Breadcrumb
Information inequalities
Supervisor: Oliver Johnson
Theme: Applied Probability
This is a project in Information Theory concerning the behaviour of discrete entropy on summation. It is a surprising fact that a sharp bound (the Entropy Power Inequality) is known for continuous variables, but little progress has been made in the apparently simpler discrete case. This project would investigate the conjectures that can be made in the discrete case, attempting to understand the special role of the Poisson distribution, both by using computers to test some of the conjectures numerically and by rigorously proving theorems.