Title: Optimization of cell suppression process on tabular data and algorithmic matching attack
Speaker: Kazuhiro Minami (The Institute of Statistical Mathematics)
Place: Live-streaming by Zoom

Attendance: 30 (Students: 17; Staffs: 11; Others: 2)

In the talk, the speaker has given a lucid introduction of cell suppression on tabular data and a concrete algorithmic matching attack. Cell suppression on tabular data is a very important process to protect confidential information on survey participants by hiding sensitive cell values. Especially when publishing statistical tables aggregated from various public surveys, the process is needed. However, there is an issue: decreasing the number of cells induces loss of information. Therefore it is desired to minimize loss of information while protected from matching attacks in practice. The speaker has introduced a recent research study to address such a problem, by formulating the cell suppression problem as an optimization problem and then applying the technique of the Benders decomposition. Moreover, its algorithm has been implemented by the speaker in the well-known statistical software R for cell suppression.
As a consequence of this talk, the speaker’s study is very important to address this practical issue. We could also learn how useful mathematical science study is in practice.