Abstract
Private Evolution (PE) is a differentially private algorithm for synthetic data generation. While it can be viewed as a Wasserstein learning algorithm, it performs much better in practice than worst-case Wasserstein analyses would predict. We recast PE as generative model-augmented Wasserstein learning. We show theoretically that when we take into account the use of a generative model that is able to capture something about the true distribution, then we can obtain much better performance bounds. For example, if the generator gives samples in the same low-dimensional space as the distribution, then sample complexity depends on intrinsic, not ambient, dimension. We also show that standard variants of PE can fail to converge on simple well-clustered instances, and propose a new geometry-aware version of PE with provable convergence on such instances. Experimentally, we show that our new algorithm is competitive with standard baselines and can improve recall.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
McMillan, A., Talwar, K., & Zhou, F. (2026). Understanding Private Evolution as Learning-Augmented Clustering. https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering
MLA 9
McMillan, Audra, et al. "Understanding Private Evolution as Learning-Augmented Clustering." https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering.
Chicago (author–date)
McMillan, Audra, Kunal Talwar, and Felix Zhou. 2026. "Understanding Private Evolution as Learning-Augmented Clustering." https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering.
Harvard
McMillan, A., Talwar, K. and Zhou, F. (2026) 'Understanding Private Evolution as Learning-Augmented Clustering', Available at: https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering.
Vancouver
McMillan A, Talwar K, Zhou F. Understanding Private Evolution as Learning-Augmented Clustering. https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering
IEEE
A. McMillan, K. Talwar, and F. Zhou, "Understanding Private Evolution as Learning-Augmented Clustering," https://omanscience.com/en/articles/understanding-private-evolution-as-learning-augmented-clustering.