Hi Folks
Recently I came across a published paper "Understanding Database Reconstruction Attacks on Public Data," by Dr. Abowd (Senior Scientist at Census) and Dr. Garfinkel (Senior Computer Scientist at Census).
The paper attacks complementary cell suppression method used by the federal statistical agencies to protect published tabular format data. This attack is an attempt to justify and promote authors' preferred method of differential privacy. The published paper has major flaws and demonstrates lack of understanding on how a complementary cell suppression is performed.
The attached pdf file containing the critique of the above mentioned paper by Dr. Krish Muralidhar from University of Oklahoma provides a clear insight into deficiencies in Census's published paper. I sincerely hope that the statistical community, with vested interest in getting good quality Census2020 data, will take notice of this analysis.
Ramesh A Dandekar
[M.S. Chem Eng., B.S. Chem Eng.]
Retired Mathematical Statistician
------------------------------
Ramesh Dandekar
------------------------------