Society can gain much value from Big Data. We can study census data to learn where to allocate public resources, medical records from hospitals to fight diseases, or data about students and teachers to evaluate the effectiveness of various approaches to learning and teaching. In all of these scenarios, we need to limit statistical disclosure: we want to release accurate statistics about the data while preserving the privacy of the individuals who contributed it. This paper gives an overview of recent advances and open challenges in the field, focusing on methods that probably limit how much an adversary can learn from a data release.
© 2012 New York Academy of Sciences.