Abstract | ||
---|---|---|
Convolutional Neural Networks (CNNs) are an alternative type of neural network that can be used to reduce spectral variations and model spectral correlations which exist in signals. Since speech signals exhibit both of these properties, CNNs are a more effective model for speech compared to Deep Neural Networks (DNNs). In this paper, we explore applying CNNs to large vocabulary speech tasks. First, we determine the appropriate architecture to make CNNs effective compared to DNNs for LVCSR tasks. Specifically, we focus on how many convolutional layers are needed, what is the optimal number of hidden units, what is the best pooling strategy, and the best input feature type for CNNs. We then explore the behavior of neural network features extracted from CNNs on a variety of LVCSR tasks, comparing CNNs to DNNs and GMMs. We find that CNNs offer between a 13-30% relative improvement over GMMs, and a 4-12% relative improvement over DNNs, on a 400-hr Broadcast News and 300-hr Switchboard task. |
Year | DOI | Venue |
---|---|---|
2013 | 10.1109/ICASSP.2013.6639347 | Acoustics, Speech and Signal Processing |
Keywords | Field | DocType |
correlation methods,neural nets,speech recognition,CNN,DNN,LVCSR tasks,broadcast news,convolutional layers,deep convolutional neural networks,hidden units,large vocabulary continuous speech recognition,pooling strategy,spectral correlations model,spectral variations reduction,speech signals,switchboard task,time 300 hr,time 400 hr,Neural Networks,Speech Recognition | Broadcasting,Pattern recognition,Convolutional neural network,Computer science,Pooling,Speech recognition,Time delay neural network,Artificial intelligence,Artificial neural network,Vocabulary,Deep neural networks | Conference |
ISSN | Citations | PageRank |
1520-6149 | 64 | 2.92 |
References | Authors | |
3 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tara N. Sainath | 1 | 3497 | 232.43 |
Abdel-rahman Mohamed | 2 | 3772 | 266.13 |
B. Kingsbury | 3 | 4175 | 335.43 |
Bhuvana Ramabhadran | 4 | 1779 | 153.83 |