%Aigaion2 BibTeX export from Idiap Publications %Saturday 21 December 2024 04:38:27 PM @TECHREPORT{Motlicek_Idiap-RR-20-2012, author = {Motlicek, Petr and Garner, Philip N. and Imseng, David and Valente, Fabio}, keywords = {acoustic modeling, Automatic Speech Recognition, Subs-ace Gaussian Mixture Models}, projects = {Idiap}, month = {7}, title = {Application of Subspace Gaussian Mixture Models in Contrastive Acoustic Scenarios}, type = {Idiap-RR}, number = {Idiap-RR-20-2012}, year = {2012}, institution = {Idiap}, address = {Rue Marconi 19, Martigny, Switzerland}, abstract = {This paper describes experimental results of applying Subspace Gaussian Mixture Models (SGMMs) in two completely diverse acoustic scenarios: (a) for Large Vocabulary Continuous Speech Recognition (LVCSR) task over (well-resourced) English meeting data and, (b) for acoustic modeling of underresourced Afrikaans telephone data. In both cases, the performance of SGMM models is compared with a conventional context-dependent HMM/GMM approach exploiting the same kind of information available during the training. LVCSR systems are evaluated on standard NIST Rich Transcription dataset. For under-resourced Afrikaans, SGMM and HMM/GMM acoustic systems are additionally compared to KL-HMM and multilingual Tandem techniques boosted using supplemental out-of-domain data. Experimental results clearly show that the SGMMapproach (having considerably less model parameters) outperforms conventional HMM/GMM system in both scenarios and for all examined training conditions. In case of under-resourced scenario, the SGMM trained only using indomain data is superior to other tested approaches boosted by data from other domain.}, pdf = {https://publications.idiap.ch/attachments/reports/2012/Motlicek_Idiap-RR-20-2012.pdf} }