Imperial College London
Alternate

DrCongLing

Faculty of Engineering, Department of Electrical and Electronic Engineering

Reader in Coding and Information Theory
 
 
 
//

Contact

 

+44 (0)20 7594 6214c.ling

 
 
//

Location

 

815Electrical EngineeringSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@inproceedings{Wang:2016:10.1109/ITW.2016.7606863,
author = {Wang, Z and Ling, C},
doi = {10.1109/ITW.2016.7606863},
publisher = {IEEE},
title = {Symmetric mettropolis-within-Gibbs algorithm for lattice Gaussian sampling},
url = {http://dx.doi.org/10.1109/ITW.2016.7606863},
year = {2016}
}
Download

RIS format (EndNote, RefMan)

TY  - CPAPER
AB  - As a key sampling scheme in Markov chain MonteCarlo   (MCMC)   methods,   Gibbs   sampling   is   widely   used   invarious  research  fields  due  to  its  elegant  univariate  conditionalsampling,  especially  in  tacking  with  multidimensional  samplingsystems. In this paper, a Gibbs-based sampler named as symmet-ric Metropolis-within-Gibbs (SMWG) algorithm is proposed forlattice Gaussian sampling. By adopting a symmetric Metropolis-Hastings (MH) step into the Gibbs update, we show the Markovchain  arising  from  it  is  geometrically  ergodic,  which  convergesexponentially  fast  to  the  stationary  distribution.  Moreover,  byoptimizing  its  symmetric  proposal  distribution,  the  convergenceefficiency  can  be  further  enhanced.
AU  - Wang,Z
AU  - Ling,C
DO  - 10.1109/ITW.2016.7606863
PB  - IEEE
PY  - 2016///
TI  - Symmetric mettropolis-within-Gibbs algorithm for lattice Gaussian sampling
UR  - http://dx.doi.org/10.1109/ITW.2016.7606863
UR  - http://hdl.handle.net/10044/1/40358
ER  -
Download