The logic of neural networks

This paper establishes the equivalence between multilayer feedforward networks and linear combinations of Lukasiewicz propositions. In this sense, multilayer forward networks have a logic interpretation, which should permit to apply logical techniques in the neural networks framework.

Bibliographic Details
Authors: Castro Peña, Juan Luis, Trillas i Gay, Enric
Format: article
Publication Date:1998
Country:España
Institution:Universitat Politècnica de Catalunya (UPC)
Repository:UPCommons. Portal del coneixement obert de la UPC
Language:English
OAI Identifier:oai:upcommons.upc.edu:2099/3501
Online Access:https://hdl.handle.net/2099/3501
Access Level:Open access
Keyword:Feedforward networks
Lukasiewicz logic
Universal aproximators
Squashing functions
Lògica matemàtica
Xarxes neuronals (Informàtica)
Classificació AMS::03 Mathematical logic and foundations::03B General logic
Description
Summary:This paper establishes the equivalence between multilayer feedforward networks and linear combinations of Lukasiewicz propositions. In this sense, multilayer forward networks have a logic interpretation, which should permit to apply logical techniques in the neural networks framework.