Exact Inference with Approximate Computation for Differentially Private Data via Perturbations

This paper discusses how two classes of approximate computation algorithms can be adapted, in a modular fashion, to achieve exact statistical inference from differentially private data products. Considered are Bayesian for inference, and Monte Carlo Expectation-Maximization likelihood inference. Up error, these is with respect the joint specification both analyst's original model, curator's differential privacy mechanism. Highlighted duality between on data, which leveraged by well-designed computational procedure