This thesis considers the entire automated speech recognition process and presents a standardised approach to LVCSR experimentation with HMMs. It also discusses various approaches to speaker adaptation such as MLLR and multiscale, and presents experimental results for cross-task speaker adaptation. An analysis of training parameters and data sufficiency for reasonable system performance estimates are also included. It is found that Maximum Likelihood Linear Regression (MLLR) supervised adaptation can result in 6% reduction (absolute) in word error rate given only one minute of adaptation data, as compared with an unadapted model set trained on a different task. The unadapted system performed at 24% WER and the adapted system at 18% WER. This is achieved with only 4 to 7 adaptation classes per speaker, as generated from a regression tree.
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:OWTU.10012/840 |
| Date | January 2000 |
| Creators | Stokes-Rees, Ian James |
| Publisher | University of Waterloo |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Thesis or Dissertation |
| Format | application/pdf, 512540 bytes, application/pdf |
| Rights | Copyright: 2000, Stokes-Rees, Ian James. All rights reserved. |
Page generated in 0.0022 seconds