Text Anchor Based Metric Learning for Small-footprint Keyword Spotting
This work addresses the problem of efficient and accurate keyword spotting for small-footprint devices, offering incremental improvements over existing methods.
The paper tackles the challenge of balancing small model size and high accuracy in keyword spotting by proposing text anchors to stabilize metric learning and a new model (LG-Net) combining 1D-CNN and self-attention for better long-term feature capture, achieving state-of-the-art accuracies of 97.67% and 96.79% on two datasets with a lighter version at 96.82% and 95.77% accuracy using only 74k parameters.
Keyword Spotting (KWS) remains challenging to achieve the trade-off between small footprint and high accuracy. Recently proposed metric learning approaches improved the generalizability of models for the KWS task, and 1D-CNN based KWS models have achieved the state-of-the-arts (SOTA) in terms of model size. However, for metric learning, due to data limitations, the speech anchor is highly susceptible to the acoustic environment and speakers. Also, we note that the 1D-CNN models have limited capability to capture long-term temporal acoustic features. To address the above problems, we propose to utilize text anchors to improve the stability of anchors. Furthermore, a new type of model (LG-Net) is exquisitely designed to promote long-short term acoustic feature modeling based on 1D-CNN and self-attention. Experiments are conducted on Google Speech Commands Dataset version 1 (GSCDv1) and 2 (GSCDv2). The results demonstrate that the proposed text anchor based metric learning method shows consistent improvements over speech anchor on representative CNN-based models. Moreover, our LG-Net model achieves SOTA accuracy of 97.67% and 96.79% on two datasets, respectively. It is encouraged to see that our lighter LG-Net with only 74k parameters obtains 96.82% KWS accuracy on the GSCDv1 and 95.77% KWS accuracy on the GSCDv2.