Publication at Faculty of Mathematics and Physics |
2014
Abstract
This paper presents an extension of the Kaldi automatic speech recognition toolkit to support on-line recognition. The resulting recogniser supports acoustic models trained using state-of-the-art acoustic modelling techniques.
As the recogniser produces word posterior lattices, it is particularly useful in statistical dialogue systems, which try to exploit uncertainty in the recognizer's output. Our experiments show that the on- line recogniser performs significantly better in terms of latency when compared to a cloud-based recogniser.