1 paper
Jeeven Balasubramaniam, Nakul Garg
Low-power speech recognition requires reducing audio acquisition and feature-processing costs as well as neural inference. We present Lyric, a speech-recognition front end that use…