Towards a Shapley Value Graph Framework for Medical peer-influence

TL;DR

This paper introduces a graph-based peer influence framework to enhance interpretability of black-box ML models by analyzing feature interactions and their consequences for interventions in medical AI.

Contribution

It proposes a novel graph framework for understanding feature-to-feature interactions and their impact on interventions, advancing explainability in medical AI models.

Findings

Framework improves interpretability of feature interactions

Enhances guidance for interventions based on explanations

Provides a new perspective on feature influence in AI models

Abstract

eXplainable Artificial Intelligence (XAI) is a sub-field of Artificial Intelligence (AI) that is at the forefront of AI research. In XAI, feature attribution methods produce explanations in the form of feature importance. People often use feature importance as guidance for intervention. However, a limitation of existing feature attribution methods is that there is a lack of explanation towards the consequence of intervention. In other words, although contribution towards a certain prediction is highlighted by feature attribution methods, the relation between features and the consequence of intervention is not studied. The aim of this paper is to introduce a new framework, called a peer influence framework to look deeper into explanations using graph representation for feature-to-feature interactions to improve the interpretability of black-box Machine Learning models and inform intervention.