Deep Unfolding for Accelerating Iterative Algorithms in Wireless Communications

  • Nguyen, The Vi; 
  • Nguyen, Thi My Tuyen; 
  • Tran, Kiet Nguyen Tuan; 
  • Kim, Gahyun; 
  • Cho, Sungrae
Citations

WEB OF SCIENCE

0
Citations

SCOPUS

0

초록

Iterative algorithms are widely used to solve optimization problems in communications and signal processing tasks. However, high iteration counts and computational complexity per iteration lead to excessive computational latency, limiting the use of these algorithms in time-sensitive applications. To overcome these challenges, a promising model-based deep learning technique, called deep unfolding, has recently been employed to accelerate iterative algorithms. In particular, deep unfolding transforms iterative algorithms into trainable neural network architectures, where each iteration corresponds to a network layer and algorithmic parameters become learnable. This approach preserves interpretability while enabling fast inference. Moreover, recent advances further reduce computational complexity per iteration with surrogate functions. These developments provide efficient and interpretable learning-based optimization frameworks suitable for real-time communication systems.

키워드

6G networks; Deep unfolding; Machine learning
제목
Deep Unfolding for Accelerating Iterative Algorithms in Wireless Communications
저자
Nguyen, The Vi; Nguyen, Thi My Tuyen; Tran, Kiet Nguyen Tuan; Kim, Gahyun; Cho, Sungrae
DOI
10.1109/ICOIN68469.2026.11480627
발행일
2026
유형
Proceedings Paper
저널명
International Conference on Information Networking
페이지
634 ~ 636