Privacy-preserving Learning
HGI » UbiCrypt » Projects

Pri­va­cy-pre­ser­ving Le­arning

1st ad­vi­sor: Simon, 2nd ad­vi­sor: May

In this pro­ject, we are con­cer­ned with tech­ni­ques that allow to per­form sta­tis­ti­cal ana­ly­sis of data in a pri­va­cy-pre­ser­ving fa­shion.

In par­ti­cu­lar, we in­ves­ti­ga­te how well the ob­jec­tives of sta­tis­ti­cal le­arning can be pur­su­ed sub­ject to the cons­traint of achie­ving ep­si­lon-dif­fe­ren­ti­al pri­va­cy.

We fur­ther­mo­re study va­ria­ti­ons in the no­ti­on of pri­va­cy-pre­ser­va­ti­on and ana­ly­ze how they re­la­te to each other.


Prof. Hans Simon

Prof. Hans Simon