Differential association and operational equivalence of discrete Hopfield networks

Likas, Aristidis; Stafylopatis, Andreas

dc.contributor.author	Likas, Aristidis	en
dc.contributor.author	Stafylopatis, Andreas	en
dc.date.accessioned	2014-03-01T01:43:37Z
dc.date.available	2014-03-01T01:43:37Z
dc.date.issued	1995	en
dc.identifier.issn	12100552	en
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/24171
dc.relation.uri	http://www.scopus.com/inward/record.url?eid=2-s2.0-0029180068&partnerID=40&md5=b41fb29b26d5ce0d7b9b244f0c13cdab	en
dc.subject.other	Associative storage	en
dc.subject.other	Equivalence classes	en
dc.subject.other	Mathematical transformations	en
dc.subject.other	Vectors	en
dc.subject.other	Binary Hopfield neural network	en
dc.subject.other	Differential association	en
dc.subject.other	Operational equivalence	en
dc.subject.other	State vector	en
dc.subject.other	Neural networks	en
dc.title	Differential association and operational equivalence of discrete Hopfield networks	en
heal.type	journalArticle	en
heal.publicationDate	1995	en
heal.abstract	An original transform is presented which, given a binary Hopfield neural network and a state vector of this network, creates a new binary Hopfield network of the same size and establishes a correspondence between the states of the two networks during operation. The derived network operates in a differential manner with respect to the initial one, in that its state vector represents the deviation of the corresponding state of the original network from the base state used for the transform. As a consequence, the energy of the second network accepts an analogous interpretation. The transform exhibits several interesting properties which are proved and discussed. Moreover, the notion of operational equivalence is introduced and it is shown that equivalence classes of binary Hopfield networks can be defined on the basis of the transform.	en
heal.publisher	VSP Int Sci Publ, Zeist, Netherlands	en
heal.journalName	Neural Network World	en
dc.identifier.volume	5	en
dc.identifier.issue	1	en
dc.identifier.spage	81	en
dc.identifier.epage	89	en