An evaluation of a spoken document retrieval baseline system in finish

This paper presents a baseline spoken document retrieval system in Finnish. Due to its agglutinative structure, Finnish speech can not be adequately transcribed using the standard large vocabulary continuous speech recognition approaches. The definition of a sufficient lexicon and the training of the statistical language models are difficult, because the words appear transformed by many inflections and compounds. In this work we apply a recently developed unlimited vocabulary speech recognition system that allows the use of n-gram language models based on morpheme-like subword units discovered in an unsupervised manner. In addition to word-based indexing, we also propose an indexing based on the subword units provided directly by our speech recognizer. In an initial evaluation of newsreading in Finnish, we obtained a fairly low recognition error rate and average document retrieval precisions close to that from human reference transcripts.

[1]  Ian H. Witten,et al.  Managing Gigabytes: Compressing and Indexing Documents and Images , 1999 .

[2]  Steve Renals,et al.  Indexing and retrieval of broadcast news , 2000, Speech Commun..

[3]  Mikko Kurimo,et al.  Morphologically Motivated Language Models in Speech Recognition , 2005 .

[4]  Eero Sormunen,et al.  A Method for Measuring Wide Range Performance of Boolean Queries in Full-Text Databases , 2000 .

[5]  Ellen M. Voorhees,et al.  The TREC Spoken Document Retrieval Track: A Success Story , 2000, TREC.

[6]  Mikko Kurimo,et al.  On lexicon creation for turkish LVCSR , 2003, INTERSPEECH.

[7]  Andreas Stolcke,et al.  SRILM - an extensible language modeling toolkit , 2002, INTERSPEECH.

[8]  M. Kurimo,et al.  Using phone durations in finnish large vocabulary continuous speech recognition , 2004, Proceedings of the 6th Nordic Signal Processing Symposium, 2004. NORSIG 2004..

[9]  William J. Byrne,et al.  On large vocabulary continuous speech recognition of highly inflectional language - czech , 2001, INTERSPEECH.

[10]  Kimmo Koskenniemi,et al.  A General Computational Model for Word-Form Recognition and Production , 1984 .

[11]  Mathias Creutz,et al.  Unsupervised Discovery of Morphemes , 2002, SIGMORPHON.

[12]  Mikko Kurimo,et al.  Unlimited vocabulary speech recognition based on morphs discovered in an unsupervised manner , 2003, INTERSPEECH.

[13]  Dietrich Klakow,et al.  Speech recognition for huge vocabularies by using optimized sub-word units , 2001, INTERSPEECH.

[14]  John H. L. Hansen,et al.  Speechfind: an experimental on-line spoken document retrieval system for historical audio archives , 2002, INTERSPEECH.