Continual Release Moment Estimation with Differential Privacy

Abstract

We propose *Joint Moment Estimation* (JME), a method for continually and privately estimating both the first and second moments of a data stream with reduced noise compared to naive approaches. JME supports the *matrix mechanism* and exploits a joint sensitivity analysis to identify a privacy regime in which the second-moment estimation incurs no additional privacy cost, thereby improving accuracy while maintaining privacy. We demonstrate JME’s effectiveness in two applications: estimating the running mean and covariance matrix for Gaussian density estimation and model training with DP-Adam.

Cite

Text

Kalinin et al. "Continual Release Moment Estimation with Differential Privacy." Advances in Neural Information Processing Systems, 2025.

Markdown

[Kalinin et al. "Continual Release Moment Estimation with Differential Privacy." Advances in Neural Information Processing Systems, 2025.](https://mlanthology.org/neurips/2025/kalinin2025neurips-continual/)

BibTeX

@inproceedings{kalinin2025neurips-continual,
  title     = {{Continual Release Moment Estimation with Differential Privacy}},
  author    = {Kalinin, Nikita and Upadhyay, Jalaj and Lampert, Christoph H.},
  booktitle = {Advances in Neural Information Processing Systems},
  year      = {2025},
  url       = {https://mlanthology.org/neurips/2025/kalinin2025neurips-continual/}
}