MULTI-LEVEL ADAPTIVE NETWORKS / BBC / British Broadcasting Corporation / NVIDIA / GPU / F. Grezl A.E. / /
Country
Italy / /
Facility
P.N. Garner / Theano library / University of Edinburgh / /
IndustryTerm
speech recognition systems / multi-genre media archives using out-of-domain data / adaptive networks / tandem systems / dependent neural network / deep belief networks / Web inventory / conversational telephone speech / large vocabulary speech recognition systems / neural networks / low resource speech recognition applications / context-dependent deep neural networks / discriminative training algorithms / tandem and hybrid deep neural networks / hybrid systems / tandem and hybrid systems / learning algorithm / deep neural networks / /
MarketIndex
set 20 / /
MusicAlbum
Speech / /
Organization
European Association for Machine Translation / University of Edinburgh / ADAPTIVE NETWORKS IN TANDEM AND HYBRID ASR SYSTEMS Peter Bell / Pawel Swietojanski / and Steve Renals Centre for Speech Technology Research / /
Person
TED CTS AMI / /
Position
also performed speaker / speaker / /
Product
DNNs / GeForce GTX 690 GPUs / /
RadioStation
DNN / /
Technology
Spoken Language Technology / speech recognition / neural network / Machine Translation / discriminative training algorithms / speech recognition system / /
MULTI-LEVEL ADAPTIVE NETWORKS IN TANDEM AND HYBRID ASR SYSTEMS Peter Bell, Pawel Swietojanski, and Steve Renals Centre for Speech Technology Research, University of Edinburgh, Edinburgh EH8 9AB, UK ABSTRACT In this paper