We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.SP
< prev | next >
new | recent | 2204

Change to browse by:

cs
cs.IT
cs.LG
eess
math

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Electrical Engineering and Systems Science > Signal Processing

Title: Model-Based Deep Learning of Joint Probabilistic and Geometric Shaping for Optical Communication

Authors: Vladislav Neskorniuk, Andrea Carnio, Domenico Marsella, Sergei K. Turitsyn, Jaroslaw E. Prilepsky, Vahid Aref
(Submitted on 5 Apr 2022)
Abstract: Autoencoder-based deep learning is applied to jointly optimize geometric and probabilistic constellation shaping for optical coherent communication. The optimized constellation shaping outperforms the 256 QAM Maxwell-Boltzmann probabilistic distribution with extra 0.05 bits/4D-symbol mutual information for 64 GBd transmission over 170 km SMF link.
Comments: 2 pages; accepted for oral presentation at CLEO 2022 in May 2022
Subjects: Signal Processing (eess.SP); Information Theory (cs.IT); Machine Learning (cs.LG)
Cite as: arXiv:2204.07457 [eess.SP]
  (or arXiv:2204.07457v1 [eess.SP] for this version)

Submission history

From: Vladislav Neskorniuk [view email]
[v1] Tue, 5 Apr 2022 19:35:40 GMT (1659kb,D)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.