Engineering
In this chapter, we briefly discussed 3 methods (other than hiding the name, etc.) of inference control. . The query set size control: don't return an answer if the set size is too small; N-respondent, k% dominance rule: don't return an answer if = k% of the population/data; - Randomization: add a small amount of random noise to the data. I. (1 pt) What is the advantage of using "randomization" compared to using "query set size control" or "N-respondent, k% dominance rule"? II. (1 pt) What is the disadvantage of using "randomization" compared to using "query set size control" or "N-respondent, k% dominance rule"? III. (1 pt) Outline an attack on a system using "query set size control". That is, outline a way of getting some information (such as average) of a small set A. IV. (1 pt) Outline an attack on a system using "N-respondent, k% dominance rule". That is, outline a way of getting some information (such as average) of a set A(with size < N) that contributes the majority (>k%) of data (let's say, set B). V. (1 pt) Outline an attack on a system using "randomization". That is, outline a way of "recovering" the original data that does not include the random noise.