Smoothly Bounding User Contributions in Differential Privacy

Abstract

A differentially private algorithm guarantees that the input of a single user won’t significantly change the output distribution of the algorithm. When a user contributes more data points, more information can be collected to improve the algorithm’s performance. But at the same time, more noise might need to be added to the algorithm in order to keep the algorithm differentially private and this might hurt the algorithm’s performance. Amin et al. (2019) initiates the study on bounding user contributions and proposes a very natural algorithm which limits the number of samples each user can contribute by a threshold.

Cite

Text

Epasto et al. "Smoothly Bounding User Contributions in Differential Privacy." Neural Information Processing Systems, 2020.

Markdown

[Epasto et al. "Smoothly Bounding User Contributions in Differential Privacy." Neural Information Processing Systems, 2020.](https://mlanthology.org/neurips/2020/epasto2020neurips-smoothly/)

BibTeX

@inproceedings{epasto2020neurips-smoothly,
  title     = {{Smoothly Bounding User Contributions in Differential Privacy}},
  author    = {Epasto, Alessandro and Mahdian, Mohammad and Mao, Jieming and Mirrokni, Vahab and Ren, Lijie},
  booktitle = {Neural Information Processing Systems},
  year      = {2020},
  url       = {https://mlanthology.org/neurips/2020/epasto2020neurips-smoothly/}
}