Successive Wyner-Ziv Coding for the Binary CEO Problem Under Logarithmic Loss

Abstract

The L-link binary Chief Executive Officer (CEO) problem under logarithmic loss is investigated in this paper. A quantization splitting technique is applied to convert the problem under consideration to a (2L − 1)-step successive Wyner-Ziv (WZ) problem, for which a practical coding scheme is proposed. In the proposed scheme, low-density generator-matrix (LDGM) codes are used for binary quantization while low-density parity-check (LDPC) codes are used for syndrome generation; the decoder performs successive decoding based on the received syndromes and produces M. Nangir was with the He is now with the DRAFT many seemingly simple sources and distortion measures, the understanding of the relevant information-theoretic limits is rather limited. A remarkable exception is a somewhat underappreciated distortion measure called logarithmic loss (log-loss). As shown by Courtade and Weissman [17], the rate-distortion region of the CEO problem under log-loss admits a single-letter characterization for arbitrary finite-alphabet sources and noisy observations. Different from the conventional distortion measures which are typically imposed on "hard" reconstructions defined over the given source alphabet, the reconstructions associated with DRAFT

Asymptotic Rate-Distortion Analysis of Symmetric Remote Gaussian Source Coding: Centralized Encoding vs. Distributed Encoding

“…Two scenarios are considered: centralized encoding (see Figure 1 ) and distributed encoding (see Figure 2 ). It is worth noting that the distributed encoding scenario is closely related to the CEO problem, which has been studied extensively [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 ].…”