Wav2Letter: an End-to-End ConvNet-based Speech Recognition System
This work addresses speech recognition for applications requiring simpler, alignment-free models, though it is incremental as it builds on existing end-to-end approaches.
The paper tackles speech recognition by introducing an end-to-end convolutional network model that outputs letters directly from transcribed speech, eliminating the need for phoneme alignment, and achieves competitive word error rates on the Librispeech corpus with MFCC features and promising results from raw waveform.
This paper presents a simple end-to-end model for speech recognition, combining a convolutional network based acoustic model and a graph decoding. It is trained to output letters, with transcribed speech, without the need for force alignment of phonemes. We introduce an automatic segmentation criterion for training from sequence annotation without alignment that is on par with CTC while being simpler. We show competitive results in word error rate on the Librispeech corpus with MFCC features, and promising results from raw waveform.