Abstract

In speech recognition domain knowledge is usually implemented by training specialized acoustic and language models. This requires large amounts of training data for the domain. When such data is not accessible there often still is external knowledge, obtained through other means, available that might be used to constrain the search for likely utterances. This paper presents a number of methods to exploit such knowledge; an adaptive language model and a lattice rescoring approach based on Bayesian updating. To decide whether external knowledge is applicable a word level confidence measures is implemented. As a special case of the general problem station­to­station travel frequencies are considered to improve recognition accuracy in a train table dialog system. Experiments are conduced to test and compare the different techniques.