Parameters’ analysis as a tool for boosting neural networks’ learning efficiency

V.G. Prokhorov

Abstract


General traits of neural networks’ learning using error back propagation algorithm were reviewed. Different networks’ parameters and their influence on learning efficiency was analyzed. Various changes in parameters set and resulting changes in the learning process of neural networks were described.

Problems in programming 2009; 4: 89-95


References


Le Cunn Y., Bottou L., Orr G.B. Neural Networks: Tricks of the trade, Springer. – 1998. – Р. 1 – 5.

Каллан Р. Основные концепции нейронных сетей. – М.: Вильямс, 2001. – С. 80 – 196.

Burges C.J.C. A Method for Training Neural Network to Recognize Character Strings, AT&T Bell Laboratories. – 1992. – Р. 1 – 8.

Simard P.Y. Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis, Microsoft . – 1998. – Р. 23 – 24.

Le Cunn Y. Efficient BackProp, Speech and Image Processing Services Research AT&T Lab. – 1998. – Р. 5 – 16.

Vaillant R. Localization of Objects in Images, Speech and Image Processing Services Research AT&T Lab. – 1994. – Р. 1 – 13.

Хайкин C. Нейронные сети. Полный курс Изд. второе (испр.). Прэнтис Холл. – 2006. – С. 239 – 298 ; 308 – 315.

Le Cunn Y., Bottou L., Haffner P. Gradient Based Learning Applied to Document Recognition, IEEE Press. – 1998 – Р. 4 – 12.

Прохоров В. Использование сверхточных сетей для распознавания рукописных символов // Проблеми програмування. – 2008. – № 2-3. – С. 669 – 674.


Refbacks

  • There are currently no refbacks.