SoK: A Review of Differentially Private Linear Models For High-Dimensional Data
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914737196892160 |
|---|---|
| author | Khanna, Amol Raff, Edward Inkawhich, Nathan |
| author_facet | Khanna, Amol Raff, Edward Inkawhich, Nathan |
| contents | Linear models are ubiquitous in data science, but are particularly prone to overfitting and data memorization in high dimensions. To guarantee the privacy of training data, differential privacy can be used. Many papers have proposed optimization techniques for high-dimensional differentially private linear models, but a systematic comparison between these methods does not exist. We close this gap by providing a comprehensive review of optimization methods for private high-dimensional linear models. Empirical tests on all methods demonstrate robust and coordinate-optimized algorithms perform best, which can inform future research. Code for implementing all methods is released online. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_01141 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SoK: A Review of Differentially Private Linear Models For High-Dimensional Data Khanna, Amol Raff, Edward Inkawhich, Nathan Machine Learning Cryptography and Security I.2 Linear models are ubiquitous in data science, but are particularly prone to overfitting and data memorization in high dimensions. To guarantee the privacy of training data, differential privacy can be used. Many papers have proposed optimization techniques for high-dimensional differentially private linear models, but a systematic comparison between these methods does not exist. We close this gap by providing a comprehensive review of optimization methods for private high-dimensional linear models. Empirical tests on all methods demonstrate robust and coordinate-optimized algorithms perform best, which can inform future research. Code for implementing all methods is released online. |
| title | SoK: A Review of Differentially Private Linear Models For High-Dimensional Data |
| topic | Machine Learning Cryptography and Security I.2 |
| url | https://arxiv.org/abs/2404.01141 |