Title :
Counting Boolean networks are universal approximators
Author_Institution :
IST, Lisbon Univ., Portugal
Abstract :
A Boolean neural model is presented, where fuzzy reasoning emerges as a macroscopic property from individual neuron Boolean counting operations and random inter-neuron connections. The main objective of this work is to demonstrate that such networks are Universal Approximators. This is achieved through well known properties of non parametric techniques (Parzen Window estimators) to estimate any probability density function
Keywords :
Boolean algebra; fuzzy neural nets; fuzzy set theory; inference mechanisms; probability; uncertainty handling; Boolean neural model; Parzen Window estimators; counting Boolean networks; fuzzy reasoning; macroscopic property; neuron Boolean counting operations; non parametric techniques; probability density function; random inter-neuron connections; universal approximators; Boolean functions; Flip-flops; Fuzzy neural networks; Fuzzy reasoning; Fuzzy sets; Fuzzy systems; Neural networks; Neurons; Probability density function; Production;
Conference_Titel :
Fuzzy Information Processing Society - NAFIPS, 1998 Conference of the North American
Conference_Location :
Pensacola Beach, FL
Print_ISBN :
0-7803-4453-7
DOI :
10.1109/NAFIPS.1998.715567
