INFERENCE DEVICE, INFERENCE METHOD AND PROGRAM

PROBLEM TO BE SOLVED: To provide an inference device, an inference method and a program wherein an input pattern can be expressed by a real number value vector and an if-then rule including a conjunction, a disjunction and a negation can be learned by a self-propagating neural network. SOLUTION: The...

Full description

Saved in:
Bibliographic Details
Main Authors SUDO AKITO, ZHANG CHENLI, TSUBOYAMA MANABU, HASEGAWA OSAMU
Format Patent
LanguageEnglish
Published 18.12.2008
Subjects
CALCULATING
COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
COMPUTING
COUNTING
PHYSICS
Online AccessGet full text

Cover

More Information
Summary:PROBLEM TO BE SOLVED: To provide an inference device, an inference method and a program wherein an input pattern can be expressed by a real number value vector and an if-then rule including a conjunction, a disjunction and a negation can be learned by a self-propagating neural network. SOLUTION: The inference device is provided with a learning part 40 for learning the if-then rule as an input pattern by using the self-propagating neural network and an inference part 50 for executing inference by using a learned result of the learning part 40. The self-propagating neural network successively inputs input patterns described by a multi-dimensional vectors, automatically propagates nodes on the basis of the input patterns and nodes described by the multi-dimensional vectors arranged in the neural network, and connects respective nodes by sides. COPYRIGHT: (C)2009,JPO&INPIT
Bibliography:Application Number: JP20070151252