Straggler-Resilient Differentially-Private Decentralized Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2022
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913406801412096 |
|---|---|
| author | Yakimenka, Yauhen Weng, Chung-Wei Lin, Hsuan-Yin Rosnes, Eirik Kliewer, Jörg |
| author_facet | Yakimenka, Yauhen Weng, Chung-Wei Lin, Hsuan-Yin Rosnes, Eirik Kliewer, Jörg |
| contents | We consider the straggler problem in decentralized learning over a logical ring while preserving user data privacy. Especially, we extend the recently proposed framework of differential privacy (DP) amplification by decentralization by Cyffers and Bellet to include overall training latency--comprising both computation and communication latency. Analytical results on both the convergence speed and the DP level are derived for both a skipping scheme (which ignores the stragglers after a timeout) and a baseline scheme that waits for each node to finish before the training continues. A trade-off between overall training latency, accuracy, and privacy, parameterized by the timeout of the skipping scheme, is identified and empirically validated for logistic regression on a real-world dataset and for image classification using the MNIST and CIFAR-10 datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2212_03080 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Straggler-Resilient Differentially-Private Decentralized Learning Yakimenka, Yauhen Weng, Chung-Wei Lin, Hsuan-Yin Rosnes, Eirik Kliewer, Jörg Machine Learning Cryptography and Security Information Theory We consider the straggler problem in decentralized learning over a logical ring while preserving user data privacy. Especially, we extend the recently proposed framework of differential privacy (DP) amplification by decentralization by Cyffers and Bellet to include overall training latency--comprising both computation and communication latency. Analytical results on both the convergence speed and the DP level are derived for both a skipping scheme (which ignores the stragglers after a timeout) and a baseline scheme that waits for each node to finish before the training continues. A trade-off between overall training latency, accuracy, and privacy, parameterized by the timeout of the skipping scheme, is identified and empirically validated for logistic regression on a real-world dataset and for image classification using the MNIST and CIFAR-10 datasets. |
| title | Straggler-Resilient Differentially-Private Decentralized Learning |
| topic | Machine Learning Cryptography and Security Information Theory |
| url | https://arxiv.org/abs/2212.03080 |