Framewise phoneme classification with bidirectional LSTM and other neural network architectures
Framewise phoneme classification with bidirectional LSTM and other neural network architectures
复制标题
DOI：
10.1016/j.neunet.2005.06.042
复制
发表时间：
2005-06-01
期刊：
影响因子：
7.8
通讯作者：
Schmidhuber, J
中科院分区：
文献类型：
--
作者：
Graves, A;Schmidhuber, J
在本文中，我们提出了双向长短期记忆（LSTM）网络，以及LSTM学习算法的修改后的全梯度版本。我们使用TIMIT数据库，在框架音素分类的基准任务上评估了双向LSTM（BLSTM）和其他几种网络架构。我们的主要发现是双向网络优于单向网络，长短期记忆（LSTM）比标准的递归神经网络（RNN）和时间窗口多层感知器（MLP）更快，也更准确。我们的研究结果支持这样的观点，即上下文信息是至关重要的语音处理，并建议BLSTM是一个有效的架构，利用它。（c）2005年爱思唯尔有限公司保留所有权利。
In this paper, we present bidirectional Long Short Term Memory (LSTM) networks, and a modified, full gradient version of the LSTM learning algorithm. We evaluate Bidirectional LSTM (BLSTM) and several other network architectures on the benchmark task of framewise phoneme classification, using the TIMIT database. Our main findings are that bidirectional networks outperform unidirectional ones, and Long Short Term Memory (LSTM) is much faster and also more accurate than both standard Recurrent Neural Nets (RNNs) and time-windowed Multilayer Perceptrons (MLPs). Our results support the view that contextual information is crucial to speech processing, and suggest that BLSTM is an effective architecture with which to exploit it. (c) 2005 Elsevier Ltd. All rights reserved.