2009 · 58 citations · 6 references
ComputingComputer ScienceComputational LinguisticsNatural Language ProcessingPhonologyLanguage ModelingLanguage ModelsN-gram AnalysisPart-of-speech TaggingConfidence EstimationGainLanguage Model PretrainingModel Confidence EstimationN-gram RepresentationN-gram StatisticsNeural Language ModelsWord N-grams
We measure the effects of a weak language model, estimated from as little as 100k words of text, on unsupervised acoustic model training and then explore the best method of using word confidences to estimate n-gram counts for unsupervised language model training. Even with 100k words of text and 10 hours of training data, unsupervised acoustic modeling is robust, with 50% of the gain recovered when compared to supervised training. For language model training, multiplying the word confidences together to get a weighted count produces the best reduction in WER by 2% over the baseline language model and 0.5% absolute over using unweighted transcripts. Oracle experiments show that a larger gain is possible, but better confidence estimation techniques are needed to identify correct n-grams.
6
Unsupervised Training on Large Amounts of Broadcast News Data
Junxia Ma, Spyros Matsoukas, Owen Kimball et al. · 2006 · 51 citations · Full text
Computing, Arabic, Computer Science +11
Unsupervised versus supervised training of acoustic models
Jeff Ma, Richard Schwartz · 2008 · 50 citations