Friday, July 17, 2009

Demo of CAT (the Cornell Anonymization Toolkit) at SIGMOD 2009

At the recent SIGMOD conference, we presented our new tool for limiting disclosure in data publishing: CAT, the Cornell Anonymization Toolkit. While there has been a lot of interesting work on different algorithms for privacy-preserving data publishing, there are few tools that combine them with a usable GUI that visualize what these algorithms achieve. CAT has an easy-to-use interface; it guides the user through the process of preparing a dataset forpublication while limiting disclosure through the identification of records that have high risk under various attacker models.

CAT currently implements l-diversity and t-closeness.

Below are a few screenshots of CAT; the code is available here. We welcome feedback from users about features that you would like to have and what aspects of the tool you think work well and which need improvement.

No comments:

Post a Comment