Bounding User Contributions: A Bias-Variance Trade-off in Differential Privacy

By Kareem Amin, Alex Kulesza, Andrés Muñoz Medina, Sergei Vassilvitskii. 2019

78 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Privacy & Security

Citations in the corpus

Advances and Open Problems in Federated Learning (2019, neutral): “Hence, similar to recent works on user-level DP in the central model [24], figuring out how to adaptively bound the contributions of users and clip the model parameters remains an interesting research direction [381, 324].”
Learning discrete distributions: user vs item-level privacy (2020, background): “The problem of bounding user contributions in user-level privacy in the context of both histogram estimation and learning machine learning models was studied in Amin et al. [2019]. Differentially private SQL with bounded user contributions was proposed in Wilson et al.”
Differentially Private Learning with Adaptive Clipping (2019, positive): “This effect was observed empirically by McMahan et al. (2018), and the clipping bias-variance tradeoff is theoretically analyzed and shown to be an inherent property of differentially private learning by Amin et al. (2019).”
A Comprehensive Survey of Privacy-preserving Federated Learning (2021, methodology): “To limit the privacy leakage from the updates or the final model, a user-level differential privacy can be applied during the FL training process [6, 44, 83].”
Differentially Private SQL with Bounded User Contribution (2019, background): “In this section we explore the relationship between utility and various parameters, which must be adjusted to balance privacy and utility [1].”
Characterizing Private Clipped Gradient Descent on Convex Generalized Linear Problems (2020, unknown): “There has been both theoretical [18, 7, 26, 41, 4] and empirical research [40, 33] providing algorithms which can be used for choosing a “near-optimal” value of the clipping norm.”
How to DP-fy ML: A Practical Guide to Machine Learning with Differential Privacy (2023, background): “The choice of the range is a critical parameter, and leads to a bias-variance trade-off (Amin et al., 2019).”