Regularizing Activation Distribution for Training Binarized Deep Networks

TL;DR

This paper introduces a distribution loss regularization method for training Binarized Neural Networks, improving accuracy and robustness without sacrificing energy efficiency.

Contribution

It proposes a systematic framework for distribution loss regularization that enhances BNN training stability and accuracy.

Findings

Distribution loss improves BNN accuracy.

Regularization enhances training robustness to hyper-parameters.

BNNs maintain energy efficiency with the proposed method.

Abstract

Binarized Neural Networks (BNNs) can significantly reduce the inference latency and energy consumption in resource-constrained devices due to their pure-logical computation and fewer memory accesses. However, training BNNs is difficult since the activation flow encounters degeneration, saturation, and gradient mismatch problems. Prior work alleviates these issues by increasing activation bits and adding floating-point scaling factors, thereby sacrificing BNN's energy efficiency. In this paper, we propose to use distribution loss to explicitly regularize the activation flow, and develop a framework to systematically formulate the loss. Our experiments show that the distribution loss can consistently improve the accuracy of BNNs without losing their energy benefits. Moreover, equipped with the proposed regularization, BNN training is shown to be robust to the selection of hyper-parameters including optimizer and learning rate.