Ushering in a New Era of Algorithm Design

Abstract

Advances in data acquisition technology, together with the imminent demise of Moore’s Law, are prompting a rethink of basic algorithm design principles. Computing with massive data sets, data streaming, coping with uncertainty, priced computation, property testing, and sublinear algorithms are all parts of the story. So is the growing trend toward using algorithms as modeling tools for natural phenomena. I will discuss some of these developments; in particular, dimension reduction, sublinear algorithms, online reconstruction, and self-improving algorithms.