Academic
Publications
Neural network learning and expert systems

Neural network learning and expert systems,Stephen I. Gallant

Neural network learning and expert systems   (Citations: 290)
BibTex | RIS | RefWorks Download
Published in 1993.
Cumulative Annual
    • ...An expert system is an algorithmic program that draws upon a database comprised of expert knowledge to guide decisions on how to proceed at different points in the assessment process (Gallant, 1993, p...

    Donald C. Mattson. An Introduction to the Computerized Assessment of Art-Based Instrument...

    • ...The second weak learner algorithm is called Pocket with a ratchet or pocket (Gallant 1993)...

    U. Rajendra Acharyaet al. An integrated diabetic index using heart rate variability signal featu...

    • ...Furthermore, an application is described of the adaptive DE algorithm with multiple trial vectors for the training of feedforward flat ANNs [16] for the classification of the parity-p problem...
    • ...The EBP algorithm is based on the gradient method and permits efficient neural network training for solving difficult problems, which often refer to nonseparable data [16]...
    • ...It is necessary to point out that the algorithm presented can also be used easily to train multioutput ANNs, ANNs with nonstandard architectures (for example, the tower architecture [16]), and networks with a nondifferentiable neuron activation function for which applications of the EBP or LM algorithms are not possible...

    Adam Slowik. Application of an Adaptive Differential Evolution Algorithm With Multi...

    • ...Gallant 1993), and both Hyperion and TM data were processed using spectral unmixing methods (Adams and Smith 1986, Boardman 1989), based on end-member spectra...

    David W. Leverington. Discrimination of sedimentary lithologies using Hyperion and Landsat T...

    • ...Although within neurosymbolic approaches there have been advanced efforts at combining neural networks with first-order logic [43], [22], [4] or with multivalued logic [33] or even with nonclassical logics [17], [18], [19], [34], those that combine symbolic rules (of propositional type) and neural networks still possess a great part [13], [50], [31], [48], [14], [24], [25] and seem to have given more applied results [38], [52], [53], [54]...
    • ...Neural networks represent a totally different approach to problemsolving,knownasconnectionism(see,e.g.,[13]).The mainadvantagesofneuralnetworksaretheirabilitytoobtain their knowledge from training examples (thus reducing the interaction with the experts to a minimum), their ability to generalize and represent complex and imprecise knowledge, and their high level of efficiency...
    • ...Weights are set in a way that makes the network to infer correctly. MACIE [12], [13] is such a system...
    • ...The adaline unit uses the LMS algorithm for learning, which more safely converges for nonlinear training sets and generalizes better than perceptron (see, e.g., [13])...
    • ...To demonstrate NPA, we use an example problem from [13] and its corresponding data set (called ACUTE)...
    • ...They refer to the following: . The neurules integrated inference engine (IIE) is compared to the inference mechanisms used in connectionist expert systems and presented in [12], [13] (MACIE) and [21] (RIE)...
    • ...The first data set (ACUTE) is the one used as an example in Section 4 and is taken from [13]...
    • ...MACIE and RIE were implemented in a simulated way and also applied to the four connectionist knowledge bases, which were created from the same data sets according to the guidelines presented in [13]...

    Ioannis Hatzilygeroudiset al. Integrated Rule-Based Learning and Inference

Sort by: