Recognition Performance of a Structured Language Model

dc.creatorChelba, Ciprian
dc.creatorJelinek, Frederick
dc.date2000-01-24
dc.date.accessioned2026-07-07T03:15:52Z
dc.date.available2026-07-07T03:15:52Z
dc.descriptionA new language model for speech recognition inspired by linguistic analysis is presented. The model develops hidden hierarchical structure incrementally and uses it to extract meaningful information from the word history - thus enabling the use of extended distance dependencies - in an attempt to complement the locality of currently used trigram models. The structured language model, its probabilistic parameterization and performance in a two-pass speech recognizer are presented. Experiments on the SWITCHBOARD corpus show an improvement in both perplexity and word error rate over conventional trigram models.
dc.description4 pages
dc.identifierhttps://arxiv.org/abs/cs/0001022
dc.identifierhttp://arxiv.org/abs/cs/0001022
dc.identifierProceedings of Eurospeech, 1999, pp. 1567-1570, Budapest, Hungary
dc.identifier.urihttp://salesiana.dossiersoluciones.com/handle/123456789/30142
dc.subjectComputation and Language
dc.subjectG.3, I.2.7, I.5.1, I.5.4
dc.titleRecognition Performance of a Structured Language Model
dc.typetext

Files

Collections