Hierarchical Neural Network Architecture In Keyword Spotting
This addresses the problem of efficient and accurate keyword spotting for ASR systems, but it appears incremental as it applies an existing HNN method to KWS.
The paper tackles the contradiction in Keyword Spotting (KWS) between needing high recall for ASR start signals and low computational complexity for real-time use by implementing a Hierarchical Neural Network (HNN). HNN outperforms traditional DNN and CNN with slightly smaller model size and lower computation, making it easy to deploy on devices.
Keyword Spotting (KWS) provides the start signal of ASR problem, and thus it is essential to ensure a high recall rate. However, its real-time property requires low computation complexity. This contradiction inspires people to find a suitable model which is small enough to perform well in multi environments. To deal with this contradiction, we implement the Hierarchical Neural Network(HNN), which is proved to be effective in many speech recognition problems. HNN outperforms traditional DNN and CNN even though its model size and computation complexity are slightly less. Also, its simple topology structure makes easy to deploy on any device.