Signed Binary Weight Networks

TL;DR

This paper introduces signed-binary networks that combine weight sparsity and binarization to enhance DNN inference efficiency, achieving high sparsity and energy savings while maintaining accuracy.

Contribution

The paper presents a novel signed-binary network method that exploits both sparsity and binarization for improved efficiency without sacrificing accuracy.

Findings

Achieves 69% sparsity on DNNs with comparable accuracy to full-precision models.

Demonstrates real speedup on general-purpose devices.

Reduces energy consumption on ASICs due to high sparsity.

Abstract

Efficient inference of Deep Neural Networks (DNNs) is essential to making AI ubiquitous. Two important algorithmic techniques have shown promise for enabling efficient inference - sparsity and binarization. These techniques translate into weight sparsity and weight repetition at the hardware-software level enabling the deployment of DNNs with critically low power and latency requirements. We propose a new method called signed-binary networks to improve efficiency further (by exploiting both weight sparsity and weight repetition together) while maintaining similar accuracy. Our method achieves comparable accuracy on ImageNet and CIFAR10 datasets with binary and can lead to 69% sparsity. We observe real speedup when deploying these models on general-purpose devices and show that this high percentage of unstructured sparsity can lead to a further reduction in energy consumption on ASICs.