Framewise Phoneme Classification with Bidirectional LSTM and Other Neural Network Architectures

In this paper, we present bidirectional Long ShortTerm Memory (LSTM) networks, and a modified, full gradientversion of the LSTM learning algorithm. We evaluate bidirectionalLSTM (BLSTM) and several other network architectureson the benchmark task of framewise phoneme classification,using the TIMIT database. Our main findings are that bidirectionalnetworks ou..