Receptive Field Refinement for Convolutional Neural Networks Reliably Improves Predictive Performance

TL;DR

This paper introduces a new receptive field analysis method for CNNs that predicts unproductive layers, enabling automated architecture optimization and leading to significant performance improvements across multiple state-of-the-art models on ImageNet1K.

Contribution

The paper formalizes receptive field expansion analysis, allowing automated identification of unproductive layers to optimize CNN architectures for better efficiency and accuracy.

Findings

Improved ImageNet1K performance across multiple CNN architectures.

Achieved new state-of-the-art results for each evaluated model class.

Method is simple, automated, and applicable to various architectures.

Abstract

Minimal changes to neural architectures (e.g. changing a single hyperparameter in a key layer), can lead to significant gains in predictive performance in Convolutional Neural Networks (CNNs). In this work, we present a new approach to receptive field analysis that can yield these types of theoretical and empirical performance gains across twenty well-known CNN architectures examined in our experiments. By further developing and formalizing the analysis of receptive field expansion in convolutional neural networks, we can predict unproductive layers in an automated manner before ever training a model. This allows us to optimize the parameter-efficiency of a given architecture at low cost. Our method is computationally simple and can be done in an automated manner or even manually with minimal effort for most common architectures. We demonstrate the effectiveness of this approach by increasing parameter efficiency across past and current top-performing CNN-architectures. Specifically, our approach is able to improve ImageNet1K performance across a wide range of well-known, state-of-the-art (SOTA) model classes, including: VGG Nets, MobileNetV1, MobileNetV3, NASNet A (mobile), MnasNet, EfficientNet, and ConvNeXt - leading to a new SOTA result for each model class.