Meta-learning for the proposed method
Acquiring and utilizing accurate channel state information (CSI) is crucial for realizing the benefits of massive multiple-input multiple-output (MIMO) technology. Current prevailing CSI feedback approaches improve precision by employing advanced deep-learning methods to learn representative CSI features for a subsequent compression process. Unlike previous works, we treat the CSI compression problem in the context of implicit neural representations. Specifically, each CSI matrix is viewed as a neural function that maps the spatial coordinates (antenna and subchannel) to the corresponding channel gains with physical significance. Rather than transmitting the parameters of the specific neural functions, we send low-cost modulations of the CSI matrix, derived through a meta-learning algorithm. These modulations are then applied to a shared base network at the receiver to reconstruct the CSI. Numerical results show that our proposed approach achieves state-of-the-art performance and showcases flexibility in feedback strategies.
(a): The NMSE performance over varying feedback lengths. (b) and (c): Loss values and NMSE during meta-learning process.
(a): Effect of quantization. (b): Coding gain from entropy coding.
@article{wu2024mimo,
title={MIMO Channel as a Neural Function: Implicit Neural Representations for Extreme CSI Compression in Massive MIMO Systems},
author={Wu, Haotian and Zhang, Maojun and Shao, Yulin and Mikolajczyk, Krystian and G{\"u}nd{\"u}z, Deniz},
journal={arXiv preprint arXiv:2403.13615},
year={2024}
}