ConvMixer: Feature Interactive Convolution with Curriculum Learning for Small Footprint and Noisy Far-field Keyword Spotting

Autor:	Ng, Dianwen, Chen, Yunqi, Tian, Biao, Fu, Qiang, Chng, Eng Siong
Rok vydání:	2022
Předmět:	Computer Science - Sound Electrical Engineering and Systems Science - Audio and Speech Processing
Druh dokumentu:	Working Paper
DOI:	10.1109/ICASSP43922.2022.9747025
Popis:	Building efficient architecture in neural speech processing is paramount to success in keyword spotting deployment. However, it is very challenging for lightweight models to achieve noise robustness with concise neural operations. In a real-world application, the user environment is typically noisy and may also contain reverberations. We proposed a novel feature interactive convolutional model with merely 100K parameters to tackle this under the noisy far-field condition. The interactive unit is proposed in place of the attention module that promotes the flow of information with more efficient computations. Moreover, curriculum-based multi-condition training is adopted to attain better noise robustness. Our model achieves 98.2% top-1 accuracy on Google Speech Command V2-12 and is competitive against large transformer models under the designed noise condition. Comment: submitted to ICASSP 2022
Databáze:	arXiv
Externí odkaz:	http://arxiv.org/abs/2201.05863 Zobrazit plný text záznamu View this record from Arxiv