学术报告

报告题目：Phaseless sampling and reconstruction

报告人：成诚（杜克大学&SAMSI，博士后）  

报告时间：2018年6月14日上午上午10：00—11：00

报告地点：数计院307报告室

数学与统计学院

2018.6.13

摘要：In this talk, we consider whether it is possible to determine thesignal $f$ in a shift-invariant space, up to a sign, from its magnitudemeasurements either on the Euclidean space or on a discrete set. This isclosely related to the phase retrieval problem of real-valued signals in aninfinite-dimensional space. For the signal living in a shift-invariant space,we introduce an undirected graph associated with it and use connectivity of thegraph to characterize its phase retrievability.We also propose a stable algorithm to reconstruct the signal from itsnoisy phaseless samples taken on a discrete set. The proposed algorithmprovides a suboptimal approximation to the original signal in magnitudemeasurements.