Student Probability

Suppose we wish to recover a signal x in R^n from m measurements of the form y_i = || , which is know as Phase Retrieval problem. A recent proposed algorithm called PhaseMax solve the above problem via linear program. We show an elementary proof the PhaseMax algorithm relies on standard probabilistic concentration and covering arguments. This talk is based on the paper '' An elementary proof of Convex Phase Retrieval in the Natrural Parameter Space via the Linear Program PhaseMax'' by P. Hand and V. Voroninski. Speaker(s): Yun Wei (UM)