Training Competitive Binary Neural Networks from Scratch

TL;DR

This paper introduces a simple training method for binary neural networks that achieves state-of-the-art results without relying on prior knowledge or complex strategies, including novel dense connection architectures.

Contribution

It presents a new approach to training binary neural networks from scratch with a simpler method and introduces dense connections for binary models, improving performance.

Findings

Achieved state-of-the-art results on benchmark datasets.

First to successfully adopt dense connections in binary networks.

Demonstrated that simpler training strategies can match complex methods.

Abstract

Convolutional neural networks have achieved astonishing results in different application areas. Various methods that allow us to use these models on mobile and embedded devices have been proposed. Especially binary neural networks are a promising approach for devices with low computational power. However, training accurate binary models from scratch remains a challenge. Previous work often uses prior knowledge from full-precision models and complex training strategies. In our work, we focus on increasing the performance of binary neural networks without such prior knowledge and a much simpler training strategy. In our experiments we show that we are able to achieve state-of-the-art results on standard benchmark datasets. Further, to the best of our knowledge, we are the first to successfully adopt a network architecture with dense connections for binary networks, which lets us improve the state-of-the-art even further.