Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Neural network”

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 1,261 records · Page 70Linked to original sources
PubMed2006

Parameter incremental learning algorithm for neural networks.

In this paper, a novel stochastic (or online) training algorithm for neural networks, named parameter incremental learning (PIL) algorithm, is proposed and developed. The main idea of the PIL strategy is that the learning algorithm should not only adapt to the newly presented input-output training pattern by adjusting parameters, but also preserve the prior results. A general PIL algorithm for feedforward neural networks is accordingly presented as the first-order approximate solution to an optimization problem, where the performance index is the combination of proper measures of preservation and adaptation. The PIL algorithms for the multilayer perceptron (MLP) are subsequently derived. Numerical studies show that for all the three benchmark problems used in this paper the PIL algorithm for MLP is measurably superior to the standard online backpropagation (BP) algorithm and the stochastic diagonal Levenberg-Marquardt (SDLM) algorithm in terms of the convergence speed and accuracy. Other appealing features of the PIL algorithm are that it is computationally as simple as the BP algorithm, and as easy to use as the BP algorithm. It, therefore, can be applied, with better performance, to any situations where the standard online BP algorithm is applicable.

Algorithms↗
PubMed2004

Robust redesign of a neural network controller in the presence of unmodeled dynamics.

This paper presents a neural network control redesign, which achieves robust stabilization in the presence of unmodeled dynamics restricted to be input to output practically stable (IOpS), without requiring any prior knowledge on any bounding function. Moreover, the state of the unmodeled dynamics is permitted to go unbounded provided that the nominal system state and/or the control input also go unbounded. The neural network controller is equipped with a resetting strategy to deal with the problem of possible division by zero, which may appear since we consider unknown input vector fields with unknown signs. The uniform ultimate boundedness of the system output to an arbitrarily small set, plus the boundedness of all other signals in the closed-loop is guaranteed.

Algorithms↗
PubMed2005

Are artificial neural networks white boxes?

In this paper, we introduce a novel Mamdani-type fuzzy model, referred to as the all-permutations fuzzy rule base (APFRB), and show that it is mathematically equivalent to a standard feedforward neural network. We describe several applications of this equivalence between a neural network and our fuzzy rule base (FRB), including knowledge extraction from and knowledge insertion into neural networks.

Algorithms↗
PubMed2002

Artificial neural network method for predicting protein secondary structure content.

In this paper, the neural network method was applied to predict the content of protein secondary structure elements that was based on 'pair-coupled amino acid composition', in which the sequence coupling effects are explicitly included through a series of conditional probability elements. The prediction was examined by a self-consistency test and an independent-dataset. Both indicated good results obtained when using the neural network method to predict the contents of alpha-helix, beta-sheet, parallel beta-sheet strand, antiparallel beta-sheet strand, beta-bridge, 3(10)-helix, pi-helix, H-bonded turn, bend, and random coil.

Neural Networks, Computer↗
PubMed1989

Neural networks--an artificial intelligence approach to the analysis of clinical data.

This paper describes an approach to the pattern recognition problem of categorizing evoked response waveforms, using that branch of artificial intelligence known as neural networking. A total of 19 subjects (38 eyes) were exposed to a reversing chequerboard pattern, and the resulting Pattern electroretinogram (PERG) recorded. This was repeated for each of the subjects at contrast levels of 75%,50% and 25%, with the resulting response becoming less pronounced as the contrast level was decreased. The subjects were also given an Arden contrast sensitivity test, and the resulting scores recorded for each eye. The averaged responses from each of the 3 contrast levels for each of the 38 eyes tested were used as the input to a computer simulated 3-input per neuron cascaded neural network. The output and weighting parameters for the individual neurons were then varied until the network was able to optimally classify the PERG results on the basis of the Arden contrast sensitivity scores. The neural network was thus taught to recognize patterns in the PERG responses in much the same way that an experienced clinician might. A second neural network was similarly taught to classify the PERG responses on the basis of a clinician's subjective rating of each of the responses.

Adult↗
PubMed1998

Identification, localization, and modulation of neural networks for walking in the mudpuppy (Necturus maculatus) spinal cord.

We tested the hypothesis that the neural networks for walking in the mudpuppy can be divided into a flexor and an extensor center, each of which contains collections of interneurons localized in the vicinity of their motoneuron pools. Combining a battery of techniques, we identified and localized the elbow flexor center and its motoneuron pool in the C2 segment and the elbow extensor center and its motoneuron pool in the C3 segment. Rhythmic flexion or extension of the limb in isolation could be induced by continuous trains of current pulses of the C2 or C3 segments, respectively. Independent activation could also occur after application of glutamate receptor agonist NMDA. Part of segment C2 in isolation generated rhythmic elbow flexor bursts, whereas part of segment C3 in isolation generated rhythmic elbow extensor bursts. An isolated region spanning the C3 roots generated both flexor and extensor bursts. The step cycle was modulated in a phase-dependent manner by stimulation of the dorsal roots, the ventral roots, or either of the two centers. The effects of ventral root stimulation were removed by deafferentation to block reafferent input attributable to muscle contraction induced by the stimulation. We conclude that the neural networks for walking contain at least a flexor and an extensor generator that are localized in close apposition to the motoneuron pools, that the two centers can work independently despite the fact that there are reciprocal inhibitory interconnections between them, and that sensory input interacts with the spinal neural networks to reset the ongoing walking rhythm in a phase-dependent manner.

Animals↗
PubMed2003

Neural network and multiple regression models for PM10 prediction in Athens: a comparative assessment.

Particulate atmospheric pollution in urban areas is considered to have significant impact on human health. Therefore, the ability to make accurate predictions of particulate ambient concentrations is important to improve public awareness and air quality management. This study examines the possibility of using neural network methods as tools for daily average particulate matter with aerodynamic diameter <10 microm (PM10) concentration forecasting, providing an alternative to statistical models widely used up to this day. Based on a data inventory, in a fixed central site in Athens, Greece, ranging over a two-year period, and using mainly meteorological variables as inputs, neural network models and multiple linear regression models were developed and evaluated. Comparison statistics used indicate that the neural network approach has an edge over regression models, expressed both in terms of prediction error (root mean square error values lower by 8.2-9.4%) and of episodic prediction ability (false alarm rate values lower by 7-13%). The results demonstrate that artificial neural networks (ANNs), if properly trained and formed, can provide adequate solutions to particulate pollution prognostic demands.

Air Pollutants↗
PubMed1995

Ovarian dysplasia in epithelial inclusion cysts. A morphometric approach using neural networks.

BACKGROUND: Ovarian dysplasia has been described in the ovarian surface epithelium by histologic and morphometric studies. This study evaluates ovarian dysplasia in epithelial inclusion cysts adjacent to overt carcinoma and also incidentally found in ovaries removed for nonneoplastic diseases, including oophorectomies for family history of ovarian cancer, using an artificial neural network. METHODS: Histologic sections from 37 ovaries of which 26 were diagnosed with dysplasia in epithelial inclusion cysts (10 adjacent to carcinoma and 16 incidental) and 11 with benign epithelial inclusion cysts were evaluated by tracing nuclear profiles and assessing measures of nuclear area, shape, and texture. These sections were analyzed using artificial neural networks and also statistically using the Kruskal-Wallis test with the Dunn procedure to compare the morphologic similarity of dysplasia found incidentally in inclusion cysts unrelated to carcinoma from that in inclusion cysts adjacent to carcinoma. RESULTS: Neither statistical nor artificial neural network analysis was able to distinguish between incidental and adjacent dysplasia. Both types differed significantly from the control cases. CONCLUSIONS: Neural networks are powerful classification tools when applied to multiple variables extracted from individual cases. In this study, they helped to substantiate the similarity between dysplasia found incidentally and that adjacent to ovarian carcinoma. Because dysplasia represents a potential precancerous lesion, its incidental finding may help identify patients at risk for developing ovarian carcinoma.

Female↗
PubMed2005

Real-time ocular artifact suppression using recurrent neural network for electro-encephalogram based brain-computer interface.

The paper presents an adaptive noise canceller (ANC) filter using an artificial neural network for real-time removal of electro-oculogram (EOG) interference from electro-encephalogram (EEG) signals. Conventional ANC filters are based on linear models of interference. Such linear models provide poorer prediction for biomedical signals. In this work, a recurrent neural network was employed for modelling the interference signals. The eye movement and eye blink artifacts were recorded by the placing of an electrode on the forehead above the left eye and an electrode on the left temple. The reference signal was then generated by the data collected from the forehead electrode being added to data recorded from the temple electrode. The reference signal was also contaminated by the EEG. To reduce the EEG interference, the reference signal was first low-pass filtered by a moving averaged filter and then applied to the ANC. Matlab Simulink was used for real-time data acquisition, filtering and ocular artifact suppression. Simulation results show the validity and effectiveness of the technique with different signal-to-noise ratios (SNRs) of the primary signal. On average, a significant improvement in SNR up to 27 dB was achieved with the recurrent neural network. The results from real data demonstrate that the proposed scheme removes ocular artifacts from contaminated EEG signals and is suitable for real-time and short-time EEG recordings.

Artifacts↗
PubMed2000

Detection of leukocytes in contact with the vessel wall from in vivo microscope recordings using a neural network.

Leukocytes play an important role in the host defense as they may travel from the blood stream into the tissue in reacting to inflammatory stimuli. The leukocyte-vessel wall interactions are studied in post capillary vessels by intravital video microscopy during in vivo animal experiments. Sequences of video images are obtained and digitized with a frame grabber. A method for automatic detection and characterization of leukocytes in the video images is developed. Individual leukocytes are detected using a neural network that is trained with synthetic leukocyte images generated using a novel stochastic model. This model makes it feasible to generate images of leukocytes with different shapes and sizes under various lighting conditions. Experiments indicate that neural networks trained with the synthetic leukocyte images perform better than networks trained with images of manually detected leukocytes. The best performing neural network trained with synthetic leukocyte images resulted in an 18% larger area under the ROC curve than the best performing neural network trained with manually detected leukocytes.

Animals↗
PubMed1999

[Artificial neural network application in spectral recognition].

An artificial neural network was devised and established for spectral matching. It was proved that when training sample spectra and unknown spectra had the same magnitude calibration, the network recognized spectra very well and had some advantages over other conventional spectral matching methods.

Computer Simulation↗
PubMed1980

A template matching model for pattern recognition: self-organization of templates and template matching by a disinhibitory neural network.

A template matching model for pattern recognition is proposed. By following a previously-proposed algorithm for synpatic modification (Hirai, 1980), the template of a stimulus pattern is self-organized as a spatial distribution pattern of matured synapses on the cells receiving modifiable synapses. Template matching is perfomed by the disinhibitory neural network cascaded beyond the neural layer composed of the cells receiving the modifiable synapses. The performance of the model has been simulated on a digital computer. After repetitive presentations of a stimulus pattern, a cell receiving the modifiable synapses comes to have the template of that pattern. And the cell in the latter layer of the disinhibitory neural network that receives the disinhibitory input from that cell becomes selectively sensitive to that pattern. Learning patterns are not restricted by previously learned ones. They can be subset or superset patterns of the ones previously learned. If an unknown pattern is presented to the model, no cell beyond the disinhibitory neural network will respond. However, if previously learned patterns are embedded in that pattern, the cells which have the templates of those patterns respond and are assumed to transmit the information to higher center. The computer simulation also show that the model can organize a clean template under a noisy environment.

Animals↗
PubMed1997

Recognition of rotating images using an automatic feature extraction technique and neural networks.

This paper presents a new automatic feature extraction technique and a neural network based classification method for recognition of rotating images. The image processing technique extracts global features of an image and converts a large size image into a one-dimensional small vector. A special advantage of the proposed technique is that the extracted features are the same even if the original image is rotated with rotation angles from 5 to 355 or rotated and a little bit distorted. The proposed technique is based on simple co-ordinate geometry, fuzzy sets and neural networks. The proposed approach is very easy in implementation and it has been developed in C++ on a Sun workstation. The experimental results have demonstrated that the proposed approach performs successfully on a variety of small as well as large scale rotated and distorted images.

Algorithms↗
PubMed1995

Assessment of neural network models using prediction analysis.

Examination of classification or confusion matrices in neural network models is often advocated as a way to investigate the effects of network architecture on classification rules and to understand the topography of the classification space. In addition, for model comparison purposes it is useful to be able to make statements concerning the statistical reliability of these patterns of hit and error cells. Prediction analysis provides a number of descriptive and inferential measures for the evaluation of hypotheses about patterns of hit and error cells. The present paper applied recent advances in prediction analysis to the examination of neural network classification matrices. Results of the application of prediction analysis to the generalized XOR problem and to data from classification of cardiovascular responses of human subjects were used to illustrate the determination of absolute and relative fit of omnibus models as well as the evaluation of specific hypotheses within these models.

Cardiovascular System↗
PubMed2001

Comparison of artificial neural networks with other statistical approaches: results from medical data sets.

BACKGROUND: In recent years, considerable attention has been given to the development of sophisticated techniques for exploring data sets. One such class of techniques is artificial neural networks (ANNs). Artificial neural networks have many attractive theoretic properties, specifically, the ability to detect non predefined relations such as nonlinear effects and/or interactions. These theoretic advantages come at the cost of reduced interpretability of the model output. Many authors have analyzed the same data set, based on these factors, with both standard statistical methods (such as logistic or Cox regression) and ANN. METHODS: The goal of this work is to review the literature comparing the performance of ANN with standard statistical techniques when applied to medium to large data sets (sample size > 200 patients). A thorough literature search was performed, with specific criteria for a published comparison to be included in this review. RESULTS: In the 28 studies included in this review, ANN outperformed regression in 10 cases (36%), was outperformed by regression in 4 cases (14%), and the 2 methods had similar performance in the remaining 14 cases (50%). However, in the 8 largest studies (sample size > 5000), regression and ANN tied in 7 cases, with regression winning in the remaining case. In addition, there is some suggestion of publication bias. CONCLUSIONS: Neither method achieves the desired performance. Both methods should continue to be used and explored in a complementary manner. However, based on the available data, ANN should not replace standard statistical approaches as the method of choice for the classification of medical data.

Humans↗
PubMed2003

[Study on the application of artificial neural network on diabetes mellitus/insulin-glucose tolerance classification].

OBJECTIVE: To discuss the potential application of artificial neural network (ANN) on the epidemiological classification of disease. METHODS: Learning vector quantification neural network (LVQNN) and discriminate analysis were applied to data from epidemiological survey in a mine in 1996. RESULTS: The structure of LVQNN was 25-->13-->3. The total veracity rates was 96.98%, and 92.45% among the abnormal blood glucose individuals. Through stepwise discriminate analysis, the discriminate equations were established including 11 variables with a total veracity rate of 87.34%, but was 85.53% in the abnormal blood glucose individuals. Further analysis on 30 cases with missing values showed that the disagreement ratio of LVQ was 1/30, lower than that of discriminate analysis of 7/30. CONCLUSIONS: Compared to the conventional statistics method, LVQ not only showed better prediction precision, but could treat data with missing values satisfactorily plus it had no limit to the type or distribution of relevant data, thus provided a new powerful method to epidemiologic prediction.

Algorithms↗
PubMedDate not supplied

Recurrent fractal neural networks: a strategy for the exchange of local and global information processing in the brain.

The regulation of biological networks relies significantly on convergent feedback signaling loops that render a global output locally accessible. Ideally, the recurrent connectivity within these systems is self-organized by a time-dependent phase-locking mechanism. This study analyzes recurrent fractal neural networks (RFNNs), which utilize a self-similar or fractal branching structure of dendrites and downstream networks for phase-locking of reciprocal feedback loops: output from outer branch nodes of the network tree enters inner branch nodes of the dendritic tree in single neurons. This structural organization enables RFNNs to amplify re-entrant input by over-the-threshold signal summation from feedback loops with equivalent signal traveling times. The columnar organization of pyramidal neurons in the neocortical layers V and III is discussed as the structural substrate for this network architecture. RFNNs self-organize spike trains and render the entire neural network output accessible to the dendritic tree of each neuron within this network. As the result of a contraction mapping operation, the local dendritic input pattern contains a downscaled version of the network output coding structure. RFNNs perform robust, fractal data compression, thus coping with a limited number of feedback loops for signal transport in convergent neural networks. This property is discussed as a significant step toward the solution of a fundamental problem in neuroscience: how is neuronal computation in separate neurons and remote brain areas unified as an instance of experience in consciousness? RFNNs are promising candidates for engaging neural networks into a coherent activity and provide a strategy for the exchange of global and local information processing in the human brain, thereby ensuring the completeness of a transformation from neuronal computation into conscious experience.

Brain↗
PubMed1999

Nonlinear transfer function encodes synchronization in a neural network from the mammalian brain.

Synchronization is one of the mechanisms by which the brain encodes information. The observed synchronization of neuronal activity has, however, several levels of fluctuations, which presumably regulate local features of specific areas. This means that biological neural networks should have an intrinsic mechanism able to synchronize the neuronal activity but also to preserve the firing capability of individual cells. Here, we investigate the input-output relationship of a biological neural network from developing mammalian brain, i.e., the hippocampus. We show that the probability of occurrence of synchronous output activity (which consists in stereotyped population bursts recorded throughout the hippocampus) is encoded by a sigmoidal transfer function of the input frequency. Under this scope, low-frequency inputs will not produce any coherent output while high-frequency inputs will determine a synchronous pattern of output activity (population bursts). We analyze the effect of the network size (N) on the parameters of the transfer function (threshold and slope). We found that sigmoidal functions realistically simulate the synchronous output activity of hippocampal neural networks. This outcome is particularly important in the application of results from neural network models to neurobiology.

Animals↗