By Topic

Proceedings 1993 The First New Zealand International Two-Stream Conference on Artificial Neural Networks and Expert Systems

24-26 Nov. 1993

Filter Results

Displaying Results 1 - 25 of 95
  • Main problems and issues in neural networks application

    Publication Year: 1993
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (111 KB)

    Summary form only given. Some of the specific problems that have been faced by the author in using neural networks in different application areas are outlined. They concern justifying the use of neural nets: the apparent ad hoc nature of the method; the lack of sufficient training data; the lack of an acceptable validation procedure; the need to relate network weights to physical parameters; the w... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Problems in expert systems development

    Publication Year: 1993, Page(s):166 - 167
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (169 KB)

    Summary form only given. The pitfalls of using prototyping and iteration in expert systems development are examined. It is argued that substituting prototyping for a thorough understanding of the process can cause many problems. Moreover, even when the analysis of expertise and the modeling of the domain are placed firmly within the framework of the life cycle, there are still problems with knowle... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integrating machine learning with knowledge-based systems

    Publication Year: 1993, Page(s):150 - 151
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (130 KB)

    Summary form only given. The integration of expert systems and learning systems that can manipulate the knowledge base appropriately in hybrid systems involving large knowledge bases is discussed. Hybrids of knowledge-based and learning systems with various functionalities are described. The benefits of studying such hybrid systems are examined.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fuzzy sets, fuzzy logic and the goals of artificial intelligence

    Publication Year: 1993
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (61 KB)

    Summary form only given. Investigates some of the goals of artificial intelligence and the limitations of the purely symbolic approach. The integration of diverse theories can result in a more powerful approach to the study of intelligent systems. The use of fuzzy sets for knowledge representation, and of fuzzy logic for inference under uncertainty is illustrated. The advantage of combining fuzzy ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design and implementation of video-rate real-time systems incorporating neural networks: experiments and experience

    Publication Year: 1993
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (80 KB)

    Summary form only given. The Hannibal system for video-rate facial feature detection is briefly described. The eventual application of the system is in ultra-low-bit-rate video telephony using model based coding and image generation techniques. The use of a hierarchy of neural nets together with temporal and spatial filters and maximum-likelihood classifiers provides a robust system.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Interference and discrimination in trained networks by the backpropagation algorithm

    Publication Year: 1993
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (52 KB)

    Summary form only given. A number of recent simulation studies have shown that when a connectionist net is trained, using backpropagation, to memorize sets of items in sequence and without negative exemplars, newly learned information seriously interferes with old. Three converging methods have been employed to show why and under what circumstances such retroactive interference arises. A geometric... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Is the fusion of fuzzy logic, neuroscience and chaos possible?

    Publication Year: 1993
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (54 KB)

    Summary form only given. Fuzzy linguistic terms, neuroscience, and chaos are defined. It is suggested that the fusion of these three concepts possibly could lead to the construction of a new type of information processing system.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • 1993 First New Zealand International Two-Stream Conference on Artificial Neural Networks and Expert Systems

    Publication Year: 1993, Page(s): 0_1
    Request permission for commercial reuse | PDF file iconPDF (35 KB)
    Freely Available from IEEE
  • A neural network based relaying scheme for distribution system high impedance fault detection

    Publication Year: 1993, Page(s):321 - 324
    Cited by:  Papers (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (252 KB)

    The authors present an artificial neural network (ANN)-FFT based novel detection scheme for high impedance faults (HIFs) in electrical distribution networks. The scheme uses an ANN and utilizes a discriminant vector of dominant harmonic sequence magnitudes of negative and zero (-, 0) components of the supply voltages and currents at the substation View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integration of knowledge acquired by different neural networks

    Publication Year: 1993, Page(s):73 - 74
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (196 KB)

    The author describes a set of experiments on decomposing a problem into smaller ones, training a network for each smaller problem and integrating the learned weight settings into a system capable of solving the original problem. Several network structures are suggested and performance comparisons are made. Integration of knowledge acquired by different neural networks not only can reduce the train... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Using data mining to support the construction and maintenance of expert systems

    Publication Year: 1993, Page(s):156 - 159
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (388 KB)

    Many expert systems are constructed from and allied with a large collection of databases that are continually being updated. The authors address the issues of how such a knowledge base can be constructed using tools that search the databases for significant, unexpected correlations and present them to the knowledge engineer for review. Once a knowledge base has been constructed, there is the probl... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A learning scheme for bipartite recurrent networks and its performance

    Publication Year: 1993, Page(s):34 - 37
    Cited by:  Patents (5)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (344 KB)

    A new learning scheme specialized to recurrent neural networks with the bipartite topology is proposed. The scheme is expected to have better convergence than the general Boltzmann machine learning. This improvement results from the restricted form of the network topology and an energy form devised to have a dominant global minimum. Compared to the recurrent backpropagation algorithm, the scheme i... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The identification of pitting and crevice corrosion using a neural network

    Publication Year: 1993, Page(s):325 - 326
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (148 KB)

    An artificial neural network (ANN) has been trained to monitor the electrochemical signals produced by electrodes of stainless steel during the initiation stage of localized corrosion. This exploratory study used changes in the current time series to monitor the onset of corrosion and determine whether the form of corrosion was pitting or crevice corrosion. A multilayer feedforward perceptron netw... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural network for connected character recognition

    Publication Year: 1993, Page(s):76 - 78
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (204 KB)

    A connected character recognition system using an architecture based on the selective attention model of a neural network is described. The discussion covers pattern recognition, segmentation, repairing imperfect patterns, attention focusing, search control, attention switching, size and position information, and computer simulation. Improvement of the system by the introduction of bend-detection ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Nonmonotonic inheritance reasoning in classifiers

    Publication Year: 1993, Page(s):160 - 162
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (300 KB)

    The authors investigate the significance of classifier systems with respect to nonmonotonic inheritance reasoning by constructing a simple classifier system. The experiments show that classifier systems can perform useful inheritance reasoning and may be a viable alternative to expert systems when dealing with potentially nonmonotonic environments View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural network encoding approach comparison: an empirical study

    Publication Year: 1993, Page(s):38 - 41
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (380 KB)

    The authors report the results of an empirical study about the effect of input encoding on the performance of a neural network in the classification of numerical data. Two types of encoding schemes were studied, namely numerical encoding and bit pattern encoding. Fisher Iris data were used to evaluate the performance of various encoding approaches. It was found that encoding approaches affect a ne... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Dynamic modeling and control of nonlinear systems using neural network

    Publication Year: 1993, Page(s):327 - 330
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (248 KB)

    The basic concept of design of dynamic neural controllers is discussed, emphasizing the control of the nonminimum phase plant. Neural networks are used to design a nonlinear controller combined with a robust linear controller. It unifies the nonlinear components in the controller and the plant together. This scheme simplifies the structure of the dynamic neural controller and modeling. Simulation ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Acquiring design knowledge from design cases

    Publication Year: 1993, Page(s):338 - 340
    Cited by:  Papers (2)  |  Patents (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (240 KB)

    Single-purpose learning algorithms are often not efficient in acquiring knowledge for expert systems in design. The authors present a method that learns design knowledge from design cases by the extensive generalization of abstract design schemas, which function as both a representation schema storing generalized knowledge and a memory organization facility indexing specific cases and aiding in re... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A neurocomputing model based on radial basis functions for stereo matching

    Publication Year: 1993, Page(s):92 - 95
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (308 KB)

    The authors describe a stereo model for combining the use of both monocular and binocular cues in stereo matching. The solution lies in mapping the two 2-D image planes onto a Gaussian sphere and working out the Jacobian determinant of such a mapping. Matching is then performed using zero-crossings associated with their orientation information. How such a model could be implemented using a neural ... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A hybrid approach for knowledge representation and reasoning

    Publication Year: 1993, Page(s):214 - 215
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (156 KB)

    Symbolic knowledge-based systems as well as neural networks offer several advantages, but suffer from several disadvantages if they are used as isolated systems. In consideration of this fact the main idea is to propose an architecture which is capable of handling both paradigms in one system. In the suggested approach neural networks are treated as an additional virtual symbolic knowledge represe... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Incorporating risk curves into a reservoir management model for irrigation supply

    Publication Year: 1993, Page(s):274 - 278
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (304 KB)

    A mathematical framework that incorporates risk attitudes into a management model for a reservoir supplying water to an irrigation district is proposed. The attitudes of the agricultural procedures are captured within a neural network which in turn is embedded into a stochastic dynamic programming (SDP) formulation. Rather than using the expectation of the return function to drive the SDP model, t... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Learning fuzzy rules through neural networks

    Publication Year: 1993, Page(s):137 - 139
    Cited by:  Papers (4)  |  Patents (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (276 KB)

    Presents a method for learning fuzzy rules through training a neural network with the backpropagation algorithm. Membership functions for the fuzzy concepts participating in the rules can also be learned through the proposed scheme. The learned fuzzy rules can then be implemented in a fuzzy inference machine, and a function which approximates the real goal function between the independent input va... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Combining engineering models with biophysical models to analyze a biological neural network: the electrosensory system of sharks, skates and rays

    Publication Year: 1993, Page(s):8 - 10
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (228 KB)

    The elasmobranch electrosensory system forms a neural representation of weak electrical field sources in the environment. Electrical noise correlated to the animal's own activity is rejected at an early stage of processing in the brain. The noise rejecting mechanism appears to be a neural analog of an adaptive array filter in the dorsal octavolateral nucleus of the medulla. The authors have applie... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Trading a NYSE-stock with a simple artificial neural network-based financial trading system

    Publication Year: 1993, Page(s):294 - 295
    Cited by:  Papers (2)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (144 KB)

    Reports hypothetical trading results of a New York Stock Exchange (NYSE) listed stock over a period of two years using an artificial neural network (ANN) based financial trading system. The system was designed, constructed and tested for its ability to predict stock prices and more importantly increase trading profit. This system is still at a preliminary stage and many of the parameters effect on... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Knowledge capture model for expert systems development

    Publication Year: 1993, Page(s):163 - 164
    Cited by:  Papers (1)
    Request permission for commercial reuse | Click to expandAbstract | PDF file iconPDF (152 KB)

    The development of expert systems requires the efficient capturing and organizing of diverse knowledge in a form suitable for encoding in an expert system shell. A knowledge capture model has been developed and applied for expertise in the selection and use of materials. Two expert systems are currently under development using the model: GeoMat on selection and use of materials for geothermal ener... View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.