Objective methods for evaluating synthetic intonation
Robert A. Clark, Kurt E Dusterhoff · 1999
This paper studies the overall effect of language modeling on perplexity and word error rate, starting from a trigram model with a standard smoothing method up to complex state-of-theart language models: (1) We compare different smoothing methods, namely linear vs. absolute discounting, interpolation vs. backing-off, and back-off functions based on relative frequencies vs. singleton events.(2) We show the effect of complex language model techniques by using distant-trigrams and automatically selected word classes and word phrases using a maximum likelihood criterion (i.e.minimum perplexity).(3) We show the overall gain of the combined application of the above techniques, as opposed to their separate assessment in past publications.(4) We give perplexity and word error rate results on the North American Business corpus (NAB) with a training text of about 240 million words and on the German Verbmobil corpus.where N() denotes the frequency of the associated word tuple in the training set.Since each event (wn 2 ; wn 1 ; wn) has its own fixed probability, the trigram model is called parameterless.For this reason, the symbol denoting the model parameters is useless and will be dropped for the rest of this paper.For speech recognition tasks with large vocabularies of, say, 20 000 words, the straightforward trigram model results into