A Survey on Explainable Artificial Intelligence (XAI): Towards Medical XAI

TL;DR

This paper reviews various interpretability methods in AI, focusing on their application in medicine to enhance transparency and trust in machine decisions, highlighting the importance of explainability in high-stakes fields.

Contribution

It categorizes interpretability approaches and applies this framework to medical AI, aiming to guide clinicians and promote better understanding and education.

Findings

Different interpretability categories reveal varied approaches from simple explanations to complex pattern analysis.

Applying interpretability frameworks to medical AI can improve clinician trust and decision-making.

The review encourages development of mathematically grounded medical education and practices.

Abstract

Recently, artificial intelligence and machine learning in general have demonstrated remarkable performances in many tasks, from image processing to natural language processing, especially with the advent of deep learning. Along with research progress, they have encroached upon many different fields and disciplines. Some of them require high level of accountability and thus transparency, for example the medical sector. Explanations for machine decisions and predictions are thus needed to justify their reliability. This requires greater interpretability, which often means we need to understand the mechanism underlying the algorithms. Unfortunately, the blackbox nature of the deep learning is still unresolved, and many machine decisions are still poorly understood. We provide a review on interpretabilities suggested by different research works and categorize them. The different categories show different dimensions in interpretability research, from approaches that provide "obviously" interpretable information to the studies of complex patterns. By applying the same categorization to interpretability in medical research, it is hoped that (1) clinicians and practitioners can subsequently approach these methods with caution, (2) insights into interpretability will be born with more considerations for medical practices, and (3) initiatives to push forward data-based, mathematically- and technically-grounded medical education are encouraged.