IEA: Inner Ensemble Average within a convolutional neural network

TL;DR

This paper introduces Inner Ensemble Average (IEA), a novel method that replaces single convolutional layers with ensemble averages of multiple layers within CNNs, leading to improved accuracy.

Contribution

The paper proposes IEA, a new approach to embed ensemble learning directly inside CNN layers, enhancing model performance without increasing inference complexity.

Findings

IEA-based CNNs outperform standard CNNs on benchmark datasets.

Feature analysis shows IEA improves feature diversity and similarity scores.

Empirical results confirm the effectiveness of IEA in various tasks.

Abstract

Ensemble learning is a method of combining multiple trained models to improve model accuracy. We propose the usage of such methods, specifically ensemble average, inside Convolutional Neural Network (CNN) architectures by replacing the single convolutional layers with Inner Average Ensembles (IEA) of multiple convolutional layers. Empirical results on different benchmarking datasets show that CNN models using IEA outperform those with regular convolutional layers. A visual and a similarity score analysis of the features generated from IEA explains why it boosts the model performance.