Speaker: Dr. Debolina Ghatak
Affiliation: Post-Doctoral Fellow at IAI, TCG CREST
Date: 09/03/2022
Title: Inference from Private Microdata
Abstract: Microdata can be found in several application areas like economic data, medical data, industrial data etc. Microdata, if released in its raw form are prone to several kinds of attacks. Several privacy guarantees have been discovered over time including k-anonymity, l-diversity, t-closeness, differential privacy and its variants. However, privacy protection is not enough for microdata. It is also very important to utilize the protected data for inferencing. The recent focus in this area of work is to develop trade-off models that find parameter values that balance between disclosure risk and utility loss parameters. In this talk, we will discuss a few models and challenges in this field of work.