Nov. 5 2007 2:00PM

Nov. 5 2007 3:30PM

An information inequality expresses nonnegativity of a linear combination of Shannon entropies of subvectors of a random vector. Recent results on such (linear, unconstraint) information inequalities will be reviewed. As an example, a new fractional Shearer lemma will be presented and used to prove an upper bound on size of the family of the independent sets in a graph. Second application features Zhang-Yeung inequality that implies new tighter bounds on rates in information networks. Third application concerns bounds on rates of the ideal secret sharing schemes.

Pravidelný seminář Inteligentní systémy (Contact person: Tomáš Kroupa)