Pruning vs XNOR-Net: A Comprehensive Study of Deep Learning for Audio Classification on Edge-devices

TL;DR

This paper compares model compression techniques like pruning, quantization, and XNOR-Net for audio classification on resource-limited edge devices, revealing trade-offs in performance, memory, and computation.

Contribution

It is the first comprehensive empirical study applying XNOR-Net to end-to-end audio classification and comparing it with pruning and quantization methods.

Findings

XNOR-Net achieves comparable accuracy with full precision models for few classes.

Memory reduction of 32-fold and computation reduction of 58-fold with XNOR-Net.

Pruning and quantization outperform XNOR-Net as class count increases, with about 8x more computation.

Abstract

Deep learning has celebrated resounding successes in many application areas of relevance to the Internet of Things (IoT), such as computer vision and machine listening. These technologies must ultimately be brought directly to the edge to fully harness the power of deep learning for the IoT. The obvious challenge is that deep learning techniques can only be implemented on strictly resource-constrained edge devices if the models are radically downsized. This task relies on different model compression techniques, such as network pruning, quantization, and the recent advancement of XNOR-Net. This study examines the suitability of these techniques for audio classification on microcontrollers. We present an application of XNOR-Net for end-to-end raw audio classification and a comprehensive empirical study comparing this approach with pruning-and-quantization methods. We show that raw audio classification with XNOR yields comparable performance to regular full precision networks for small numbers of classes while reducing memory requirements 32-fold and computation requirements 58-fold. However, as the number of classes increases significantly, performance degrades, and pruning-and-quantization based compression techniques take over as the preferred technique being able to satisfy the same space constraints but requiring approximately 8x more computation. We show that these insights are consistent between raw audio classification and image classification using standard benchmark sets. To the best of our knowledge, this is the first study to apply XNOR to end-to-end audio classification and evaluate it in the context of alternative techniques. All codes are publicly available on GitHub.