Quantization Design for Deep Learning-Based CSI Feedback

2025/03/11 by Yin, Manru, Han, Shengqian, Yang, Chenyang · 2 citations
#FOS: Electrical engineering #Signal Processing (eess.SP) #electronic engineering #information engineering

paper · doi:10.48550/arxiv.2503.08125

Abstract

Deep learning-based autoencoders have been employed to compress and reconstruct channel state information (CSI) in frequency-division duplex systems. Practical implementations require judicious quantization of encoder outputs for digital transmission. In this paper, we propose a novel quantization module with bit allocation among encoder outputs and develop a method for joint training the module and the autoencoder. To enhance learning performance, we design a loss function that adaptively weights the quantization loss and the logarithm of reconstruction loss. Simulation results show the performance gain of the proposed method over existing baselines.

Cited by

• NEFT: A Unified Transformer Framework for Efficient Near-Field CSI Feedback in XL-MIMO Systems
• CSI Compression Beyond Latents: End-to-End Hybrid Attention-CNN Networks with Entropy Regularization

Related

• Learning Representations for CSI Adaptive Quantization and Feedback
• Quantization Adaptor for Bit-Level Deep Learning-Based Massive MIMO CSI Feedback
• Information-Preserving CSI Feedback: Invertible Networks with Endogenous Quantization and Channel Error Mitigation
• Bit-level Optimized Neural Network for Multi-antenna Channel Quantization
• Changeable Rate and Novel Quantization for CSI Feedback Based on Deep Learning