TY - GEN
T1 - Framewise phoneme classification with bidirectional LSTM networks
AU - Graves, Alex
AU - Schmidhuber, Jürgen
N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14
PY - 2005/12/1
Y1 - 2005/12/1
N2 - In this paper, we apply bidirectional training to a Long Short Term Memory (LSTM) network for the first time. We also present a modified, full gradient version of the LSTM learning algorithm. We discuss the significance of framewise phoneme classification to continuous speech recognition, and the validity of using bidirectional networks for online causal tasks. On the TIMIT speech database, we measure the framewise phoneme classification scores of bidirectional and unidirectional variants of both LSTM and conventional Recurrent Neural Networks (RNNs). We find that bidirectional LSTM outperforms both RNNs and unidirectional LSTM. © 2005 IEEE.
AB - In this paper, we apply bidirectional training to a Long Short Term Memory (LSTM) network for the first time. We also present a modified, full gradient version of the LSTM learning algorithm. We discuss the significance of framewise phoneme classification to continuous speech recognition, and the validity of using bidirectional networks for online causal tasks. On the TIMIT speech database, we measure the framewise phoneme classification scores of bidirectional and unidirectional variants of both LSTM and conventional Recurrent Neural Networks (RNNs). We find that bidirectional LSTM outperforms both RNNs and unidirectional LSTM. © 2005 IEEE.
UR - http://ieeexplore.ieee.org/document/1556215/
UR - http://www.scopus.com/inward/record.url?scp=33750097476&partnerID=8YFLogxK
U2 - 10.1109/IJCNN.2005.1556215
DO - 10.1109/IJCNN.2005.1556215
M3 - Conference contribution
SN - 0780390482
SP - 2047
EP - 2052
BT - Proceedings of the International Joint Conference on Neural Networks
ER -