PrivDPR generates synthetic graphs under node-level DP by adding Gaussian noise only to the gradient of a node embedding in a deep PageRank model, with depth used to control sensitivity.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
PrivDPR: Synthetic Graph Publishing with Deep PageRank under Differential Privacy
PrivDPR generates synthetic graphs under node-level DP by adding Gaussian noise only to the gradient of a node embedding in a deep PageRank model, with depth used to control sensitivity.