Stability of Learning in Classes of Recurrent and Feedforward Networks Wilson, William Hulme en_US 2021-11-25T13:11:44Z 2021-11-25T13:11:44Z 1995 en_US
dc.description.abstract This paper concerns a class of recurrent neural networks related to Elman networks (simple recurrent networks) and Jordan networks and a class of feedforward networks architecturally similar to Waibel’s TDNNs. The recurrent nets used herein, unlike standard Elman/Jordan networks, may have more than one state vector. It is known that such multi-state Elman networks have better learning performance on certain tasks than standard Elman networks of similar weight complexity. The task used involves learning the graphotactic structure of a sample of about 400 English words. Learning performance was tested using regimes in which the state vectors are, or are not, zeroed between words: the former results in larger minimum total error, but without the large oscillations in total error observed when the state vectors are not periodically zeroed. Learning performance comparisons of the three classes of network favour the feedforward nets. en_US
dc.language English
dc.language.iso EN en_US
dc.publisher University of Sydney, Electrical Engineering en_US
dc.rights CC BY-NC-ND 3.0 en_US
dc.rights.uri en_US
dc.source Legacy MARC en_US
dc.subject.other Elman network en_US
dc.subject.other simple recurrent network en_US
dc.subject.other tower network en_US
dc.subject.other Jordan network en_US
dc.subject.other Neural Networks, Genetic Algorithms and Fuzzy Logic (280212) en_US
dc.title Stability of Learning in Classes of Recurrent and Feedforward Networks en_US
dc.type Conference Paper en
dcterms.accessRights open access
dspace.entity.type Publication en_US
unsw.description.notePublic Further publications of this author can be found at en_US
unsw.identifier.doi Sydney, Australia en_US
unsw.relation.faculty Other UNSW
unsw.relation.ispartofconferenceLocation Sydney, Australia en_US
unsw.relation.ispartofconferenceName Australian Conference on Neural Networks en_US
unsw.relation.ispartofconferenceProceedingsTitle Proceedings of the Sixth Australian Conference on Neural Networks (ACNN’95) en_US
unsw.relation.ispartofconferenceYear 1995 en_US
unsw.relation.ispartofpagefrompageto 142-145 en_US
unsw.relation.originalPublicationAffiliation Wilson, William Hulme, UNSW en_US
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
wilson stability.pdf
98.88 KB
Resource type