[KSIĄŻKA][B] Efficient processing of deep neural networks

V Sze, YH Chen, TJ Yang, JS Emer - 2020 - Springer
This book provides a structured treatment of the key principles and techniques for enabling
efficient processing of deep neural networks (DNNs). DNNs are currently widely used for …

Recurrent neural networks: An embedded computing perspective

NM Rezk, M Purnaprajna, T Nordström… - IEEE Access, 2020 - ieeexplore.ieee.org
Recurrent Neural Networks (RNNs) are a class of machine learning algorithms used for
applications with time-series and sequential data. Recently, there has been a strong interest …

Vocell: A 65-nm Speech-Triggered Wake-Up SoC for 10- W Keyword Spotting and Speaker Verification

JSP Giraldo, S Lauwereins, K Badami… - IEEE Journal of Solid …, 2020 - ieeexplore.ieee.org
The use of speech-triggered wake-up interfaces has grown significantly in the last few years
for use in ubiquitous and mobile devices. Since these interfaces must always be active …

A 510-nW wake-up keyword-spotting chip using serial-FFT-based MFCC and binarized depthwise separable CNN in 28-nm CMOS

W Shan, M Yang, T Wang, Y Lu, H Cai… - IEEE Journal of Solid …, 2020 - ieeexplore.ieee.org
We propose a sub-μW always-ON keyword spotting (μKWS) chip for audio wake-up
systems. It is mainly composed of a neural network (NN) and a feature extraction (FE) circuit …

14.1 A 510nW 0.41 V low-memory low-computation keyword-spotting chip using serial FFT-based MFCC and binarized depthwise separable convolutional neural …

W Shan, M Yang, J Xu, Y Lu, S Zhang… - … Solid-State Circuits …, 2020 - ieeexplore.ieee.org
Ultra-low power is a strong requirement for always-on speech interfaces in wearable and
mobile devices, such as Voice Activity Detection (VAD) and Keyword Spotting (KWS)[1]-[5] …

A 22nm, 10.8 μ W/15.1 μ W Dual Computing Modes High Power-Performance-Area Efficiency Domained Background Noise Aware Keyword- Spotting Processor

B Liu, H Cai, Z Wang, Y Sun, Z Shen… - … on Circuits and …, 2020 - ieeexplore.ieee.org
This paper proposes a high power-performance-area efficient background noise aware
keyword-spotting (KWS) processor based on an optimized binarized weight network (BWN) …

Tinyvers: A tiny versatile system-on-chip with state-retentive eMRAM for ML inference at the extreme edge

V Jain, S Giraldo, J De Roose, L Mei… - IEEE Journal of Solid …, 2023 - ieeexplore.ieee.org
Extreme edge devices or Internet-of-Things (IoT) nodes require both ultra-low power (ULP)
always-on (AON) processing as well as the ability to do on-demand sampling and …

An 8.93 TOPS/W LSTM recurrent neural network accelerator featuring hierarchical coarse-grain sparsity for on-device speech recognition

D Kadetotad, S Yin, V Berisha… - IEEE Journal of Solid …, 2020 - ieeexplore.ieee.org
Long short-term memory (LSTM) is a type of recurrent neural networks (RNNs), which is
widely used for time-series data and speech applications, due to its high accuracy on such …

Hardware acceleration for embedded keyword spotting: Tutorial and survey

JSP Giraldo, M Verhelst - ACM Transactions on Embedded Computing …, 2021 - dl.acm.org
In recent years, Keyword Spotting (KWS) has become a crucial human–machine interface
for mobile devices, allowing users to interact more naturally with their gadgets by leveraging …

A 148-nW reconfigurable event-driven intelligent wake-up system for AIoT nodes using an asynchronous pulse-based feature extractor and a convolutional neural …

Z Wang, Y Liu, P Zhou, Z Tan, H Fan… - IEEE Journal of Solid …, 2021 - ieeexplore.ieee.org
This article presents a 148-nW always-on wake-up system that drastically reduces the
system power consumption of Internet of Things (IoT) sensor nodes while oftentimes …