Demystifying and Generalizing BinaryConnect

TL;DR

This paper investigates the theoretical foundations of BinaryConnect, revealing its connections to dual averaging and conditional gradient algorithms, and introduces ProxConnect as a generalized framework with competitive experimental results.

Contribution

It clarifies the inner workings of BinaryConnect, proposes ProxConnect as a natural generalization, and provides convergence analysis and empirical validation.

Findings

Existing quantization algorithms are surprisingly similar.

ProxConnect achieves competitive performance on CIFAR-10 and ImageNet.

ProxConnect is a natural extension of BinaryConnect with proven convergence.

Abstract

BinaryConnect (BC) and its many variations have become the de facto standard for neural network quantization. However, our understanding of the inner workings of BC is still quite limited. We attempt to close this gap in four different aspects: (a) we show that existing quantization algorithms, including post-training quantization, are surprisingly similar to each other; (b) we argue for proximal maps as a natural family of quantizers that is both easy to design and analyze; (c) we refine the observation that BC is a special case of dual averaging, which itself is a special case of the generalized conditional gradient algorithm; (d) consequently, we propose ProxConnect (PC) as a generalization of BC and we prove its convergence properties by exploiting the established connections. We conduct experiments on CIFAR-10 and ImageNet, and verify that PC achieves competitive performance.