# When Does Synthetic Data Augmentation Improve Score-Based Imbalanced Classification?

arXiv: https://arxiv.org/abs/2606.26053 · HTML: https://mlsanity.com/paper/2606.26053/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 45 references that are in the map.
- First posted on arXiv 24 Jun 2026.

## Builds on (most cited first, 11 of 11)

- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Synthetic Oversampling: Theory and A Practical Approach Using LLMs to Address Data Imbalance](https://mlsanity.com/paper/2406.03628/) (2024)
- [Conditional Data Synthesis Augmentation](https://mlsanity.com/paper/2504.07426/) (2025)
- [GLM Inference with AI-Generated Synthetic Data Using Misspecified Linear Regression](https://mlsanity.com/paper/2503.21968/) (2025)
- [Bias-Corrected Data Synthesis for Imbalanced Learning](https://mlsanity.com/paper/2510.26046/) (2025)
- [Imbalance-aware Presence-only Loss Function for Species Distribution Modeling](https://mlsanity.com/paper/2403.07472/) (2024)
- [Concentration and excess risk bounds for imbalanced classification with synthetic oversampling](https://mlsanity.com/paper/2510.20472/) (2025)
- [Synthetic Augmentation in Imbalanced Learning: When It Helps, When It Hurts, and How Much to Add](https://mlsanity.com/paper/2601.16120/) (2026)
- [Classification Imbalance as Transfer Learning](https://mlsanity.com/paper/2601.10630/) (2026)
- [Generative Synthetic Data for Causal Inference: Pitfalls, Remedies, and Opportunities](https://mlsanity.com/paper/2604.23904/) (2026)

## Abstract (arXiv)

Synthetic data augmentation is widely used to mitigate class imbalance, but its theoretical effects on score-based classification remain poorly understood. This paper develops a framework for characterizing when synthetic minority augmentation can improve threshold-integrated and threshold-optimized metrics, including AUROC, AUPRC, best-threshold balanced accuracy, and best-threshold (F1) score. We separate the effect of augmentation into two components: a change in effective class weighting and a discrepancy between the synthetic and true minority distributions. Under well-specified score models, the raw estimator already targets the likelihood-ratio ordering, which is population-optimal for the metrics considered. Consequently, augmentation cannot provide a fundamental population-level improvement beyond possible finite-sample variance reduction, and may introduce additional bias through synthetic distributional error. We further establish minimax lower bounds showing that the raw estimator already achieves the optimal metric-regret rate in the well-specified regime. Under misspecification, however, augmentation can play a qualitatively different role: by changing the effective class balance, it can alter the restricted-class projection and correct ranking errors induced by the raw imbalanced objective. We provide explicit improvement bounds quantifying the roles of approximation error, finite-sample estimation error, and synthetic distributional error. Simulation studies corroborate the theory, demonstrating limited gains under well-specification and nontrivial but nonmonotone improvements under misspecification.