Weighted information and weighted entropy

The talk will be about the weighted information (WI) and weighted
entropy (WE) where outcomes of the same probability contribute
according to their weight/utility. Think of two events: a tsunami
somewhere far away and a cracking tree in a garden next door. Both are rare and hence quite informative (say they convey an equal amount of info). But they clearly are of different importance
to us. The emerging concepts of WI/WE have some peculiar properties;
this is what we try to explore. Math results include generalisations
of Ky Fan and Hadamard inequalities and extensions of the
Shannon-McMillan-Breiman theorem.

No preliminary knowledge of Probability/Info Theory will be assumed.

• Speaker: I Stuhl (Denver/Debrecen) and Y Suhov (Penn State/Cambridge)
• Monday 01 August 2016, 14:00–15:00
• Venue: MR12, CMS, Wilberforce Road, Cambridge, CB3 0WB.
• Series: Probability; organiser: Julia Blackwell.