Dimension Extractors and Optimal Decompression

A dimension extractor is an algorithm designed to increase the effective dimension—i.e., the amount of computational randomness—of an infinite binary sequence, in order to turn a “partially random” sequence into a “more random” sequence. Extractors are exhibited for various effective dimensions, including constructive, computable, space-bounded, time-bounded, and finite-state dimension. Using similar techniques, the Ku?era-Gács theorem is examined from the perspective of decompression, by showing that every infinite sequence S is Turing reducible to a Martin-Löf random sequence R such that the asymptotic number of bits of R needed to compute n bits of S, divided by n, is precisely the constructive dimension of S, which is shown to be the optimal ratio of query bits to computed bits achievable with Turing reductions. The extractors and decompressors that are developed lead directly to new characterizations of some effective dimensions in terms of optimal decompression by Turing reductions.