Saved in:
Bibliographic Details
Main Authors: Monteiro, Adriana Laurindo, Loubes, Jean-Michel
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.15867
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914498684649472
author Monteiro, Adriana Laurindo
Loubes, Jean-Michel
author_facet Monteiro, Adriana Laurindo
Loubes, Jean-Michel
contents The growing use of Machine Learning (ML) tools comes with critical challenges, such as limited model explainability. We propose a global explainability framework that leverages Optimal Transport and Distributionally Robust Optimization to analyze how ML algorithms respond to constrained data perturbations. Our approach enforces constraints on feature-level statistics (e.g., brightness, age distribution), generating realistic perturbations that preserve semantic structure. We provide a model-agnostic diagnostic bench that applies to both tabular and image domains with solid theoretical guarantees. We validate the approach on real-world datasets providing interpretable robustness diagnostics that complement standard evaluation and fairness auditing tools.
format Preprint
id arxiv_https___arxiv_org_abs_2603_15867
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Evaluating Black-Box Vulnerabilities with Wasserstein-Constrained Data Perturbations
Monteiro, Adriana Laurindo
Loubes, Jean-Michel
Machine Learning
The growing use of Machine Learning (ML) tools comes with critical challenges, such as limited model explainability. We propose a global explainability framework that leverages Optimal Transport and Distributionally Robust Optimization to analyze how ML algorithms respond to constrained data perturbations. Our approach enforces constraints on feature-level statistics (e.g., brightness, age distribution), generating realistic perturbations that preserve semantic structure. We provide a model-agnostic diagnostic bench that applies to both tabular and image domains with solid theoretical guarantees. We validate the approach on real-world datasets providing interpretable robustness diagnostics that complement standard evaluation and fairness auditing tools.
title Evaluating Black-Box Vulnerabilities with Wasserstein-Constrained Data Perturbations
topic Machine Learning
url https://arxiv.org/abs/2603.15867