mlsanityData to 14 Oct 2026

Counterfactual Marginalisation: Framework for Evaluating Robustness to Nuisance Variables

arXiv 2609.10778 ↗First posted 9 Sep 2026cs.LG cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Machine learning models can achieve strong test performance while relying on demographic or acquisition-related shortcuts. We propose counterfactual (CF) marginalisation as a test-time evaluation procedure for assessing robustness of classification models to such variables. Given a CF image generator, we intervene on nuisance parent variables such as age or sex, generate CF versions of each test image, and average predictions over a target intervention distribution. This produces intervention-aware predictions that marginalise demographic effects while preserving patient-specific latent information. We use these predictions to define metrics for CF risk, calibration, stability and worst-case sensitivity. We demonstrate this framework's utility for quantitative robustness evaluation.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 28 references that are in the map. First posted on arXiv 9 Sep 2026.

Builds on

18 of 28 references.

References · 18

  1. 2017On Calibration of Modern Neural Networkscited by 4,960
  2. 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,423
  3. 2020Sharpness-Aware Minimization for Efficiently Improving Generalizationcited by 1,417
  4. 2019Unsupervised Data Augmentation for Consistency Trainingcited by 1,060
  5. 2014Semi-Supervised Learning with Deep Generative Modelscited by 496
  6. 2020Bayesian Deep Learning and a Probabilistic Perspective of Generalizationcited by 480
  7. 2016Ladder Variational Autoencoderscited by 318
  8. 2020Deep Structural Causal Models for Tractable Counterfactual Inferencecited by 211
  9. 2022Metrics reloaded: Recommendations for image analysis validationcited by 195
  10. 2021Reading Race: AI Recognises Patient's Racial Identity In Medical Imagescited by 156
  11. 2022Diffusion Causal Models for Counterfactual Estimationcited by 87
  12. 2017Ensembles of Multiple Models and Architectures for Robust Brain Tumour Segmentationcited by 58
6 more
  1. 2023High Fidelity Image Counterfactuals with Probabilistic Causal Modelscited by 53
  2. 2021Analyzing Overfitting under Class Imbalance in Neural Networks for Image Segmentationcited by 34
  3. 2022Addressing Fairness Issues in Deep Learning-Based Medical Image Analysis: A Systematic Reviewcited by 28
  4. 2024Understanding Test-Time Augmentationcited by 23
  5. 2023No Fair Lunch: A Causal Perspective on Dataset Bias in Machine Learning for Medical Imagingcited by 11
  6. 2024Semi-Supervised Learning for Deep Causal Generative Modelscited by 6

Built on by

Not cited yet.