ResidualPlanner+: a scalable matrix mechanism for marginals and beyond

ArXi:2305.08175v4 Announce Type: replace-cross Noisy marginals are a common form of confidentiality protecting data release and are useful for many downstream tasks such as contingency table analysis, construction of Bayesian networks, and even synthetic data generation. Privacy mechanisms that provide unbiased noisy answers to linear queries (such as marginals) are known as matrix mechanisms. We propose ResidualPlanner and ResidualPlanner+, two highly scalable matrix mechanisms.