Low-Rank Decomposition for Rate-Adaptive Deep Joint Source-Channel Coding

Man Xu, Chan Tong Lam, Yuanhui Liang, Benjamin Ng, Sio Kei Im

研究成果: Conference contribution同行評審

1 引文 斯高帕斯(Scopus)

摘要

Deep joint source-channel coding (DJSCC) has received extensive attention in the communications community. However, the high computational costs and storage requirements prevent the DJSCC model from being effectively deployed on embedded systems and mobile devices. Recently, convolutional neural network (CNN) compression via low-rank decomposition has achieved remarkable performance. In this paper, we conduct a comparative study of low-rank decomposition for lowering the computational complexity and storage requirement for Rate-Adaptive DJSCC, including CANDECOMP/PARAFAC (CP) de-composition, Tucker (TK) decomposition, and Tensor-train (TT) decomposition. We evaluate the compression ratio, speedup ratio, and Peak Signal-to-Noise Ratio (PSNR) performance loss for the CP, TK, and TT decomposition with fine-tuning and pruning. From the experimental results, we found that compared with the TT decomposition, CP decomposition with fine-tuning lowers the PSNR performance degradation at the expense of higher compression and speedup ratio.

原文English
主出版物標題2022 IEEE 8th International Conference on Computer and Communications, ICCC 2022
發行者Institute of Electrical and Electronics Engineers Inc.
頁面58-64
頁數7
ISBN(電子)9781665450515
DOIs
出版狀態Published - 2022
事件8th IEEE International Conference on Computer and Communications, ICCC 2022 - Virtual, Online, China
持續時間: 9 12月 202212 12月 2022

出版系列

名字2022 IEEE 8th International Conference on Computer and Communications, ICCC 2022

Conference

Conference8th IEEE International Conference on Computer and Communications, ICCC 2022
國家/地區China
城市Virtual, Online
期間9/12/2212/12/22

指紋

深入研究「Low-Rank Decomposition for Rate-Adaptive Deep Joint Source-Channel Coding」主題。共同形成了獨特的指紋。

引用此