Training fully recurrent neural networks with complex weights

Author

Kechriotis, George ; Manolakos, Elias S.

Author_Institution

Dept. of Electr. & Comput. Eng., Northeastern Univ., Boston, MA, USA

Volume

41

Issue

3

fYear

1994

fDate

3/1/1994 12:00:00 AM

Firstpage

235

Lastpage

238

Abstract

In this brief paper, the Real Time Recurrent Learning (RTRL) algorithm for training fully recurrent neural networks in real time, is extended for the case of a recurrent neural network whose inputs, outputs, weights and activation functions are complex. A practical definition of the complex activation function is adopted and the complex form of the conventional RTRL algorithm is derived. The performance of the proposed algorithm is demonstrated with an application in complex communication channel equalization

Keywords

learning (artificial intelligence); recurrent neural nets; telecommunication channels; transfer functions; RTRL algorithm; Real Time Recurrent Learning; activation functions; complex communication channel equalization; complex weights; fully recurrent neural networks; network inputs; network outputs; network weights; Adaptive equalizers; Backpropagation algorithms; Communication channels; Digital signal processing; Limit-cycles; Neural networks; Recurrent neural networks; Signal processing algorithms; Speech processing; State-space methods;

fLanguage

English

Journal_Title

Circuits and Systems II: Analog and Digital Signal Processing, IEEE Transactions on

Publisher

ieee

ISSN

1057-7130

Type

jour

DOI

10.1109/82.279210

Filename

279210