Technical Program

Paper Detail

Paper Title Towards an Algebraic Network Information Theory: Distributed Lossy Computation of Linear Functions
Paper IdentifierTH1.R4.3
Authors Sung Hoon Lim, Korea Institute of Ocean Science and Technology (KIOST), Korea (South); Chen Feng, University of British Columbia, Canada; Adriano Pastore, CTTC, Canada; Bobak Nazer, Boston University, United States; Michael Gastpar, École polytechnique fédérale de Lausanne (EPFL), Switzerland
Session Multiterminal Source Coding
Location Odéon, Level 3
Session Time Thursday, 11 July, 09:50 - 11:10
Presentation Time Thursday, 11 July, 10:30 - 10:50
Manuscript  Click here to download the manuscript
Abstract Consider the important special case of the K-user distributed source coding problem where the decoder only wishes to recover one or more linear combinations of the sources. The work of Korner and Marton demonstrated that, in some cases, the optimal rate region is attained by random linear codes, and strictly improves upon the best-known achievable rate region established via random i.i.d. codes. Recent efforts have sought to develop a framework for characterizing the achievable rate region for nested linear codes via joint typicality encoding and decoding. Here, we make further progress along this direction by proposing an achievable rate region for simultaneous joint typicality decoding of nested linear codes. Our approach generalizes the results of Korner and Marton to computing an arbitrary number of linear combinations and to the lossy computation setting.