...n of weak sources. von-Neumann considered sources where each bit from the source is independent, but biased, i.e. the probability of the bit 1 being 1 is p, where p may not be 2 . Santha and Vazirani =-=[SV84]-=- considered sources where each bit is biased and the bias of the bit depends on the bits that have come before it. The most general notion of a weak source to date was first introduced by Nisan and Zu...

...t: def • L = [Fq ] k def • R = [Fq ] def • G = {(x, y)|�i s.t. E(x)i = y}, i.e., for any element x ≤ L, x is connected to the n elements E(x)1, E(x)2, ..., E(x)n in R. q^k x n i Ta-Shma and Zuckerman =-=[TSZ01]-=- observed that this graph is a good extractor exactly when the corresponding code is a good list decodable code! In fact, when the graph is a good extractor, the corresponding code has a property that...

...red sources where each bit is biased and the bias of the bit depends on the bits that have come before it. The most general notion of a weak source to date was first introduced by Nisan and Zuckerman =-=[NZ93]-=-. Here, we bound the min-entropy of the source: Definition 2 The min-entropy of a source D is defined to be: def H�(D) = − log(max(D(x)) x�D We will call a distribution over {0, 1} n that has min-entr...