TentacleNet: A Pseudo-Ensemble Template for Accurate Binary Convolutional Neural Networks

TL;DR

TentacleNet is a novel binary CNN template that leverages ensemble-inspired parallelization to enhance accuracy while maintaining low memory usage, making binarized models more practical.

Contribution

It introduces TentacleNet, a new end-to-end trainable binary CNN template that improves accuracy through parallelization inspired by ensemble learning, with efficient memory utilization.

Findings

Outperforms classical binary models in accuracy

Achieves significant memory savings compared to binary ensemble methods

Effective on multiple realistic benchmarks

Abstract

Binarization is an attractive strategy for implementing lightweight Deep Convolutional Neural Networks (CNNs). Despite the unquestionable savings offered, memory footprint above all, it may induce an excessive accuracy loss that prevents a widespread use. This work elaborates on this aspect introducing TentacleNet, a new template designed to improve the predictive performance of binarized CNNs via parallelization. Inspired by the ensemble learning theory, it consists of a compact topology that is end-to-end trainable and organized to minimize memory utilization. Experimental results collected over three realistic benchmarks show TentacleNet fills the gap left by classical binary models, ensuring substantial memory savings w.r.t. state-of-the-art binary ensemble methods.