LGMay 29, 2025

Generalizability vs. Counterfactual Explainability Trade-Off

arXiv:2505.23225v1h-index: 20
Originality Highly original
AI Analysis

This work addresses the trade-off between generalization and explainability for machine learning practitioners, offering a theoretical and empirical framework.

The paper investigates the relationship between model generalization and counterfactual explainability, showing that poor generalization increases the ease of generating counterfactual explanations, with empirical validation using ε-VCP as a proxy for overfitting.

In this work, we investigate the relationship between model generalization and counterfactual explainability in supervised learning. We introduce the notion of $\varepsilon$-valid counterfactual probability ($\varepsilon$-VCP) -- the probability of finding perturbations of a data point within its $\varepsilon$-neighborhood that result in a label change. We provide a theoretical analysis of $\varepsilon$-VCP in relation to the geometry of the model's decision boundary, showing that $\varepsilon$-VCP tends to increase with model overfitting. Our findings establish a rigorous connection between poor generalization and the ease of counterfactual generation, revealing an inherent trade-off between generalization and counterfactual explainability. Empirical results validate our theory, suggesting $\varepsilon$-VCP as a practical proxy for quantitatively characterizing overfitting.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes