Large Language Model Empowered CSI Feedback in Massive MIMO Systems

2026-03-03

AI Summary

This work addresses the challenge of efficient channel state information (CSI) compression and feedback in frequency-division duplexing massive MIMO systems by introducing large language models (LLMs) to the CSI feedback task for the first time. The problem is reformulated as a masked token prediction task, aligning closely with the pretraining objective of LLMs. Guided by an information-theoretic self-information metric, the method dynamically selects high-information CSI elements for masking and feedback. This approach not only significantly reduces feedback overhead but also enhances CSI reconstruction accuracy, thereby achieving more efficient channel state feedback.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Despite the success of large language models (LLMs) across domains, their potential for efficient channel state information (CSI) compression and feedback in frequency division duplex (FDD) massive multiple-input multiple-output (mMIMO) systems remains largely unexplored yet increasingly important. In this paper, we propose a novel LLM-based framework for CSI feedback to exploit the potential of LLMs. We first reformulate the CSI compression feedback task as a masked token prediction task that aligns more closely with the functionality of LLMs. Subsequently, we design an information-theoretic mask selection strategy based on self-information, identifying and selecting CSI elements with the highest self-information at the user equipment (UE) for feedback. This ensures that masked tokens correspond to elements with lower self-information, while visible tokens correspond to elements with higher self-information, thus maximizing the accuracy of LLM predictions.

Problem

• CSI feedback
• massive MIMO
• FDD
• channel state information
• compression

Innovation

• Large Language Model
• CSI Feedback
• Masked Token Prediction
• Self-Information
• Massive MIMO