ResidualPlanner provides an optimal scalable matrix mechanism for Gaussian noise on marginal queries that optimizes convex loss functions of variances, with ResidualPlanner+ extending support to combined marginal and range/prefix-sum workloads while outperforming HDMM.
For d beween 8 and 15, HDMM decides to add noise to all 3-way marginals and nothing else (even though the workload is all ≤ 3 marginals)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2023 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ResidualPlanner+: a scalable matrix mechanism for marginals and beyond
ResidualPlanner provides an optimal scalable matrix mechanism for Gaussian noise on marginal queries that optimizes convex loss functions of variances, with ResidualPlanner+ extending support to combined marginal and range/prefix-sum workloads while outperforming HDMM.