Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies

Seppo Enarvi; Peter Smit; Sami Virpioja; Mikko Kurimo

doi:10.1109/TASLP.2017.2743344

Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies

Seppo Enarvi, Peter Smit, Sami Virpioja, Mikko Kurimo

Research output: Contribution to journal › Article › Scientific › peer-review

11 Citations (Scopus)

220 Downloads (Pure)

Abstract

Today, the vocabulary size for language models in large vocabulary speech recognition is typically several hundreds of thousands of words. While this is already sufficient in some applications, the out-of-vocabulary words are still limiting the usability in others. In agglutinative languages the vocabulary for conversational speech should include millions of word forms to cover the spelling variations due to colloquial pronunciations, in addition to the word compounding and inflections. Very large vocabularies are also needed, for example, when the recognition of rare proper names is important.

Original language	English
Pages (from-to)	2085-2097
Number of pages	13
Journal	IEEE/ACM Transactions on Audio, Speech, and Language Processing
Volume	25
Issue number	11
Early online date	23 Aug 2017
DOIs	https://doi.org/10.1109/TASLP.2017.2743344
Publication status	Published - Nov 2017
MoE publication type	A1 Journal article-refereed

Access to Document

10.1109/TASLP.2017.2743344

Accepted version, with edits, IEEE copyright, and erratum.
(c) 2017 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.
Accepted author manuscript, 1.9 MBLicence: Other

Fingerprint Dive into the research topics of 'Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies'. Together they form a unique fingerprint.

Science-IT
Mikko Hakala (Manager)
School of Science
Facility/equipment: Facility

Cite this

@article{740669405e5d4208af53e61615e0603c,

title = "Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies",

abstract = "Today, the vocabulary size for language models in large vocabulary speech recognition is typically several hundreds of thousands of words. While this is already sufficient in some applications, the out-of-vocabulary words are still limiting the usability in others. In agglutinative languages the vocabulary for conversational speech should include millions of word forms to cover the spelling variations due to colloquial pronunciations, in addition to the word compounding and inflections. Very large vocabularies are also needed, for example, when the recognition of rare proper names is important.",

author = "Seppo Enarvi and Peter Smit and Sami Virpioja and Mikko Kurimo",

year = "2017",

month = nov,

doi = "10.1109/TASLP.2017.2743344",

language = "English",

volume = "25",

pages = "2085--2097",

journal = "IEEE/ACM Transactions on Audio, Speech, and Language Processing",

issn = "2329-9290",

publisher = "IEEE Advancing Technology for Humanity",

number = "11",

}

TY - JOUR

T1 - Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies

AU - Enarvi, Seppo

AU - Smit, Peter

AU - Virpioja, Sami

AU - Kurimo, Mikko

PY - 2017/11

Y1 - 2017/11

N2 - Today, the vocabulary size for language models in large vocabulary speech recognition is typically several hundreds of thousands of words. While this is already sufficient in some applications, the out-of-vocabulary words are still limiting the usability in others. In agglutinative languages the vocabulary for conversational speech should include millions of word forms to cover the spelling variations due to colloquial pronunciations, in addition to the word compounding and inflections. Very large vocabularies are also needed, for example, when the recognition of rare proper names is important.

AB - Today, the vocabulary size for language models in large vocabulary speech recognition is typically several hundreds of thousands of words. While this is already sufficient in some applications, the out-of-vocabulary words are still limiting the usability in others. In agglutinative languages the vocabulary for conversational speech should include millions of word forms to cover the spelling variations due to colloquial pronunciations, in addition to the word compounding and inflections. Very large vocabularies are also needed, for example, when the recognition of rare proper names is important.

U2 - 10.1109/TASLP.2017.2743344

DO - 10.1109/TASLP.2017.2743344

M3 - Article

VL - 25

SP - 2085

EP - 2097

JO - IEEE/ACM Transactions on Audio, Speech, and Language Processing

JF - IEEE/ACM Transactions on Audio, Speech, and Language Processing

SN - 2329-9290

IS - 11

ER -

Automatic Speech Recognition with Very Large Conversational Finnish and Estonian Vocabularies

Abstract

Access to Document

Science-IT

Cite this