Polarized Dropout: A Novel Deep Joint Source Channel Coding Scheme for Erasure Channels

This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted dropout technique in endowing deep neural networks with resilience against erasures. However, directly applying dropout at the channel layer...

Full description

Saved in:
Bibliographic Details
Published in:IEEE communications letters Vol. 28; no. 9; pp. 1986 - 1990
Main Authors: Ren, Zichang, Wang, Yiru, Zhao, Yuping
Format: Journal Article
Language:English
Published: New York IEEE 01.09.2024
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Artificial neural networks
Channels
Coding
Decoding
deep learning
erasure channel
Joint source-channel coding
Number theory
Performance degradation
polar code
Receivers
Symbols
Training
Transfer functions
Transforms
Vectors
ISSN:1089-7798, 1558-2558
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This letter investigates the challenges encountered by deep joint source-channel coding in erasure channels. We explore the effectiveness of the widely adopted dropout technique in endowing deep neural networks with resilience against erasures. However, directly applying dropout at the channel layer introduces uncertainty into the neural network's training process, leading to performance degradation. To address this issue, we introduce the Polarized Dropout scheme and a novel network architecture that encodes analog symbols using the Walsh-Hadamard transform based on real-number field computation. Leveraging the polarization of symbol recovery probabilities, for a given erasure rate, a determined set of neurons will be assigned a dropout rate of 1, while the remainder are assigned a dropout rate of 0. Simulation results indicate a maximum enhancement of nearly 6dB in communication performance.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1089-7798
1558-2558
DOI:10.1109/LCOMM.2024.3422176