Neural Networks with Mixed Activation Functions

다양한 활성 함수를 사용하는 신경회로망의 구성

  • Lee, Chung-Yeol (School of Electrical Engineering and Computer Science Korea Advanced Institute of Science and Technology) ;
  • Park, Cheol-Hoon (School of Electrical Engineering and Computer Science Korea Advanced Institute of Science and Technology)
  • 이충열 (한국과학기술원 전자전산학부) ;
  • 박철훈 (한국과학기술원 전자전산학부)
  • Published : 2008.06.18

Abstract

When we apply the neural networks to applications, we need to select proper architecture of the network and the activation function of the network is one of most important characteristics. In this research, we propose a method to make a network using multiple activation functions. The performance of the proposed method is investigated through the computer simulations on various regression problems.

Keywords