Research interest within KHeaT
Our research within "health technology" targets at privacy of medical data (patient data in particular) and the well-known privacy-utility tradeoff. For instance, we have studied
•reidentification, i.e., how well can respective individuals be identified from perturbed versions of data describing them,
•differentially private data-analysis tasks (e.g., clustering, explanations),
•meaningful quantification of the utility of data lost with better privacy,
•query processing and other data-analysis tasks over encrypted data,
•generation of synthetic data (for settings where patient data is needed but is not available).
The last thread, albeit being very important, is currently unfinished and is the one we currently seek to continue.