A theoretical and empirical study finds that unbiased second-moment estimates in private Adam can be harmful in high dimensions, and that scale-then-privatize outperforms the alternatives on a small transformer task.
Improved dif- ferential privacy for sgd via optimal private linear operators on adaptive streams.Advances in Neural Information Processing Systems, 35:5910–5924, 2022
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
On Design Principles for Private Adaptive Optimizers
A theoretical and empirical study finds that unbiased second-moment estimates in private Adam can be harmful in high dimensions, and that scale-then-privatize outperforms the alternatives on a small transformer task.