Dyn-D2P dynamically adjusts DP noise and gradient clipping in decentralized learning, with a 1/sqrt(n) utility rate on top of an unquantified clipping bias.
Our data, ourselves: Privacy via distributed noise generation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Dyn-D$^2$P: Dynamic Differentially Private Decentralized Learning with Provable Utility Guarantee
Dyn-D2P dynamically adjusts DP noise and gradient clipping in decentralized learning, with a 1/sqrt(n) utility rate on top of an unquantified clipping bias.