Organizer 组织者:

• • Fuhui Zhou (Professor), Nanjing University of Aeronautics and Astronautics, China
  周福辉，南京航空航天大学，教授
• • Wei Wu (Professor), Nanjing University of Posts and Telecommunications, China
  吴伟，南京邮电大学，教授
• • Yuhang Wu (Postdoctor), Nanjing University of Aeronautics and Astronautics, China
  吴雨航，南京航空航天大学，博士后

Introduction 论坛简介:

The rapid advancement of Large Language Models (LLMs) has created transformative opportunities in wireless communications, particularly in spectrum management and semantic communications. In spectrum management, LLMs excel at analyzing complex electromagnetic environments, enabling real-time spectrum sharing decisions through advanced pattern recognition and predictive analytics. For semantic communications, LLMs redefine traditional architectures by introducing context-aware compression and generative reconstruction techniques that preserve meaning while optimizing bandwidth, shifting the paradigm from bit-level transmission to knowledge-level communication and enhancing spectral efficiency with reduced latency. However, integrating LLMs presents challenges such as computational complexity optimization for real-time operation, robustness in dynamic wireless channels, and privacy-preserving distributed learning frameworks. Addressing these requires cross-disciplinary approaches combining information theory, machine learning, and wireless networking to develop efficient deployment strategies. The forum aims to facilitate discussions on the latest research achievements, technical bottlenecks, and future trends, fostering deep integration between LLMs and wireless communications to advance next-generation intelligent networks.

大语言模型（LLM）的快速发展为无线通信领域，特别是频谱管理和语义通信，带来了变革性的机遇。在频谱管理方面，LLM擅长分析复杂的电磁环境，通过先进的模式识别和预测分析实现实时频谱共享决策。在语义通信方面，LLM通过引入上下文感知压缩和生成式重建技术来重新定义传统架构，这些技术在优化带宽的同时保留语义，将范式从比特级传输转向知识级通信，并以更低的延迟提高频谱效率。然而，LLM的集成也带来了挑战，例如实时操作的计算复杂度优化、动态无线信道中的鲁棒性以及保护隐私的分布式学习框架。解决这些问题需要结合信息论、机器学习和无线网络的跨学科方法，以制定高效的部署策略。本论坛旨在促进关于最新研究成果、技术瓶颈和未来趋势的讨论，推动LLM与无线通信的深度融合，以推进下一代智能网络的发展。

Topics 主题范围:

• LLM-based dynamic spectrum access and sharing
  基于LLM的动态频谱接入与共享
• Foundation Models for Spectrum Intelligence
  面向频谱智能的基础模型
• Multi-Task Oriented Spectrum Foundation Model
  面向多任务的频谱基础模型
• Foundation Model-based Spectrum Prediction and Decision
  基于基础模型的频谱预测与决策
• LLM-enabled cognitive radio networks
  LLM赋能的认知无线电网络
• Federated learning for privacy-preserving spectrum coordination
  用于保护隐私的频谱协调的联邦学习
• LLM-assisted beamforming and MIMO optimization
  LLM辅助的波束成形与MIMO优化
• Semantic communication architectures with LLMs
  基于LLM的语义通信架构
• Context-aware semantic encoding and decoding
  上下文感知的语义编码与解码
• Knowledge distillation for efficient semantic transmission
  面向高效语义传输的知识蒸馏
• Generative AI for wireless signal reconstruction
  用于无线信号重建的生成式AI
• Energy-efficient LLM deployment in wireless networks
  无线网络中高能效的LLM部署
• LLM-driven resource allocation in wireless networks
  LLM驱动的无线网络资源分配
• Cross-modal semantic communication
  跨模态语义通信
• Edge AI for real-time LLM inference in wireless systems
  无线系统中用于实时LLM推理的边缘AI
• Adversarial robustness in LLM-based communications
  基于LLM的通信中的对抗鲁棒性
• LLM-enabled intent-based networking
  LLM赋能的基于意图的网络
• Standardization and regulatory challenges for AI in spectrum
  AI在频谱中的标准化与监管挑战