Tuning of learning rate and momentum on back-propagation

Kamiyama, Noriaki; Iijima, N.; Taguchi, Akira; Mitsui, H.; Yoshida, Yasuhiko; Sone, M.

doi:10.1109/iccs.1992.254895

Cited by 7 publications

(4 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several methods have been introduced to determine LPs. Kamiyama et al (1992) defined the learning and momentum coefficients as a linear function based on different learning and momentum coefficients data. Zaghwl and Dong (1994) utilized conjugate gradient methods for optimization of the learning and momentum coefficients.…”

Section: Related Workmentioning

confidence: 99%

Design of highly effective multilayer feedforward neural network by using genetic algorithm

2020

View full text Add to dashboard Cite

This paper presents a highly effective and precise neural network method for choosing the activation functions (AFs) and tuning the learning parameters (LPs) of a multilayer feedforward neural network by using a genetic algorithm (GA). The performance of the neural network mainly depends on the learning algorithms and the network structure. The backpropagation learning algorithm is used for tuning the network connection weights, and the LPs are obtained by the GA to provide both fast and reliable learning. Also, the AFs of each neuron in the network are automatically chosen by a GA. The present study consists of 10 different functions to accomplish a better convergence of the desired input–output mapping. Test studies are performed to solve a set of two‐dimensional regression problems for the proposed genetic‐based neural network (GNN) and conventional neural network having sigmoid AFs and constant learning parameters. The proposed GNN has also been tested by applying it to three real problems in the fields of environment, medicine, and economics. Obtained results prove that the proposed GNN is more effective and reliable when compared with the classical neural network structure.

show abstract

Section: Related Workmentioning

confidence: 99%

Design of highly effective multilayer feedforward neural network by using genetic algorithm

2020

View full text Add to dashboard Cite

show abstract

“…Regarding topology, it has been shown by Hornik, Stinchombe and White [19] that only one hidden layer is sufficient to approximate virtually any function to any degree of accuracy. However, no definitive procedure exists which specifies the optimal choice of the number of hidden nodes, although suggestions based upon theoretical considerations have been advanced by Kurkova [20] [21], and based on empirical consideration by Kamiyama et al [22], among others. In accordance with the findings of Miasek and Lin [8], we examined four different configuration, as follow: Case a: 28-10-4; Case b: 28-15-4; Case C: 28-5-4; and Case d: 28-2-4.…”

Section: Application Of the Clinical Matrixmentioning

confidence: 99%

Application of the clinical matrix to the differential diagnosis of intestinal diseases

Hedgepeth

Lin

Pawar

et al. 1999

Journal of Discrete Mathematical Sciences and Cryptography

View full text Add to dashboard Cite

We outline a methodology by which a medical diagnosis can be arrived at by the presentation of a list of signs and symptoms. The Clinical Matrix is trained using the Backpropagation paradigm. The test results for four Intestinal Diseases are self-consistent and in most cases def'mitivf!. Ambigtlities can be resolved by a single, simple laboratory test. A field test was conducted for empirical validation.

show abstract

“…To our knowledge, only a few research studies have analyzed the interdependencies among the backpropagation paradigm parameters regarding their tuning and scaling for different applications. Some parameters are maintained constant, while relationships among the others, representing their influences on learning, are established (Fahlman 1988;Kamiyama et al 1992;Sundararajan et al 1993;Tesauro and Janssens 1988). Only a fraction of the paradigm parameters are then considered in these relationships.…”

Section: Franpis Michaud and Ruben Gonzalez Rubiomentioning

confidence: 99%

Autonomous Design of Artificial Neural Networks by Neurex

Michaud

Rubio

1996

Neural Computation

View full text Add to dashboard Cite

Artificial neural networks (ANN) have been demonstrated to be increasingly more useful for complex problems difficult to solve with conventional methods. With their learning abilities, they avoid having to develop a mathematical model or acquiring the appropriate knowledge to solve a task. The difficulty now lies in the ANN design process. A lot of choices must be made to design an ANN, and there are no available design rules to make these choices directly for a particular problem. Therefore, the design for an ANN demands a certain number of iterations, mainly guided by the expertise and the intuition of the developer. To automate the ANN design process, we have developed Neurex, composed of an expert system and an ANN simulator. Neurex autonomously guides the iterative ANN design process. Its structure tries to reproduce the design steps done by a human expert in conceiving an ANN. As a whole, the Neurex structure serves as a framework to implement this expertise for different learning paradigms. This article presents the system's general characteristics and its use in designing ANN using the standard backpropagation learning law.

show abstract

Tuning of learning rate and momentum on back-propagation

Cited by 7 publications

References 1 publication

Design of highly effective multilayer feedforward neural network by using genetic algorithm

Design of highly effective multilayer feedforward neural network by using genetic algorithm

Application of the clinical matrix to the differential diagnosis of intestinal diseases

Autonomous Design of Artificial Neural Networks by Neurex

Contact Info

Product

Resources

About