Paper Info
Title
Expansion of training texts to generate a topic-dependent language model for meeting speech recognition
Abstract
This paper proposes expansion methods for training texts (baseline) to generate a topic-dependent language model for more accurate recognition of meeting speech. To prepare a universal language model that can cope with the variety of topics discussed in meetings is very difficult. Our strategy is to generate topic-dependent training texts based on two methods. The first is text collection from web pages using queries that consist of topic-dependent confident terms; these terms were selected from preparatory recognition results based on the TF-IDF (TF; Term Frequency, IDF; Inversed Document Frequency) values of each term. The second technique is text generation using participants' names. Our topic-dependent language model was generated using these new texts and the baseline corpus. The language model generated by the proposed strategy reduced the perplexity by 16.4% and out-of-vocabulary rate by 37.5%, respectively, compared with the language model that used only the baseline corpus. This improvement was confirmed through meeting speech recognition as well.
Year
Venue
Keywords
2012
Signal & Information Processing Association Annual Summit and Conference
speech recognition,vocabulary,TF-IDF,Web page,baseline corpus,meeting speech recognition,term frequency inversed document frequency,text collection,text generation technique,topic-dependent confident term query,topic-dependent language model,topic-dependent training text
Field
DocType
ISSN
Speech corpus,Perplexity,Noisy text analytics,Speech analytics,Computer science,Speech recognition,Speaker recognition,Universal language,Artificial intelligence,Natural language processing,Vocabulary,Language model
Conference
2309-9402
ISBN
Citations 
PageRank 
978-1-4673-4863-8
0
0.34
References 
Authors
0
4
Name
Order
Citations
PageRank
Egashira, K.100.34
Kensuke Kojima2123.69
Masaru Yamashita3286.46
Katsuya Yamauchi4111.28