confounding

Confound-isolating cross-validation approach to control for a confounding effect in a predictive model.

Conquering confounds and covariates: methods, library and guidance

NeurIPS 2024 (spotlight): A Textbook Remedy for Domain Shifts Knowledge Priors for Medical Image Analysis

☯︎[ACMMM'22] Official PyTorch Implementation of Towards Unbiased Visual Emotion Recognition via Causal Intervention

CLI that applies the Clever Materials falsification test to materials-science ML datasets: detects confounding via a bibliographic fingerprint (year/author/journal leaking the target). Built on Jablonka 2026 (arXiv:2602.17730) + Huang&Cole / Nandy datasets (Scientific Data).

Code for a probabilistic sensitivity analysis of an unmeasured confounder

End-to-end empirical implementation of model-based off-policy evaluation and pessimistic policy selection for confounded POMDPs (Hong, Qi & Xu, ICML 2024)

Recurrence risk of remedial actions on real openFDA MAUDE data. A null result, and the target-leakage episode that produced the wrong answer first.

Tools Finding Confounders

Guardrails and causal-epi utilities that operationalize 'ditch the factors associated with' analyses.

Audit whether a medical-ML result measures biology or the lab that produced the data.

Checks whether frozen pathology embeddings reveal which site a slide came from: site probe, k-NN and confound checks

Machine-learning companion to a mechanistic tenocyte model. An audit of what public tendon transcriptomics can actually support: on GSE26051 an anatomical-site baseline (AUC 0.931) outscores the whole transcriptome (0.888) and a mechanistic gene panel (0.864), and nobody reports that baseline.

When Shift Happens - Confounding is to Blame (ICLR 2026)

Confound-matched controls for classifier claims: 1:1 matching, bootstrap AUROC, anchored verdicts, and a refusal when the control is vacuous.