Class n-gram models for very large vocabulary speech recognition of Finnish and Estonian

Matti Varjokallio*, Mikko Kurimo, Sami Virpioja

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

4 Citations (Scopus)

Abstract

We study class n-gram models for very large vocabulary speech recognition of Finnish and Estonian. The models are trained with vocabulary sizes of several millions of words using automatically derived classes. To evaluate the models on Finnish and an Estonian broadcast news speech recognition task, we modify Aalto University’s LVCSR decoder to operate with the class n-grams and very large vocabularies. Linear interpolation of a standard n-gram model and a class n-gram model provides relative perplexity improvements of 21.3% for Finnish and 12.8% for Estonian over the n-gram model. The relative improvements in word error rates are 5.5% for Finnish and 7.4% for Estonian. We also compare our word-based models to a state-of-the-art unlimited vocabulary recognizer utilizing subword n-gram models, and show that the very large vocabulary word-based models can perform equally well or better.

Original languageEnglish
Title of host publicationProceedings of the 4th International Conference on Statistical Language and Speech Processing, SLSP 2016
Pages133-144
Number of pages12
Volume9918 LNCS
DOIs
Publication statusPublished - 2016
MoE publication typeA4 Article in a conference publication
EventInternational Conference on Statistical Language and Speech Processing - Pilsen, Czech Republic
Duration: 11 Oct 201612 Oct 2016
Conference number: 4
http://grammars.grlmc.com/SLSP2016/

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9918 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Conference

ConferenceInternational Conference on Statistical Language and Speech Processing
Abbreviated titleSLSP
CountryCzech Republic
CityPilsen
Period11/10/201612/10/2016
Internet address

Keywords

  • Class n-gram models
  • Language modelling
  • Morphologically rich languages
  • Speech recognition

Fingerprint Dive into the research topics of 'Class n-gram models for very large vocabulary speech recognition of Finnish and Estonian'. Together they form a unique fingerprint.

Cite this