Compression of Acoustic Event Detection Models with Low-rank Matrix Factorization and Quantization Training
This enables deployment of acoustic event detection in resource-constrained applications, but it is incremental as it combines existing compression techniques.
The paper tackled the problem of compressing neural network models for acoustic event detection by combining low-rank matrix factorization and quantization training, reducing a three-layer LSTM model size to 1% with negligible accuracy loss.
In this paper, we present a compression approach based on the combination of low-rank matrix factorization and quantization training, to reduce complexity for neural network based acoustic event detection (AED) models. Our experimental results show this combined compression approach is very effective. For a three-layer long short-term memory (LSTM) based AED model, the original model size can be reduced to 1% with negligible loss of accuracy. Our approach enables the feasibility of deploying AED for resource-constraint applications.