2013 | OriginalPaper | Buchkapitel
The Use of Several Language Models and Its Impact on Word Insertion Penalty in LVCSR
verfasst von : Gregor Donaj, Zdravko Kačič
Erschienen in: Speech and Computer
Verlag: Springer International Publishing
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
This paper investigates the influence of hypothesis length in
N
-best list rescoring. It is theoretically explained why language models prefer shorter hypotheses. This bias impacts on the word insertion penalty used in continuous speech recognition. The theoretical findings are confirmed by experiments. Parameter optimization performed on the Slovene Broadcast News database showed why optimal word insertion penalties tend be greater when two language models are used in speech recognition. This paper also presents a two-pass speech recognition algorithm. Two types of language models were used, a standard trigram word-based language model and a trigram model of morpho-syntactic description tags. A relative decrease of 2.02 % in word error rate was achieved after parameter optimization. Statistical tests were performed to confirm the significance of the word error rate decrease.