GRAPHICAL MODEL REPRESENTATIONS OF WORD LATTICES
- 1 January 2006
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE) in 2006 IEEE Spoken Language Technology Workshop
- p. 162-165
- https://doi.org/10.1109/slt.2006.326842
Abstract
We introduce a method for expressing word lattices within a dynamic graphical model. We describe a variety of choices for doing this, including a technique to relax the time information associated with lattice nodes in a way that trades off hypothesis expansion with presumed segmentation boundary accuracy. Our approach uses a set of time-inhomogeneous and algorithmically expressed conditional probability tables to encode the lattice. The approach was implemented as part of the graphical model toolkit, and word error rate improvements on the Switchboard corpus indicate that our technique is a viable means to incorporate large state space speech recognition systems into a graphical model.Keywords
This publication has 5 references indexed in Scilit:
- Development of the CU-HTK 2004 Broadcast News Transcription SystemsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2006
- The IBM 2004 Conversational Telephony System for Rich TranscriptionPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2006
- Graphical model architectures for speech recognitionIEEE Signal Processing Magazine, 2005
- Multi-speaker language modelingPublished by Association for Computational Linguistics (ACL) ,2004
- A review of large-vocabulary continuous-speechIEEE Signal Processing Magazine, 1996