Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
The paper offers algorithms for oblivious sampling for application in differentially private data analysis. The algorithms are simple but the analyses are non-trivial and the results are new. The proposed techniques provide a useful and practical solution for implementing DP algorithms involving privacy amplification by sampling. On the other hand, as pointed out by the reviewers, the experimental component is rather limited and can be improved.