Title :
On reliable computation with formal neurons
Author :
Venkatesh, Santosh S. ; Psaltis, Demetri
Author_Institution :
Moore Sch. of Electr. Eng., Pennsylvania Univ., Philadelphia, PA, USA
fDate :
1/1/1992 12:00:00 AM
Abstract :
The authors investigate the computing capabilities of formal McCulloch-Pitts neurons when errors are permitted in decisions. They assume that m decisions are to be made on a randomly specified m set of points in n space and that an error tolerance of εm decision errors is allowed, with 0⩽ε<1/2. The authors are interested in how large an m can be selected such that the neuron makes reliable decisions within the prescribed error tolerance. Formal results for two protocols for error-tolerance-a random error protocol and an exhaustive error protocol-are obtained. The results demonstrate that a formal neuron has a computational capacity that is linear in n and that this rate of capacity growth persists even when errors are tolerated in the decisions
Keywords :
neural nets; protocols; computing capabilities; decision errors; error tolerance; exhaustive error protocol; formal McCulloch-Pitts neurons; neural nets; random error protocol; Biological system modeling; Biology computing; Boolean functions; Computer errors; Computer networks; Fault tolerance; Neurons; Pattern recognition; Protocols; Reliability theory;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
