Preserving Individual Privacy in Serial Data Publishing

This is an important paper, talks about updates made to the current dataset over time

But this targets only updates to the current dataset, not additions to the dataset

Also only the sensitive attribute can change here, the other qids are assumed to be the same, again this is different from our dynamic qids that have been used

Talk about a medical dataset

we find that no previous work has sufficient protection provided for sensitive values that can change over time, which should be the more common case. In this work, we propose to study the privacy guarantee for such transient sensitive values, which we call the global guarantee.

Local guarantee means in the table the l diversity is maintained

But global means across releases → this is needed for temporal attacks

But problem is how do you anticipate the nature of incoming data

Main idea used in this paper is to create groups and see what is the probability that a particular tuple can be associated to a particular group, that probability should be less than k, that is threshold that has been used to perform the task