A Separable Temporal Convolution Neural Network with Attention for Small-Footprint Keyword Spotting
This work addresses the need for efficient keyword spotting models on mobile devices, offering a significant reduction in parameters while maintaining competitive performance, though it is incremental in nature.
The paper tackled the problem of reducing memory footprint for keyword spotting on mobile devices by proposing a separable temporal convolution neural network with attention, achieving 95.7% accuracy on the Google Speech Commands dataset with only 32.2K parameters, close to the state-of-the-art Res15 model with 239K parameters.
Keyword spotting (KWS) on mobile devices generally requires a small memory footprint. However, most current models still maintain a large number of parameters in order to ensure good performance. To solve this problem, this paper proposes a separable temporal convolution neural network with attention, it has a small number of parameters. Through the time convolution combined with attention mechanism, a small number of parameters model (32.2K) is implemented while maintaining high performance. The proposed model achieves 95.7% accuracy on the Google Speech Commands dataset, which is close to the performance of Res15(239K), the state-of-the-art model in KWS at present.