Wednesday, October 7th 2015
Title: Neural Networks in Speech Recognition for Acoustic Modeling of Tonal Language
The baseline system of an automatic speech recognition (ASR) normally uses Mel- Frequency Cepstral Coefficients (MFCC) as feature vectors. However, for tonal language like Thai, tone information is one of the important features which can be used to improve the accuracy of recognition. This topic related to a method of building an acoustic model for Thai-ASR using a combination of MFCC and tone information as an input feature vector. In addition, Artificial Neural Network (ANN) multilayer perceptrons is appled to estimate the posterior probabilities of a class model given a sequence of observation input. The performance of the ANN approach is compared with the Gaussian Mixture Model (GMM) with the Hidden Markov Model Toolkit (HTK). The experiments were carried out with 2-grams and 3-grams of a language model. The training and test data sets were recorded from male and female speakers. The results showed that the combination method for ANN input can be used to improve the performance of Thai-ASR in terms of reducing word error rate.
Colloquium coordinators are Prof.dr. M. Aiello (e-mail :
Prof.dr. M. Biehl (e-mail:
This year, the University of Groningen has submitted four research projects to compete for the national Klokhuis Science Prize. The aim of this prize is to introduce a young and wide audience to academic research. The winning project will be...
Het project WIJS, het initiatief dat de Gemeente Groningen en de Hanzehogeschool in 2014 startten, is uitgebreid met vier partners: WIJ-Groningen, de Rijksuniversiteit Groningen, Alfa-college en Noorderpoort. Vandaag wordt de komst van de vier...
Non-executive directors (hereafter: directors) have to take a critical stance towards the top managers they supervise. This has been the dominant perspective among researchers and the media after the financial crisis of 2008 and recent major...