# A Unifying Causal Framework for Analyzing Dataset Shift-stable Learning Algorithms

arXiv: https://arxiv.org/abs/1905.11374 · HTML: https://mlsanity.com/paper/1905.11374/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 May 2019.

## Built on by (most cited first, 17 of 17)

- [Contrastive ACE: Domain Generalization Through Alignment of Causal Mechanisms](https://mlsanity.com/paper/2106.00925/) (2021)
- [Data Augmentations for Improved (Large) Language Model Generalization](https://mlsanity.com/paper/2310.12803/) (2023)
- [Learning Counterfactually Invariant Predictors](https://mlsanity.com/paper/2207.09768/) (2022)
- [Why Do Probabilistic Clinical Models Fail To Transport Between Sites?](https://mlsanity.com/paper/2311.04787/) (2023)
- [Malign Overfitting: Interpolation Can Provably Preclude Invariance](https://mlsanity.com/paper/2211.15724/) (2022)
- [Open-Source Drift Detection Tools in Action: Insights from Two Use Cases](https://mlsanity.com/paper/2404.18673/) (2024)
- [Robust Learning via Conditional Prevalence Adjustment](https://mlsanity.com/paper/2310.15766/) (2023)
- [Adapting to Shifting Correlations with Unlabeled Data Calibration](https://mlsanity.com/paper/2409.05996/) (2024)
- [A New Perspective On AI Safety Through Control Theory Methodologies](https://mlsanity.com/paper/2506.23703/) (2025)
- [Toward Calibrated Mixture-of-Experts Under Distribution Shift](https://mlsanity.com/paper/2606.20544/) (2026)
- [Repeated Environment Inference for Invariant Learning](https://mlsanity.com/paper/2207.12876/) (2022)
- [AI for Anticipatory Action: Moving Beyond Climate Forecasting](https://mlsanity.com/paper/2307.15727/) (2023)
- [A causal viewpoint on prediction model performance under changes in case-mix: discrimination and calibration respond differently for prognosis and diagnosis predictions](https://mlsanity.com/paper/2409.01444/) (2024)
- [Considerations for Distribution Shift Robustness of Diagnostic Models in Healthcare](https://mlsanity.com/paper/2410.19575/) (2024)
- [An Expectation-Maximization Algorithm for Domain Adaptation in Gaussian Causal Models](https://mlsanity.com/paper/2601.03459/) (2026)
- [A Causal Framework for Mitigating Data Shifts in Healthcare](https://mlsanity.com/paper/2603.13595/) (2026)
- [Prediction-Intervention Games and Invariant Sets](https://mlsanity.com/paper/2605.16828/) (2026)

## Abstract (arXiv)

Recent interest in the external validity of prediction models (i.e., the problem of different train and test distributions, known as dataset shift) has produced many methods for finding predictive distributions that are invariant to dataset shifts and can be used for prediction in new, unseen environments. However, these methods consider different types of shifts and have been developed under disparate frameworks, making it difficult to theoretically analyze how solutions differ with respect to stability and accuracy. Taking a causal graphical view, we use a flexible graphical representation to express various types of dataset shifts. Given a known graph of the data generating process, we show that all invariant distributions correspond to a causal hierarchy of graphical operators which disable the edges in the graph that are responsible for the shifts. The hierarchy provides a common theoretical underpinning for understanding when and how stability to shifts can be achieved, and in what ways stable distributions can differ. We use it to establish conditions for minimax optimal performance across environments, and derive new algorithms that find optimal stable distributions. Using this new perspective, we empirically demonstrate that that there is a tradeoff between minimax and average performance.