LANGUAGE MODELS FOR AUTOMATIC SPEECH RECOGNITION OF CZECH LECTURES
暂无分享,去创建一个
This paper describes improvements in Automatic Speech Recognition (ASR) of Czech lectures obtained by enhancing language models. Our baseline is a statistical trigram language model with Good-Turing smoothing, trained on half billion words from newspapers, books etc. The overall improvement from adding more training data is over 10% in accuracy absolute, while using advanced language modeling techniques mainly neural networks yields another 3%. Perplexity improvements and OOV reduction are discussed too.
[1] Klaus A J Riederer. 1 LARGE VOCABULARY CONTINUOUS SPEECH RECOGNITION , 2000 .
[2] Yoshua Bengio,et al. A Neural Probabilistic Language Model , 2003, J. Mach. Learn. Res..
[3] Andreas Stolcke,et al. SRILM - an extensible language modeling toolkit , 2002, INTERSPEECH.
[4] Joshua Goodman,et al. The State of the Art in Language Modeling , 2003, NAACL.
[5] Jean-Luc Gauvain,et al. Training Neural Network Language Models on Very Large Corpora , 2005, HLT.