Some Approaches of Improving the Quality of Artificial Neural Networks Training

EasyChair Preprint 2087

Abstract

The paper is devoted to a problem of regular improving the quality of artificial neural networks’ (ANN) training. The object of study is a complex neural network consists of 2-dimensional Kohonen network and Wilshaw and von der Malsburg network. These networks are applied to timetable problem for transport systems. The main existing results of using optimal control theory for ANN training are analyzed; authors suggest a new technique based on direct neural control. Authors give comparative values of error during training process for traditional methods and  the new approach. It is presented that the new technique is better than traditional one for considered neural networks.

Keyphrases: Artificial Neural Network, Error level, algorithm, direct control, optimal control

@booklet{EasyChair:2087,
  author    = {Yefim Rozenberg and Alexey Olshansky and Ignat Dovgerd and Gleb Dovgerd and Alexander Ignatenkov and Paul Ignatenkov},
  title     = {Some Approaches of Improving the Quality of Artificial Neural Networks Training},
  howpublished = {EasyChair Preprint 2087},
  year      = {EasyChair, 2019}}