Hybrid neural coded modulation: Design and training methods
We propose a hybrid coded modulation scheme which composes of inner and outer codes. The outer-code can be any standard binary linear code with efficient soft decoding capability (e.g. low-density parity-check (LDPC) codes). The inner code is designed using a deep neural network (DNN) which takes th...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2022-03-01
|
| Series: | ICT Express |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405959522000182 |
| _version_ | 1818288601708888064 |
|---|---|
| author | Sung Hoon Lim Jiyong Han Wonjong Noh Yujae Song Sang-Woon Jeon |
| author_facet | Sung Hoon Lim Jiyong Han Wonjong Noh Yujae Song Sang-Woon Jeon |
| author_sort | Sung Hoon Lim |
| collection | DOAJ |
| description | We propose a hybrid coded modulation scheme which composes of inner and outer codes. The outer-code can be any standard binary linear code with efficient soft decoding capability (e.g. low-density parity-check (LDPC) codes). The inner code is designed using a deep neural network (DNN) which takes the channel coded bits and outputs modulated symbols. For training the DNN, we propose to use a loss function that is inspired by the generalized mutual information. The resulting constellations are shown to outperform the conventional quadrature amplitude modulation (QAM) based coding scheme for modulation order 16 and 64 with 5G standard LDPC codes. |
| first_indexed | 2024-12-13T01:58:59Z |
| format | Article |
| id | doaj.art-3bfac48e4873453188d781e33c04bf6c |
| institution | Directory Open Access Journal |
| issn | 2405-9595 |
| language | English |
| last_indexed | 2024-12-13T01:58:59Z |
| publishDate | 2022-03-01 |
| publisher | Elsevier |
| record_format | Article |
| series | ICT Express |
| spelling | doaj.art-3bfac48e4873453188d781e33c04bf6c2022-12-22T00:03:19ZengElsevierICT Express2405-95952022-03-01812530Hybrid neural coded modulation: Design and training methodsSung Hoon Lim0Jiyong Han1Wonjong Noh2Yujae Song3Sang-Woon Jeon4School of Software, Hallym University, Chuncheon, Republic of KoreaSchool of Software, Hallym University, Chuncheon, Republic of KoreaSchool of Software, Hallym University, Chuncheon, Republic of KoreaDepartment of Computer Software Engineering, Kumoh National Institute of Technology, Gumi 39177, Republic of KoreaDepartment of Electrical and Electronic Engineering, Hanyang University, Ansan, Republic of Korea; Corresponding author.We propose a hybrid coded modulation scheme which composes of inner and outer codes. The outer-code can be any standard binary linear code with efficient soft decoding capability (e.g. low-density parity-check (LDPC) codes). The inner code is designed using a deep neural network (DNN) which takes the channel coded bits and outputs modulated symbols. For training the DNN, we propose to use a loss function that is inspired by the generalized mutual information. The resulting constellations are shown to outperform the conventional quadrature amplitude modulation (QAM) based coding scheme for modulation order 16 and 64 with 5G standard LDPC codes.http://www.sciencedirect.com/science/article/pii/S2405959522000182Machine learningNeural networksModulationChannel codingGeneralized mutual information |
| spellingShingle | Sung Hoon Lim Jiyong Han Wonjong Noh Yujae Song Sang-Woon Jeon Hybrid neural coded modulation: Design and training methods ICT Express Machine learning Neural networks Modulation Channel coding Generalized mutual information |
| title | Hybrid neural coded modulation: Design and training methods |
| title_full | Hybrid neural coded modulation: Design and training methods |
| title_fullStr | Hybrid neural coded modulation: Design and training methods |
| title_full_unstemmed | Hybrid neural coded modulation: Design and training methods |
| title_short | Hybrid neural coded modulation: Design and training methods |
| title_sort | hybrid neural coded modulation design and training methods |
| topic | Machine learning Neural networks Modulation Channel coding Generalized mutual information |
| url | http://www.sciencedirect.com/science/article/pii/S2405959522000182 |
| work_keys_str_mv | AT sunghoonlim hybridneuralcodedmodulationdesignandtrainingmethods AT jiyonghan hybridneuralcodedmodulationdesignandtrainingmethods AT wonjongnoh hybridneuralcodedmodulationdesignandtrainingmethods AT yujaesong hybridneuralcodedmodulationdesignandtrainingmethods AT sangwoonjeon hybridneuralcodedmodulationdesignandtrainingmethods |