Paper Info
Title
Language modeling structures in audio transcription for retrieval of historical speeches
Abstract
In this paper we apply speech recognition for automatic tran- script generation for spoken document retrieval. The tran- scripts are used to compute an index for an archive of his- torical speeches and to provide the index, speech, and tran- scripts available for query based retrieval and browsing. I n addition to acoustic variability, the task is challenging, be- cause it covers a broad spectrum of different speaking styles and use of language. Language modeling is important for speech recognition to determine the prior probabilities of the compared word and sentence candidates in decoding. Vari- ous large text corpora are available in electronic format fo r language model training, but the open question is what and how should we include to improve the audio transcripts of this task. In this work we compare large overall language models to focused ones trained on selected subsets of the data, and to combinations between both. With respect to the potential index terms, improvements were obtained for tran- scripts that did not fit well to the scope of the large overall language model.
Year
Venue
Keywords
2004
EUSIPCO
information retrieval,probability,speech recognition,acoustic variability,audio transcription,automatic transcript generation,historical speeches,language modeling structures,large text corpora,query based retrieval,spoken document retrieval
Field
DocType
ISBN
Speech corpus,Cache language model,Speech analytics,Computer science,Audio mining,Speech recognition,Transcription (software),Natural language processing,Artificial intelligence,Document retrieval,Language model,Acoustic model
Conference
978-320-0001-65-7
Citations 
PageRank 
References 
1
0.36
10
Authors
4
Name
Order
Citations
PageRank
Mikko Kurimo190893.37
Bowen Zhou22212246.21
Rongqing Huang314110.27
John H. L. Hansen43215365.75