But ASR system for BABEL Surprise evaluation 2014

Author

Karafiat, Martin ; Vesely, Karel ; Szoke, Igor ; Burget, Lukas ; Grezl, Frantisek ; Hannemann, Mirko ; Cernocky, Jan

Author_Institution

IT4I Center of Excellence, Brno Univ. of Technol., Brno, Czech Republic

fYear

2014

Firstpage

501

Lastpage

506

Abstract

The paper describes Brno University of Technology (BUT) ASR system for 2014 BABEL Surprise language evaluation (Tamil). While being largely based on our previous work, two original contributions were brought: (1) speaker-adapted bottle-neck neural network (BN) features were investigated as an input to DNN recognizer and semi-supervised training was found effective. (2) Adding of noise to training data outperformed a classical de-noising technique while dealing with noisy test data was found beneficial, and the performance of this approach was verified on a relatively clean training/test data setup from a different language. All results are reported on BABEL 2014 Tamil data.

Keywords

learning (artificial intelligence); neural nets; signal denoising; speaker recognition; ASR system; BABEL surprise language evaluation; BN features; DNN recognizer; classical denoising technique; deep neural networks; noisy test data; semisupervised training; speaker-adapted bottleneck neural network; Abstracts; Artificial neural networks; Feature extraction; Hidden Markov models; Laboratories; Noise measurement; Training; adaptation of neural networks; bottle-neck neural networks; deep neural networks; discriminative training; noisy speech; speech recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Spoken Language Technology Workshop (SLT), 2014 IEEE

Type

conf

DOI

10.1109/SLT.2014.7078625

Filename

7078625