Yiyin Zhou

TL;DR: This paper presents an ultra-low-power voice activity detector (VAD) that uses analog signal processing for acoustic feature extraction (AFE) directly on the microphone output, approximate event-driven analog-to-digital conversion (ED-ADC), and digital deep neural network (DNN) for speech/non-speech classification.

TL;DR: A 1μW VAD system utilizing AFE and a digital BNN classifier with an event-encoding A/D interface is presented and the whole AFE is 9.4x more power-efficient than the prior art [5] and 7.9x than the state-of-the-art digital filter bank [6].

TL;DR: It is demonstrated that neural spiking is akin to taking noisy measurements on the stimulus both for time-Varying and space-time-varying stimuli and the quality of the reconstruction degrades gracefully as the threshold variability of the neurons increases.

TL;DR: In this article, the authors used FAPESP Research, Disseminations and Innovation Center for Neuromathematics (CNPq) for postdoctoral research.

TL;DR: In this article, a recurrent neural network (RNN) was used to reconstruct a signal encoded with a time encoding machine (TEM) using a recurrent Neural Network including receiving a TEM-encoded signal, processing the TEM encoded signal, and reconstructing the reconstructed signal with a RNN.