MatchboxNet: 1D Time-Channel Separable Convolutional Neural Network Architecture for Speech Commands Recognition

Abstract

We present an MatchboxNet - an end-to-end neural network for speech command recognition. MatchboxNet is a deep residual network composed from blocks of 1D time-channel separable convolution, batch-normalization, ReLU and dropout layers. MatchboxNet reaches state-of-the-art accuracy on the Google Speech Commands dataset while having significantly fewer parameters than similar models. The small footprint of MatchboxNet makes it an attractive candidate for devices with limited computational resources. The model is highly scalable, so model accuracy can be improved with modest additional memory and compute. Finally, we show how intensive data augmentation using an auxiliary noise dataset improves robustness in the presence of background noise.

References in corpus (6)

• Improved Regularization of Convolutional Neural Networks with Cutout
• Depthwise Separable Convolutions for Neural Machine Translation
• NeMo: a toolkit for building AI applications using Neural Modules
• QuartzNet: Deep Automatic Speech Recognition with 1D Time-Channel Separable Convolutions
• On Feature Normalization and Data Augmentation
• Training Keyword Spotters with Limited and Synthesized Speech Data

Cited by in corpus (7)

• Streaming keyword spotting on mobile devices
• Neural Architecture Search For Keyword Spotting
• AST: Audio Spectrogram Transformer
• MarbleNet: Deep 1D Time-Channel Separable Convolutional Neural Network for Voice Activity Detection
• Encoder-Decoder Neural Architecture Optimization for Keyword Spotting
• An Integrated Framework for Two-pass Personalized Voice Trigger
• Attention-Free Keyword Spotting