Modeling Time-Frequency Patterns with LSTM vs. Convolutional Architectures for LVCSR Tasks.

Tara N. Sainath, Bo Li

Browse the full Interspeech paper archive.