|
Probability Theory assisted by
Information Theory |
|
The purpose of the course is to prove
the most important theorems in probability theory using methods from
information theory. The chosen theorems will be formulated in stronger
versions than normally found in the literature and the proofs will be more
direct. Finally, this approach will be closer to statistical practice than
one normally finds in probability theory. Subjects
A detailed list of topics which have been
discussed in the lectures can be found here. Prerequests
Knowledge of elementary
probability theory and measure theory. Time
and place
Wednesday 8-10 in Aud. 7 at H. C.
Ørstedsinstituttet. back to top Notes
An updated version of the notes
is available here. Any feed-back on the notes is deeply appreciated.
Syllabus
3/9: Coding and Kraft’s
inequality. Entropy and divergence. 10/9: Inequalities, convexity and
continuity. 17/9: Pinsker’s inequality and
Poisson's law. 24/9: No lecture. 1/10: Universal source coding and
information projection. 8/10: Information
projection. 22/10: Strong Law of Large
Numbers. 29/10: Increasing and
decreasing information. 5/11: Markov chains 12/11: Reversible Markov
chains and Ergodicity. 19/11: The Conditional Limit
Theorem and Sanov property. Stein's Lemma and Chernoff information. 26/11: Convergence of direct and
inverse martingales. 3/12: Hewit-Savage 0-1 Law.
Central Limit Theorem. 10/12: The Central Limit Theorem. 17/12: Edgeworth expansion. A Mathematical
Theory of Communication by Claude Shannon 1948. Entropy on the world wide
web. Information
Theory Society contains a lot of links to relevant material. Lecture
notes on information theory made by Flemming Topsøe. Lecture
notes on inequalities of information theory made by Flemming Topsøe. |
|
|
Last
modified: 12/11 2003 by Peter Harremoës - moes@math.ku.dk |