Fully Convolutional Speech Recognition
This work addresses speech recognition for general applications by introducing a novel end-to-end convolutional method, though it builds on existing advances in acoustic and language modeling.
The paper tackles speech recognition by proposing a fully convolutional neural network approach that processes raw waveforms end-to-end, eliminating feature extraction steps, and achieves state-of-the-art performance on Wall Street Journal and Librispeech benchmarks, matching or surpassing models like Deep Speech 2 trained with more data.
Current state-of-the-art speech recognition systems build on recurrent neural networks for acoustic and/or language modeling, and rely on feature extraction pipelines to extract mel-filterbanks or cepstral coefficients. In this paper we present an alternative approach based solely on convolutional neural networks, leveraging recent advances in acoustic models from the raw waveform and language modeling. This fully convolutional approach is trained end-to-end to predict characters from the raw waveform, removing the feature extraction step altogether. An external convolutional language model is used to decode words. On Wall Street Journal, our model matches the current state-of-the-art. On Librispeech, we report state-of-the-art performance among end-to-end models, including Deep Speech 2 trained with 12 times more acoustic data and significantly more linguistic data.