TinySpeech: Attention Condensers for Deep Speech Recognition Neural Networks on Edge Devices

Abstract

Advances in deep learning have led to state-of-the-art performance across a multitude of speech recognition tasks. Nevertheless, the widespread deployment of deep neural networks for on-device speech recognition remains a challenge, particularly in edge scenarios where the memory and computing resources are highly constrained (e.g., low-power embedded devices) or where the memory and computing budget dedicated to speech recognition is low (e.g., mobile devices performing numerous tasks besides speech recognition). In this study, we introduce the concept of attention condensers for building low-footprint, highly-efficient deep neural networks for on-device speech recognition on the edge. An attention condenser is a self-attention mechanism that learns and produces a condensed embedding characterizing joint local and cross-channel activation relationships, and performs selective attention accordingly. To illustrate its efficacy, we introduce TinySpeech, low-precision deep neural networks comprising largely of attention condensers tailored for on-device speech recognition using a machine-driven design exploration strategy, with one tailored specifically with microcontroller operation constraints. Experimental results on the Google Speech Commands benchmark dataset for limited-vocabulary speech recognition showed that TinySpeech networks achieved significantly lower architectural complexity (as much as $507\times$ fewer parameters), lower computational complexity (as much as $48\times$ fewer multiply-add operations), and lower storage requirements (as much as $2028\times$ lower weight memory requirements) when compared to previous work. These results not only demonstrate the efficacy of attention condensers for building highly efficient networks for on-device speech recognition, but also illuminate its potential for accelerating deep learning on the edge and empowering TinyML applications.

arXiv comment10 pages

References in corpus (12)

• Neural Machine Translation by Jointly Learning to Align and Translate
• Deep Residual Learning for Image Recognition
• Self-Attention Generative Adversarial Networks
• Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition
• AtomNet: A Deep Convolutional Neural Network for Bioactivity Prediction in Structure-based Drug Discovery
• Achieving Human Parity in Conversational Speech Recognition
• CBAM: Convolutional Block Attention Module
• ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware
• Natural Adversarial Examples
• FermiNets: Learning generative machines to generate efficient neural networks via generative synthesis
• All SMILES Variational Autoencoder
• Honk: A PyTorch Reimplementation of Convolutional Neural Networks for Keyword Spotting

Cited by in corpus (7)

• Machine Learning for Microcontroller-Class Hardware: A Review
• Tiny Machine Learning: Progress and Futures
• MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers
• Hardware Aware Training for Efficient Keyword Spotting on General Purpose and Specialized Hardware
• AttendNets: Tiny Deep Image Recognition Neural Networks for the Edge via Visual Attention Condensers
• AttendSeg: A Tiny Attention Condenser Neural Network for Semantic Segmentation on the Edge
• TinyDefectNet: Highly Compact Deep Neural Network Architecture for High-Throughput Manufacturing Visual Quality Inspection