Demystifying and Generalizing BinaryConnect
This work addresses the problem of neural network quantization for efficient deployment, offering a theoretical generalization that is incremental but clarifies and extends prior methods.
The paper tackled the limited understanding of BinaryConnect and its variations by unifying existing quantization algorithms, proposing proximal maps as a natural quantizer family, and introducing ProxConnect as a generalization with proven convergence, achieving competitive performance on CIFAR-10 and ImageNet.
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.