Synthetic Data Augmentation for Functional Data: Wasserstein Risk Bounds and Diffusion-Based Conditional Generation
Abstract
Functional data are costly to collect; samples of tens to hundreds of curves are common. Synthetic data augmentation can help, but its theory is developed for finite-dimensional data and relies on total variation, which is uninformative for functions. We study augmentation with a class-conditional functional diffusion generator and state guarantees in the Wasserstein-1 distance. For stable downstream estimators, we bound the risk of training on synthetic curves, from above and below, by the risk of training on as many real curves plus a generation penalty, and derive the best number of synthetic curves and a condition under which synthetic data outperform real data. For allocating synthetic curves under class imbalance, we bound the excess risk of training on real and synthetic curves together, without sample splitting. In applications to temperature, hyperspectral and gait curves, augmentation improved classification, function-to-scalar regression and function-to-function regression, and directing synthetic curves to the minority class gave the largest gain under class imbalance.
Related
- Verifier-Guided Synthetic Augmentation for 3D Human Shape Generation
- On the Reliability of Generative Augmentation: A Wasserstein-Based Theoretical and Empirical Study
- Minimax Rates for Conditional Generative Learning under Wasserstein Loss
- Maximum Likelihood Estimation for Entity Ranking under Iterative Synthetic Data Augmentation
- Intrinsic-Dimensional Wasserstein Guarantees for Private Synthetic Measures
BibTeX
@misc{qi2026synthetic,
title = {{Synthetic Data Augmentation for Functional Data: Wasserstein Risk Bounds and Diffusion-Based Conditional Generation}},
author = {Yuhao Qi and Wenlin Dai and Xuehu Zhu},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.34120},
eprint = {2610.34120},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.34120},
}
RIS
TY - UNPB
TI - Synthetic Data Augmentation for Functional Data: Wasserstein Risk Bounds and Diffusion-Based Conditional Generation
AU - Yuhao Qi
AU - Wenlin Dai
AU - Xuehu Zhu
PY - 2026
DA - 2026/10/25
PB - arXiv
DO - 10.48550/arxiv.2610.34120
UR - https://arxiv.org/abs/2610.34120
AB - Functional data are costly to collect; samples of tens to hundreds of curves are common. Synthetic data augmentation can help, but its theory is developed for finite-dimensional data and relies on total variation, which is uninformative for functions. We study augmentation with a class-conditional functional diffusion generator and state guarantees in the Wasserstein-1 distance. For stable downstream estimators, we bound the risk of training on synthetic curves, from above and below, by the risk of training on as many real curves plus a generation penalty, and derive the best number of synthetic curves and a condition under which synthetic data outperform real data. For allocating synthetic curves under class imbalance, we bound the excess risk of training on real and synthetic curves together, without sample splitting. In applications to temperature, hyperspectral and gait curves, augmentation improved classification, function-to-scalar regression and function-to-function regression, and directing synthetic curves to the minority class gave the largest gain under class imbalance.
ER -