Publication Details
Convolutive Bottleneck Network Features for LVCSR
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Grézl František, Ing., Ph.D. (DCGM FIT BUT)
Bottleneck features, Tandem LVCSR system, linear bottleneck, Convolutional Bottleneck Network
Workshop Article about novel features for tandem LVCSR system, which are based on Convolutive Bottleneck Network. It extends the previous work on Universal Context network by using linear bottleneck and expansion to Convolutive Bottleneck Network,
In this paper, we focus on improvements of the bottleneck ANN in a Tandem LVCSR system. First, the influence of training set size and the ANN size is evaluated. Second, a very positive effect of linear bottleneck is shown. Finally a Convolutive Bottleneck Network is proposed as extension of the current stateof- the-art Universal Context Network. The proposed training method leads to 5.5% relative reduction of WER, compared to the Universal Context ANN baseline. The relative improvement compared to the 5-layer single-bottleneck network is 17.7%. The dataset ctstrain07 composed of more than 2000 hours of English Conversational Telephone Speech was used for the experiments. The TNet toolkit with CUDA GPGPU implementation was used for fast training.
@INPROCEEDINGS{FITPUB9763, author = "Karel Vesel\'{y} and Martin Karafi\'{a}t and Franti\v{s}ek Gr\'{e}zl", title = "Convolutive Bottleneck Network Features for LVCSR", pages = "42--47", booktitle = "Proceedings of ASRU 2011", year = 2011, location = "Big Island, Hawaii, US", publisher = "IEEE Signal Processing Society", ISBN = "978-1-4673-0366-8", language = "english", url = "https://www.fit.vut.cz/research/publication/9763" }