A Cascade Architecture for Keyword Spotting on Mobile Devices

Abstract

We present a cascade architecture for keyword spotting with speaker verification on mobile devices. By pairing a small computational footprint with specialized digital signal processing (DSP) chips, we are able to achieve low power consumption while continuously listening for a keyword.

31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, CA, USA

Cited by in corpus (2)

• Surprisal-Triggered Conditional Computation with Neural Networks
• Stacked 1D convolutional networks for end-to-end small footprint voice trigger detection