Paraconsistent artificial neural networks: An introduction

Abstract

In this work we sketch a new theory of artificial neural network, based on a paraconsistent annotated logic Et. Such theory, called Paraconsistent Artificial Neural Network - PANN - is built from the algorithm Para-analyzer and has as characteristics the capability of mainly manipulating uncertainty, inconsistent and paracomplete concepts. Some aspects such as capability of adaptation, velocity processing, and other useful characteristics make the PANN a promising theory. © Springer-Verlag 2004.