An Interaction-based Convolutional Neural Network (ICNN) Towards Better Understanding of COVID-19 X-ray Images

TL;DR

This paper introduces ICNN, a novel CNN variant that uses influence scores for better interpretability and achieves state-of-the-art accuracy in classifying COVID-19 X-ray images.

Contribution

The paper presents a new interaction-based CNN that leverages influence scores for improved interpretability without losing prediction accuracy.

Findings

Achieved 99.8% accuracy on COVID-19 X-ray classification

Produced a model that is both highly accurate and interpretable

Set a new benchmark for XAI in large-scale image data

Abstract

The field of Explainable Artificial Intelligence (XAI) aims to build explainable and interpretable machine learning (or deep learning) methods without sacrificing prediction performance. Convolutional Neural Networks (CNNs) have been successful in making predictions, especially in image classification. However, these famous deep learning models use tens of millions of parameters based on a large number of pre-trained filters which have been repurposed from previous data sets. We propose a novel Interaction-based Convolutional Neural Network (ICNN) that does not make assumptions about the relevance of local information. Instead, we use a model-free Influence Score (I-score) to directly extract the influential information from images to form important variable modules. We demonstrate that the proposed method produces state-of-the-art prediction performance of 99.8% on a real-world data set classifying COVID-19 Chest X-ray images without sacrificing the explanatory power of the model. This proposed design can efficiently screen COVID-19 patients before human diagnosis, and will be the benchmark for addressing future XAI problems in large-scale data sets.