From Hashing to CNNs: Training BinaryWeight Networks via Hashing

TL;DR

This paper introduces BWNH, a novel method for training Binary Weight Networks using hashing techniques, significantly reducing memory and computational requirements while maintaining high performance on standard datasets.

Contribution

The paper reveals the connection between hashing and binary weight networks and proposes an alternating optimization approach to improve training effectiveness.

Findings

BWNH outperforms state-of-the-art methods on CIFAR10, CIFAR100, and ImageNet.

Training binary weights as a hashing problem improves accuracy and efficiency.

Extensive experiments validate the effectiveness of the proposed approach.

Abstract

Deep convolutional neural networks (CNNs) have shown appealing performance on various computer vision tasks in recent years. This motivates people to deploy CNNs to realworld applications. However, most of state-of-art CNNs require large memory and computational resources, which hinders the deployment on mobile devices. Recent studies show that low-bit weight representation can reduce much storage and memory demand, and also can achieve efficient network inference. To achieve this goal, we propose a novel approach named BWNH to train Binary Weight Networks via Hashing. In this paper, we first reveal the strong connection between inner-product preserving hashing and binary weight networks, and show that training binary weight networks can be intrinsically regarded as a hashing problem. Based on this perspective, we propose an alternating optimization method to learn the hash codes instead of directly learning binary weights. Extensive experiments on CIFAR10, CIFAR100 and ImageNet demonstrate that our proposed BWNH outperforms current state-of-art by a large margin.