AclNet: efficient end-to-end audio classification CNN

TL;DR

AclNet is an efficient end-to-end CNN for audio classification that achieves high accuracy with significantly reduced computational complexity, enabling energy-efficient, always-on inference.

Contribution

The paper introduces AclNet, a novel CNN architecture optimized for audio classification that balances accuracy and efficiency, outperforming prior models.

Findings

Achieved 85.65% accuracy on ESC-50 dataset.

Reduced model size to 155k parameters with high accuracy.

Enabled energy-efficient, always-on audio classification.

Abstract

We propose an efficient end-to-end convolutional neural network architecture, AclNet, for audio classification. When trained with our data augmentation and regularization, we achieved state-of-the-art performance on the ESC-50 corpus with 85:65% accuracy. Our network allows configurations such that memory and compute requirements are drastically reduced, and a tradeoff analysis of accuracy and complexity is presented. The analysis shows high accuracy at significantly reduced computational complexity compared to existing solutions. For example, a configuration with only 155k parameters and 49:3 million multiply-adds per second is 81:75%, exceeding human accuracy of 81:3%. This improved efficiency can enable always-on inference in energy-efficient platforms.