In the paper an agent system of evolving neural networks, being an
example of collective intelligence, is presented. A concept of decentralised evolutionary computation realised as an evolutionary multi-agent systém (EMAS) may help to avoid some of the shortcomings of classical evolutionary optimisation techniques. As an addition, several methods of managing such a collective intelligent systém are mentioned. General considerations are illustrated by a particular system dedicated to the problém of time-series prediction. Selected experimental results conclude the work.
This research investigates the way in which Neural Networks (NNs)
can be used to forecast insolvency. The research aimed at forecasting, oiie for three years in advance, a clear disclosure of tlie legal condition of insolvency in a jointstock company based in Italy and operating in the textile industry, throngh the analysis of the official balance sheet. The results refer to experiments which have been carried out for three years, concerning a sample of about 500 conipanies in the textile sector, their balance sheets for a three-year period (1990-1992) and the state of the art in the following two years (1993-1994). The research has pointed out how the problem of insolvency can be dealt with by using NNs in a satisfactory way and has proved useful to show how important the method of collecting data for a further analysis throngh NNs can be.
This research investigates the way in which Neural Networks (NNs)
can be used to forecast insolvency. The research aimed at forecasting, one for three years in advance, a clear disclosure of tlie legal condition of insolvency in a jointstock company based in Italy and operating in the textile industry, through the analysis of the official balance sheet. The results refer to experiments which have been carried out for three years, concerning a sample of abont 500 coinpanies” in the textile sector, their balance sheets for a three-year period (1990-1992) and the State of the art in the following two years (1993-1994). The research has pointed out how the problem of insolvency can be dealt with by nsing NNs in a satisfactory way and has proved useful to show how important the method of collecting data for a further analysis throngh NNs can be.
This paper presents a two stage novel technique for fingerprint feature extraction and classification. Fingerprint images are considered as texture patterns and Multi Layer Perceptron (MLP) is proposed as a feature extractor. The same fingerprint patterns are applied as input and output of MLP. The characteristics output is taken from single hidden layer as the properties of the fingerprints. These features are applied as an input to the classifier to classify the features into five broad classes. The preliminary experiments were conducted on small benchmark database and the found results were promising. The results were analyzed and compared with other similar existing techniques.
An on-chip learning Artificial Neural Network (ANN) implementation
using the Pulse Width Modulatioii (PWM) technique is proposed in this paper. Synapse and neuron are analog circuits, while digital counters are utilized to store the weights. Through the PWM circuit, the digital weight is converted into a pulse signal as the input of the analog synapse circuit. The analog modified quantity of weight is transformed into a weight-update pulse signal whose width is proportion to the value of the weight modiřication quantity. The learning rule is bcised on the weight perturbation algorithrn. In this way, the weight can be long-terni-stored and ecisily modified, thereas the synapse and the neuron are of a small size in the silicon area and the learning Circuit is feasible for implementation. Taking the advantages of both the analog and the digital realizations of the ANN, this method is a meaningful way to the implementation of on-chip ANN and fuzzy processors.
In the future, speech unquestionably will become the primary means of communication between humans and machines. New applications of artificial neural networks are capable of recognizing human speech and analyze the meaning of the recognized text. The condition of the effectiveness of two-way human-machine voice communication is to apply the mechanisms of command verification and correctness. In this paper there is a review of the selected issues on recognition and safety estimations of voice commands in natural language given by the operator of the technological device. A view is offered of the complexity of the recognition process of the operator‘s words and commands using neural networks made of a few layers of neurons. There is also an intelligent system of two-way voice communication between the technological device and the operator presented, which consists of the intelligent mechanism of operator identification, word and command recognition, command syntax and result analysis, command safety assessment, technological process supervision as well as operator reaction assessment. The paper presents research results of speech recognition and automatic command recognition as well as command safety estimation with artificial neural networks. and Obsahuje seznam literatury
The paper presents iterated algorithm for parameter estimation of non-linear regression model. The non-linear model is firstly approximated by a polynomial. Afterwards, parameter estimation based on measured data is taken as the initial value for the proposed iterated algorithm. As the estimation method, the well-known Least Square Estimation (LSE), artificial neural networks (ANN) or Bayesian methodology (BM) can be used. With respect to the knowledge of initial parameters the measured data are transformed to meet best the non-linear regression criteria (orthogonal data projection). The original and transformed data are used in the next step of the designed iterated algorithm to receive better parameter estimation. The iteration is repeated until the algorithm converges into a final result. The proposed methodology can be applied on all non-linear models that could be approximated by a polynomial function. The illustrative examples show the convergence of the designed iterated algorithm.
This paper presents a lossy compression scheme for biomedical images by using a new method. Image data compression using Vector Quantization (VQ) has received a lot of attention because of its simplicity and adaptability. VQ requires the input image to be processed as vectors or blocks of image pixels. The Finite-state vector quantization (FSVQ) is known to give better performance than the memory less vector quantization (VQ). This paper presents a novel combining technique for image compression based on the Hierarchical Finite State Vector Quantization (HFSVQ) and the neural network. The algorithm performs nonlinear restoration of diffraction-limited images concurrently with quantization. The neural network is trained on image pairs consisting of a lossless compression named hierarchical vector quantization. Simulations results are presented that demonstrate improvements in visual quality and peak signal-to-noise ratio of the restored images.
Intelligent materials, structures and structronic (structure + electronic)
Systems can function autonomously in response to the varying environmental and operating conditions [1-4]. This paper describes the application of artificial neural networks as adaptive controllers for elastic waves transformations. The results presented are connected to the modelling of srnart composite materials proposed as an integral, mechanical, neural network and electronic system which has been named as the Matrix Electronic Materials (MEM). A converse piezoelectric effect is used to suppress the amplitudes or to modify the frequency of elastic waves which propagate along the thickness of a laminated metal-ceramic plate when on the front surface of the plate the oscillating pressure is applied. The investigation of such problems relating to the elaboration of smart materials and structures, can be used for diverse technical applications, in particular, for suppression of vibrations and noise.
Heat generation in the cutting zone occurs as a result of the work done in metal cutting. In this study, in order to measure the temperature generated in the chip-forming zone, numerous experiments were carried out for different cutting regimes. During these experiments, the chip's top temperature was measured using an infrared camera. Collected data were analyzed, and temperature dependence on various cutting regimes was formulated. After that, measured data were modelled using the various techniques: response surface methodology, various types of artificial neural networks and neuro-fuzzy system. The accuracy of the proposed models is presented as well as their suitability for the considered problem. Finally, the system for the adaptive control of the cutting temperature, based on the proposed models, is presented.