Local Decoding in Distributed Compression

TL;DR

This paper investigates the limits of local decoding in distributed compression, showing that for certain sources, strong locality cannot be achieved if one source is compressed below its entropy, unlike the single-source case.

Contribution

It demonstrates the fundamental difference in local decoding capabilities between single-source and multi-source lossless compression, especially for confusable sources.

Findings

Strong locality is achievable for single sources with vanishing error.

For confusable sources, strong locality fails if a source is compressed below entropy.

Non-confusable sources can achieve strong locality even below entropy.

Abstract

It was recently shown that the lossless compression of a single source $X^n$ is achievable with a notion of strong locality; any $X_i$ can be decoded from a constant number of compressed bits, with a vanishing in $n$ probability of error. By contrast, we show that for two separately encoded sources $(X^n,Y^n)$, lossless compression and strong locality is generally not possible. Specifically, we show that for the class of ``confusable'' sources, strong locality cannot be achieved whenever one of the sources is compressed below its entropy. Irrespective of $n$, for some index $i$ the probability of error of decoding $(X_i,Y_i)$ is lower bounded by $2^{-O(d)}$, where $d$ denotes the number of compressed bits accessed by the local decoder. Conversely, if the source is not confusable, strong locality is possible even if one of the sources is compressed below its entropy. Results extend to an arbitrary number of sources.