LAB: Learnable Activation Binarizer for Binary Neural Networks

TL;DR

This paper introduces LAB, a learnable activation binarizer that replaces traditional sign() in binary neural networks, enabling more effective information propagation and achieving state-of-the-art performance on ImageNet.

Contribution

The paper proposes a universal, learnable binarization module called LAB that improves BNN performance by replacing sign() with a learnable kernel, adaptable to various architectures.

Findings

LAB improves BNN accuracy on ImageNet

Plugging LAB into existing BNNs boosts performance

LAB-based BNNs are competitive with state-of-the-art methods

Abstract

Binary Neural Networks (BNNs) are receiving an upsurge of attention for bringing power-hungry deep learning towards edge devices. The traditional wisdom in this space is to employ sign() for binarizing featuremaps. We argue and illustrate that sign() is a uniqueness bottleneck, limiting information propagation throughout the network. To alleviate this, we propose to dispense sign(), replacing it with a learnable activation binarizer (LAB), allowing the network to learn a fine-grained binarization kernel per layer - as opposed to global thresholding. LAB is a novel universal module that can seamlessly be integrated into existing architectures. To confirm this, we plug it into four seminal BNNs and show a considerable performance boost at the cost of tolerable increase in delay and complexity. Finally, we build an end-to-end BNN (coined as LAB-BNN) around LAB, and demonstrate that it achieves competitive performance on par with the state-of-the-art on ImageNet.