# Preventing dataset shift from breaking machine-learning biomarkers

arXiv: https://arxiv.org/abs/2107.09947 · HTML: https://mlsanity.com/paper/2107.09947/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 93 references that are in the map.
- First posted on arXiv 21 Jul 2021.

## Builds on (most cited first, 15 of 15)

- [Learning Transferable Features with Deep Adaptation Networks](https://mlsanity.com/paper/1502.02791/) (2015)
- [Adversarial Discriminative Domain Adaptation](https://mlsanity.com/paper/1702.05464/) (2017)
- [Out-of-Distribution Generalization via Risk Extrapolation (REx)](https://mlsanity.com/paper/2003.00688/) (2020)
- [Return of Frustratingly Easy Domain Adaptation](https://mlsanity.com/paper/1511.05547/) (2015)
- [Fairness Without Demographics in Repeated Loss Minimization](https://mlsanity.com/paper/1806.08010/) (2018)
- [Learning Models with Uniform Performance via Distributionally Robust Optimization](https://mlsanity.com/paper/1810.08750/) (2018)
- [Data Augmentation Generative Adversarial Networks](https://mlsanity.com/paper/1711.04340/) (2017)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [Emergence of Invariance and Disentanglement in Deep Representations](https://mlsanity.com/paper/1706.01350/) (2017)
- [A review of domain adaptation without target labels](https://mlsanity.com/paper/1901.05335/) (2019)
- [Variance-based regularization with convex objectives](https://mlsanity.com/paper/1610.02581/) (2016)
- [Distributionally Robust Losses for Latent Covariate Mixtures](https://mlsanity.com/paper/2007.13982/) (2020)
- [Machine Learning with Multi-Site Imaging Data: An Empirical Study on the Impact of Scanner Effects](https://mlsanity.com/paper/1910.04597/) (2019)
- [Risk of Training Diagnostic Algorithms on Data with Demographic Bias](https://mlsanity.com/paper/2005.10050/) (2020)
- [Kernel Distributionally Robust Optimization](https://mlsanity.com/paper/2006.06981/) (2020)

## Built on by (most cited first, 15 of 15)

- [Generative models improve fairness of medical classifiers under distribution shifts](https://mlsanity.com/paper/2304.09218/) (2023)
- [How Robust is Unsupervised Representation Learning to Distribution Shift?](https://mlsanity.com/paper/2206.08871/) (2022)
- [Geodesic Optimization for Predictive Shift Adaptation on EEG data](https://mlsanity.com/paper/2407.03878/) (2024)
- [Statistical quantification of confounding bias in predictive modelling](https://mlsanity.com/paper/2111.00814/) (2021)
- [Advances in Automated Fetal Brain MRI Segmentation and Biometry: Insights from the FeTA 2024 Challenge](https://mlsanity.com/paper/2505.02784/) (2025)
- [Deployment of Image Analysis Algorithms under Prevalence Shifts](https://mlsanity.com/paper/2303.12540/) (2023)
- [FetMRQC: a robust quality control system for multi-centric fetal brain MRI](https://mlsanity.com/paper/2311.04780/) (2023)
- [Physics-informed and Unsupervised Riemannian Domain Adaptation for Machine Learning on Heterogeneous EEG Datasets](https://mlsanity.com/paper/2403.15415/) (2024)
- [Evaluating Synthetic Data Generation for Domain Generalization in Fetal Brain MRI Segmentation](https://mlsanity.com/paper/2411.06842/) (2024)
- [Improving generalization of machine learning-identified biomarkers with causal modeling: an investigation into immune receptor diagnostics](https://mlsanity.com/paper/2204.09291/) (2022)
- [Automatic quality control in multi-centric fetal brain MRI super-resolution reconstruction](https://mlsanity.com/paper/2503.10156/) (2025)
- [Tracking the Discriminative Axis: Dual Prototypes for Test-Time OOD Detection Under Covariate Shift](https://mlsanity.com/paper/2603.15213/) (2026)
- [Development and evaluation of CADe systems in low-prevalence setting: The RARE25 challenge for early detection of Barrett's neoplasia](https://mlsanity.com/paper/2604.11171/) (2026)
- [Prior-matched evaluation of operational Earth-observation classifiers: a three-number reporting method demonstrated on Sentinel-1 internal-wave detection](https://mlsanity.com/paper/2607.07146/) (2026)
- [When Participant Disjoint Validation Is Not Enough to Establish Generalization in Parkinson's Gait Classification](https://mlsanity.com/paper/2610.16220/) (2026)

## Abstract (arXiv)

Machine learning brings the hope of finding new biomarkers extracted from cohorts with rich biomedical measurements. A good biomarker is one that gives reliable detection of the corresponding condition. However, biomarkers are often extracted from a cohort that differs from the target population. Such a mismatch, known as a dataset shift, can undermine the application of the biomarker to new individuals. Dataset shifts are frequent in biomedical research, e.g. because of recruitment biases. When a dataset shift occurs, standard machine-learning techniques do not suffice to extract and validate biomarkers. This article provides an overview of when and how dataset shifts breaks machine-learning extracted biomarkers, as well as detection and correction strategies.