Speaker
Mario Ullrich
(Johannes Kepler University, Linz, Austria)
Description
We show that the maximal gain of adaption and randomization is limited when considering approximation of functions from convex sets based on arbitrary linear measurements in a worst-case setting.
We also discuss the situation when arbitrary non-linear, Lipschitz-continuous measurements are allowed, where some (surprising) improvements hold.
Primary author
Mario Ullrich
(Johannes Kepler University, Linz, Austria)
