Show simple item record

dc.contributor.authorRenals, Steve
dc.contributor.authorHochberg, Mike
dc.contributor.authorRobinson, Tony
dc.coverage.spatial8en
dc.date.accessioned2006-05-18T14:28:04Z
dc.date.available2006-05-18T14:28:04Z
dc.date.issued1994
dc.identifier.citationAdvances in Neural Information Processing Systems, vol. 6, pp. 1051-1058.en
dc.identifier.isbn1558603220
dc.identifier.urihttp://hdl.handle.net/1842/1109
dc.description.abstractHybrid connectionist/HMM systems model time using both a Markov chain and through properties of a connectionist network. In this paper, we discuss the nature of the time dependence currently employed in our systems using recurrent networks (RNs) and feed-forward multi-layer perceptrons (MLPs). In particular, we introduce local recurrences into an MLP to produce an enhanced input representation. This is in the form of an adaptive gamma filter and incorporates an automatic approach for learning temporal dependencies. We have experimented on a speaker-independent phone recognition task using the TIMIT database. Results using the gamma filtered input representation have shown improvement over the baseline MLP system. Improvements have been obtained through merging the baseline and gamma filter models.en
dc.format.extent774550 bytes
dc.format.mimetypeapplication/pdf
dc.language.isoen
dc.publisherMorgan Kauffmanen
dc.titleLearning temporal dependencies in connectionist speech recognitionen
dc.typeConference Paperen


Files in this item

This item appears in the following Collection(s)

Show simple item record