Information distance and applications

Abstract

We know how to measure distance from Turku to Toronto. However, do you know how to measure the distance between two information carrying entities? For example: two genomes, two music scores, two programs, two articles, two emails, two concepts, or from a question to an answer? Furthermore, such a distance measure must be application-independent, must be universal in the sense it is provably better than all other distances, and must be applicable. From a simple and accepted assumption in thermodynamics, we have developed such a theory. I will present this theory and will talk about some new applications of this theory, including a question answering system. © Springer-Verlag Berlin Heidelberg 2007.