanisms. The first, synaptogenesis, creates new, excitatory, feed-forward connections. The second, associative modification, adjusts the strength of existing synapses. Our previous implementations of synaptogenesis only incorporated a postsynaptic regulatory process, receptivity to new innervation (Adelsberger-Mangan and Levy 1993a, b). In the present study, a presynaptic regulatory process, presynaptic avidity, which regulates the tendency of a presynaptic neuron to participate in a new synaptic connection as a function of its total synaptic weight, is incorporated into the synaptogenesis process. In addition, we investigate a third mechanism, selective synapse removal. This process removes synapses between neurons whose firing is poorly correlated. Networks that are constructed with the presynaptic regulatory process maintain more information and remove more statistical dependence than networks constructed with postsynaptic receptivity and associative modification alone. Selective synapse removal also improves network performance, but only when implemented in conjunction with the presynaptic regulatory process.

Computer Simulation↗
PubMed2006

Parameter space structure of continuous-time recurrent neural networks.

A fundamental challenge for any general theory of neural circuits is how to characterize the structure of the space of all possible circuits over a given model neuron. As a first step in this direction, this letter begins a systematic study of the global parameter space structure of continuous-time recurrent neural networks (CTRNNs), a class of neural models that is simple but dynamically universal. First, we explicitly compute the local bifurcation manifolds of CTRNNs. We then visualize the structure of these manifolds in net input space for small circuits. These visualizations reveal a set of extremal saddle node bifurcation manifolds that divide CTRNN parameter space into regions of dynamics with different effective dimensionality. Next, we completely characterize the combinatorics and geometry of an asymptotically exact approximation to these regions for circuits of arbitrary size. Finally, we show how these regions can be used to calculate estimates of the probability of encountering different kinds of dynamics in CTRNN parameter space.

Animals↗
PubMed1998

On numerical simulations of integrate-and-fire neural networks.

It is shown that very small time steps are required to reproduce correctly the synchronization properties of large networks of integrate-and-fire neurons when the differential system describing their dynamics is integrated with the standard Euler or second-order Runge-Kutta algorithms. The reason for that behavior is analyzed, and a simple improvement of these algorithms is proposed.

Action Potentials↗
PubMedDate not supplied

eLoom and Flatland: specification, simulation and visualization engines for the study of arbitrary hierarchical neural architectures.

eLoom is an open source graph simulation software tool, developed at the University of New Mexico (UNM), that enables users to specify and simulate neural network models. Its specification language and libraries enables users to construct and simulate arbitrary, potentially hierarchical network structures on serial and parallel processing systems. In addition, eLoom is integrated with UNM's Flatland, an open source virtual environments development tool to provide real-time visualizations of the network structure and activity. Visualization is a useful method for understanding both learning and computation in artificial neural networks. Through 3D animated pictorially representations of the state and flow of information in the network, a better understanding of network functionality is achieved. ART-1, LAPART-II, MLP, and SOM neural networks are presented to illustrate eLoom and Flatland's capabilities.

Computer Simulation↗
PubMed1994

Use of artificial neural networks within deterministic logic for the computer ECG diagnosis of inferior myocardial infarction.

An investigation into the use of software-based artificial neural networks for the electrocardiographic (ECG) detection of inferior myocardial infarction was made. A total of 592 clinically validated subjects, including 208 with inferior myocardial infarction, 300 normal subjects, and 84 left ventricular hypertrophy cases, were used in this study. A total of 200 ECGs (100 from patients with inferior myocardial infarction and 100 from normal subjects) were fed to 66 supervised feedforward neural networks for training using a back-propagation algorithm. QRS and ST-T wave measurements were used as the input parameters for the neural networks. The best performing network using QRS measurements only and the best using QRS and ST-T data were selected by assessing a test set of 292 ECGs (108 from patients with inferior myocardial infarction, 84 from patients with left ventricular hypertrophy, and 100 from normal subjects). These two networks were then implanted separately into the deterministic Glasgow program for further study. After the implementation, it was found necessary to include a small inferior Q criterion to improve the specificity of reporting inferior myocardial infarction, thereby producing a small loss of sensitivity as compared with use of the network alone. The use of an artificial neural network within the deterministic logic performed better than either alone in the diagnosis of inferior myocardial infarction, producing a 20% gain in sensitivity with 2% loss in overall specificity compared with the original deterministic logic.

Diagnosis, Computer-Assisted↗
PubMed2005

Intelligent initialization of resource allocating RBF networks.

In any neural network system, proper parameter initialization reduces training time and effort, and generally leads to compact modeling of the process under examination, i.e. less complex network structures and better generalization. However, in cases of multi-dimensional data, parameter initialization is both difficult and time consuming. In the proposed scheme a novel, multi-dimensional, unsupervised clustering method is used to properly initialize neural network architectures, focusing on resource allocating networks (RAN); both the hidden and output layer parameters are determined by the output of the clustering process, without the need for any user interference. The main contribution of this work is that the proposed approach leads to network structures that are compact, efficient and achieve best classification results, without the need for manual selection of suitable initial network parameters. The efficiency of the proposed method has been tested on several classes of publicly available data, such as iris, Wisconsin and ionosphere data.

Algorithms↗
PubMed2005

A primal-dual neural network for online resolving constrained kinematic redundancy in robot motion control.

This paper proposes a primal-dual neural network with a one-layer structure for online resolution of constrained kinematic redundancy in robot motion control. Unlike the Lagrangian network, the proposed neural network can handle physical constraints, such as joint limits and joint velocity limits. Compared with the existing primal-dual neural network, the proposed neural network has a low complexity for implementation. Compared with the existing dual neural network, the proposed neural network has no computation of matrix inversion. More importantly, the proposed neural network is theoretically proved to have not only a finite time convergence, but also an exponential convergence rate without any additional assumption. Simulation results show that the proposed neural network has a faster convergence rate than the dual neural network in effectively tracking for the motion control of kinematically redundant manipulators.

Algorithms↗
PubMed1995

An artificial neural system for closed loop control of locomotion produced via neuromuscular electrical stimulation.

The use of neuromuscular electrical stimulation for restoration of gait in spinal cord injured subjects has been seriously pursued by many investigators for the past 15 years. By and large, however, systems to date require the intervention of a person, be it the patient or an observer, and are restricted to control of stimulation onset and termination. Further, existing systems are not adaptable to environmental and patient variations. This work proposes a system that relies on neural computing to determine proper muscle activation patterns from biomechanical signals. The intelligent system is trained to perform gait under supervision, after which it can be used to control muscle stimulation in an unknown environment. Computer simulations suggest that the best neural architecture for control of gait is a neural network including units corresponding to movement history. Separate networks for the stance and swing phases, respectively, were found to work better than a single neural network trained on the entire gait cycle. The artificial neural device proposed here also includes a voice recognition system that will allow for voluntary locomotion. A safety circuit has been designed to preclude acceptance of unwanted vocal commands in the latter system.

Computer Simulation↗
PubMed2002

Automatic recognition of cell layers in corneal confocal microscopy images.

A confocal microscope can produce gray-scale images of the different layers of the cornea. We have addressed the problem of classifying these images, i.e. recognizing the layer displayed, using the shape of the cells contained, which is uniquely related to each specific layer. A first method was designed, based first on the binarization of the image and then on the description of the cell shape by means of Hu variables (central moments). An artificial neural network was used to classify each image according to the values assumed by these variables. A Matlab prototype of the classification system was developed, considering images of three corneal layers (Bowman membrane, stroma, endothelium) in normal subjects. The system was tested on 46 images, and good results were obtained. To avoid the critical step of binarization, an alternative cell shape description was investigated, based on Zernike moments, and a new network was developed and trained. The results achieved were better than those obtained with the previous technique, and also, no binarization was necessary.

Cell Size↗
PubMed2006

Globally asymptotic stability of a class of neutral-type neural networks with delays.

Several stability conditions for a class of systems with retarded-type delays are presented in the literature. However, no results have yet been presented for neural networks with neutral-type delays. Accordingly, this correspondence investigates the globally asymptotic stability of a class of neutral-type neural networks with delays. This class of systems includes Hopfield neural networks, cellular neural networks, and Cohen-Grossberg neural networks. Based on the Lyapunov stability method, two delay-independent sufficient stability conditions are derived. These stability conditions are easily checked and can be derived from the connection matrix and the network parameters without the requirement for any assumptions regarding the symmetry of the interconnections. Two illustrative examples are presented to demonstrate the validity of the proposed stability criteria.

Algorithms↗
PubMed2000

Computer-assisted visualizations of neural networks: expanding the field of view using seamless confocal montaging.

Microscopic analysis of anatomic relationships within the neural networks of adult and developing tissues often requires sampling large spatial regions of neuronal architecture. To accomplish this, there are two common imaging approaches: (1) image the entire area at once with low spatial resolution; or (2) image small sections at higher magnification/resolution and then join the sections back together by mosaic reconstruction (photomontaging). Low magnification imaging is relatively rapid to perform, resulting in a visualization that encompasses a large field of view with an extended depth of field. However, for fluorescence microscopy, low magnification visualizations are often plagued by poor spatial resolution. High magnification imaging possesses superior spatial resolution, but it produces an image with limited depth of field. When creating a larger field of view, the final image is also fragmented at the boundaries where multiple images are stitched together. Using confocal microscopy as well as features of common image processing programs, we outline a new method to transform individual, spatially contiguous z-series into a montage with a seamless field of view and an extended depth of field. In addition, we show that the manual alignment of images our method requires does not introduce significant errors into the final image. We illustrate our method for visualizing neural networks using tissues from the adult gastropod mollusc, Tritonia diomedea, and the developing zebrafish, Danio rerio.

Animals↗
PubMed1997

Basic concepts of artificial neural networks (ANN) modeling in the application to pharmaceutical development.

Artificial neural networks (ANN) methodology is a new modeling method that has not been broadly applied to pharmaceutical sciences up to now. The aim of this paper is to give a detailed description of the associating networks as well as a description of less well-known networks (i.e., feature-extracting and nonadaptive networks) and their scope of application in pharmaceutical sciences. The descriptions include the historical origin and the basic concepts behind the computing. ANN are based on the attempt to model the neural networks of the brain. Learning algorithms for associating ANN use mathematical procedures usually derived from the gradient descent method whereas feature-extracting ANN map multidimensional input data sets onto two-dimensional spaces. Nonadaptive ANN map data sets and are able to reconstruct their patterns when presented with corrupted or noisy samples. Associating networks can typically be applied in the pharmaceutical field as an alternative to traditional response surface methodology, feature-extracting networks as alternative to principal component analysis, and nonadaptive networks for image recognition. Based on these abilities, the potential application fields of the ANN methodology in the pharmaceutical sciences is broad, ranging from clinical pharmacy through biopharmacy, drug and dosage form design, to interpretation of analytical data. The few applications presented in the pharmaceutical technology area seem promising and should be investigated in more detail.

Algorithms↗
PubMed2006

Analog-digital simulations of full conductance-based networks of spiking neurons with spike timing dependent plasticity.

We introduce and test a system for simulating networks of conductance-based neuron models using analog circuits. At the single-cell level, we use custom-designed analog circuits (ASICs) that simulate two types of spiking neurons based on Hodgkin-Huxley like dynamics: "regular spiking" excitatory neurons with spike-frequency adaptation, and "fast spiking" inhibitory neurons. Synaptic interactions are mediated by conductance-based synaptic currents described by kinetic models. Connectivity and plasticity rules are implemented digitally through a real time interface between a computer and a PCI board containing the ASICs. We show a prototype system of a few neurons interconnected with synapses undergoing spike-timing dependent plasticity (STDP), and compare this system with numerical simulations. We use this system to evaluate the effect of parameter dispersion on the behavior of small circuits of neurons. It is shown that, although the exact spike timings are not precisely emulated by the ASIC neurons, the behavior of small networks with STDP matches that of numerical simulations. Thus, this mixed analog-digital architecture provides a valuable tool for real-time simulations of networks of neurons with STDP. They should be useful for any real-time application, such as hybrid systems interfacing network models with biological neurons.

Action Potentials↗