The topics of private data analysis and streaming data management have both been separately the focus of much study within the data management community for many years. However, more recently there have been studies which bring these two previously isolated topics together. Although data streams and privacy might not seem to have much in common, it turns out that they share a symbiotic relationship. Within data streams, a common pattern is to design a compact data structure that summarizes the input that has been seen so far, and which can be updated quickly in order to reflect a small change. Within privacy, it is often helpful to build an intermediate representation of a data set, so that a small change to the input does not change the representation much. Moreover, both areas make use of the tools of randomized algorithms and probability to prove that their mechanisms give strong guarantees. Thus, data stream summaries can sometimes inspire private algorithms, and vice-versa.
This file was generated by bibtex2html 1.92.