Precision Scaling of Neural Networks for Efficient Audio Processing
This work addresses efficiency issues for real-time audio processing applications, but it is incremental as it applies existing precision scaling methods to specific audio tasks.
The paper tackles the challenge of high computation and memory demands in deep neural networks for audio processing by scaling precision, finding that lower bit precision reduces processing time by up to 30x with minimal performance impact (<3.14%) in classification tasks like voice-activity detection.
While deep neural networks have shown powerful performance in many audio applications, their large computation and memory demand has been a challenge for real-time processing. In this paper, we study the impact of scaling the precision of neural networks on the performance of two common audio processing tasks, namely, voice-activity detection and single-channel speech enhancement. We determine the optimal pair of weight/neuron bit precision by exploring its impact on both the performance and processing time. Through experiments conducted with real user data, we demonstrate that deep neural networks that use lower bit precision significantly reduce the processing time (up to 30x). However, their performance impact is low (< 3.14%) only in the case of classification tasks such as those present in voice activity detection.