English

Source Coding for Synthesizing Correlated Randomness

Information Theory 2021-05-03 v2 math.IT

Abstract

We consider a scenario wherein two parties Alice and Bob are provided X1nX_{1}^{n} and X2nX_{2}^{n} -- samples that are IID from a PMF PX1X2P_{X_1 X_2}. Alice and Bob can communicate to Charles over (noiseless) communication links of rate R1R_1 and R2R_2 respectively. Their goal is to enable Charles generate samples YnY^{n} such that the triple (X1n,X2n,Yn)(X_{1}^{n},X_{2}^{n},Y^{n}) has a PMF that is close, in total variation, to PX1X2Y\prod P_{X_1 X_2 Y}. In addition, the three parties may posses pairwise shared common randomness at rates C1C_1 and C2C_2. We address the problem of characterizing the set of rate quadruples (R1,R2,C1,C2)(R_1,R_2,C_1,C_2) for which the above goal can be accomplished. We provide a set of sufficient conditions, i.e. an inner bound to the achievable rate region, and necessary conditions, i.e. an outer bound to the rate region for this three party setup. We provide a joint-typicality based random coding argument involving encoding and decoding operations to perform soft covering and a pertinent relaxation of the PMF requirement for the encoders.

Cite

@article{arxiv.2004.03651,
  title  = {Source Coding for Synthesizing Correlated Randomness},
  author = {Touheed Anwar Atif and Arun Padakandla and S. Sandeep Pradhan},
  journal= {arXiv preprint arXiv:2004.03651},
  year   = {2021}
}
R2 v1 2026-06-23T14:43:26.698Z