Improving the Privacy and Practicality of Objective Perturbation for Differentially Private Linear Learners
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2023
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910285215825920 |
|---|---|
| author | Redberg, Rachel Koskela, Antti Wang, Yu-Xiang |
| author_facet | Redberg, Rachel Koskela, Antti Wang, Yu-Xiang |
| contents | In the arena of privacy-preserving machine learning, differentially private stochastic gradient descent (DP-SGD) has outstripped the objective perturbation mechanism in popularity and interest. Though unrivaled in versatility, DP-SGD requires a non-trivial privacy overhead (for privately tuning the model's hyperparameters) and a computational complexity which might be extravagant for simple models such as linear and logistic regression. This paper revamps the objective perturbation mechanism with tighter privacy analyses and new computational tools that boost it to perform competitively with DP-SGD on unconstrained convex generalized linear problems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2401_00583 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Improving the Privacy and Practicality of Objective Perturbation for Differentially Private Linear Learners Redberg, Rachel Koskela, Antti Wang, Yu-Xiang Machine Learning Cryptography and Security In the arena of privacy-preserving machine learning, differentially private stochastic gradient descent (DP-SGD) has outstripped the objective perturbation mechanism in popularity and interest. Though unrivaled in versatility, DP-SGD requires a non-trivial privacy overhead (for privately tuning the model's hyperparameters) and a computational complexity which might be extravagant for simple models such as linear and logistic regression. This paper revamps the objective perturbation mechanism with tighter privacy analyses and new computational tools that boost it to perform competitively with DP-SGD on unconstrained convex generalized linear problems. |
| title | Improving the Privacy and Practicality of Objective Perturbation for Differentially Private Linear Learners |
| topic | Machine Learning Cryptography and Security |
| url | https://arxiv.org/abs/2401.00583 |