Randomness extractors and data storage

Abstract

Deterministic randomness extractors are functions E : {0, 1}n → {0, 1}m which refine imperfect sources of randomness in the following sense: For every probability distribution X in some “interesting family” of distributions over {0,1}n, applying E on a sample from X yields a distribution that is (close to) the uniform distribution. Randomness extractors have many applications in various areas of computer science. Recently, Shpilka [Shp13] showed how to apply randomness extractors to solve problems in the area of data storage. Following work by Shpilka [Shp14] and Gabizon and Shaltiel [GS12b] build on this connection and extend Shpilka's original paper. In this article, we give some relevant background on randomness extractors and explain how extractors (and closely related dispersers) can be applied to solve problems in data storage.