RSB decoupling property of MAP estimators
Bereyhi A, Müller R, Schulz-Baldes H (2016)
Publication Language: English
Publication Type: Conference contribution, Original article
Publication year: 2016
Publisher: IEEE
Conference Proceedings Title: 2016 IEEE Information Theory Workshop (ITW)
Event location: Cambridge, London
Open Access Link: https://arxiv.org/pdf/1611.02629
Abstract
The large-system decoupling property of a MAP estimator is studied when it estimates the i.i.d. vector x from the observation y=Ax+z with A being chosen from a wide range of matrix ensembles, and the noise vector z being i.i.d. and Gaussian. Using the replica method, we show that the marginal joint distribution of any two corresponding input and output symbols converges to a deterministic distribution which describes the input-output distribution of a single user system followed by a MAP estimator. Under the bRSB assumption, the single user system is a scalar channel with additive noise where the noise term is given by the sum of an independent Gaussian random variable and b correlated interference terms. As the bRSB assumption reduces to RS, the interference terms vanish which results in the formerly studied RS decoupling principle.
Authors with CRIS profile
Ali Bereyhi
Lehrstuhl für Digitale Übertragung
Ralf Müller
Professur für Informationsübertragung
Hermann Schulz-Baldes
Professur für Mathematik (Mathematische Physik)
How to cite
APA:
Bereyhi, A., Müller, R., & Schulz-Baldes, H. (2016). RSB decoupling property of MAP estimators. In 2016 IEEE Information Theory Workshop (ITW). Cambridge, London, GB: IEEE.
MLA:
Bereyhi, Ali, Ralf Müller, and Hermann Schulz-Baldes. "RSB decoupling property of MAP estimators." Proceedings of the 2016 IEEE Information Theory Workshop (ITW), Cambridge, London IEEE, 2016.
BibTeX: Download