Learning Multiscale Features Directly From Waveforms

TL;DR

This paper introduces a multiscale convolutional approach to learn features directly from waveforms, significantly improving speech recognition performance over traditional spectral methods.

Contribution

It proposes a multiscale convolutional method for end-to-end waveform feature learning, surpassing spectral representations in speech recognition accuracy.

Findings

20.7% relative reduction in word error rate

Multiscale learning improves efficiency and accuracy

Enhanced temporal and frequency resolution benefits recognition

Abstract

Deep learning has dramatically improved the performance of speech recognition systems through learning hierarchies of features optimized for the task at hand. However, true end-to-end learning, where features are learned directly from waveforms, has only recently reached the performance of hand-tailored representations based on the Fourier transform. In this paper, we detail an approach to use convolutional filters to push past the inherent tradeoff of temporal and frequency resolution that exists for spectral representations. At increased computational cost, we show that increasing temporal resolution via reduced stride and increasing frequency resolution via additional filters delivers significant performance improvements. Further, we find more efficient representations by simultaneously learning at multiple scales, leading to an overall decrease in word error rate on a difficult internal speech test set by 20.7% relative to networks with the same number of parameters trained on spectrograms.