Data Efficient Subset Training with Differential Privacy
Fuente:
arXiv
Salvato in:
| Autori principali: | , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912266952114176 |
|---|---|
| author | Gandhi, Ninad Jayesh Harsha, Moparthy Venkata Subrahmanya Sri |
| author_facet | Gandhi, Ninad Jayesh Harsha, Moparthy Venkata Subrahmanya Sri |
| contents | Private machine learning introduces a trade-off between the privacy budget and training performance. Training convergence is substantially slower and extensive hyper parameter tuning is required. Consequently, efficient methods to conduct private training of models is thoroughly investigated in the literature. To this end, we investigate the strength of the data efficient model training methods in the private training setting. We adapt GLISTER (Killamsetty et al., 2021b) to the private setting and extensively assess its performance. We empirically find that practical choices of privacy budgets are too restrictive for data efficient training in the private setting. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_06732 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Data Efficient Subset Training with Differential Privacy Gandhi, Ninad Jayesh Harsha, Moparthy Venkata Subrahmanya Sri Machine Learning Private machine learning introduces a trade-off between the privacy budget and training performance. Training convergence is substantially slower and extensive hyper parameter tuning is required. Consequently, efficient methods to conduct private training of models is thoroughly investigated in the literature. To this end, we investigate the strength of the data efficient model training methods in the private training setting. We adapt GLISTER (Killamsetty et al., 2021b) to the private setting and extensively assess its performance. We empirically find that practical choices of privacy budgets are too restrictive for data efficient training in the private setting. |
| title | Data Efficient Subset Training with Differential Privacy |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2503.06732 |