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.
This paper develops constructive approximation and learning guarantees for shallow neural models with infinite-dimensional inputs, separating errors into coordinate-truncation, network width, and sample size components for a unified theoretical analysis.
Proves a tight approximation ratio for the greedy algorithm in myopic Bayesian active learning for linear regression, identifying the maximum initial leverage score as a key quantity.