Generalizability vs. Counterfactual Explainability Trade-Off

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Veglianti, Fabiano, Giorgi, Flavio, Silvestri, Fabrizio, Tolomei, Gabriele
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912402044354560
author Veglianti, Fabiano
Giorgi, Flavio
Silvestri, Fabrizio
Tolomei, Gabriele
author_facet Veglianti, Fabiano
Giorgi, Flavio
Silvestri, Fabrizio
Tolomei, Gabriele
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2505_23225
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Generalizability vs. Counterfactual Explainability Trade-Off
Veglianti, Fabiano
Giorgi, Flavio
Silvestri, Fabrizio
Tolomei, Gabriele
Machine Learning
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.
title Generalizability vs. Counterfactual Explainability Trade-Off
topic Machine Learning
url https://arxiv.org/abs/2505.23225