Normalized Information Distance is Not Semicomputable

TL;DR

This paper proves that the theoretical normalized information distance (NID), based on Kolmogorov complexity, cannot be approximated from above or below by any computable process, highlighting fundamental limits of its computability.

Contribution

It establishes that the normalized information distance (NID) is neither upper nor lower semicomputable, clarifying its theoretical limitations.

Findings

NID is not upper semicomputable.

NID is not lower semicomputable.

This limits practical approximation of NID.

Abstract

Normalized information distance (NID) uses the theoretical notion of Kolmogorov complexity, which for practical purposes is approximated by the length of the compressed version of the file involved, using a real-world compression program. This practical application is called 'normalized compression distance' and it is trivially computable. It is a parameter-free similarity measure based on compression, and is used in pattern recognition, data mining, phylogeny, clustering, and classification. The complexity properties of its theoretical precursor, the NID, have been open. We show that the NID is neither upper semicomputable nor lower semicomputable.