Tag
This paper introduces G2I, a two-stage greedy framework that uses explainable GNN explanations to generate actionable intervention hypotheses, improving efficiency over existing mask-based counterfactual methods for applications in public health and social sciences.