Neural Architecture Search For Keyword Spotting

Abstract

Deep neural networks have recently become a popular solution to keyword spotting systems, which enable the control of smart devices via voice. In this paper, we apply neural architecture search to search for convolutional neural network models that can help boost the performance of keyword spotting based on features extracted from acoustic signals while maintaining an acceptable memory footprint. Specifically, we use differentiable architecture search techniques to search for operators and their connections in a predefined cell search space. The found cells are then scaled up in both depth and width to achieve competitive performance. We evaluated the proposed method on Google's Speech Commands Dataset and achieved a state-of-the-art accuracy of over 97% on the setting of 12-class utterance classification commonly reported in the literature.

arXiv commentwill be presented in INTERSPEECH 2020

References in corpus (11)

• Neural Architecture Search with Reinforcement Learning
• Neural Architecture Search: A Survey
• DARTS: Differentiable Architecture Search
• Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition
• Large-Scale Evolution of Image Classifiers
• Depthwise Separable Convolutions for Neural Machine Translation
• Progressive Differentiable Architecture Search: Bridging the Depth Gap between Search and Evaluation
• Temporal Convolution for Real-time Keyword Spotting on Mobile Devices
• Honk: A PyTorch Reimplementation of Convolutional Neural Networks for Keyword Spotting
• MatchboxNet: 1D Time-Channel Separable Convolutional Neural Network Architecture for Speech Commands Recognition
• Small-footprint Keyword Spotting with Graph Convolutional Network

Cited by in corpus (8)

• MicroNets: Neural Network Architectures for Deploying TinyML Applications on Commodity Microcontrollers
• AutoKWS: Keyword Spotting with Differentiable Architecture Search
• Discriminatory and orthogonal feature learning for noise robust keyword spotting
• Ultra-Low Power Keyword Spotting at the Edge
• A Lightweight dynamic filter for keyword spotting
• UniKW-AT: Unified Keyword Spotting and Audio Tagging
• Encoder-Decoder Neural Architecture Optimization for Keyword Spotting
• Partially-Connected Differentiable Architecture Search for Deepfake and Spoofing Detection