📄 中文摘要
多智能体深度学习（MADL），包括多智能体深度强化学习（MADRL）、分布式/联邦训练和图结构神经网络，正成为无线系统中决策和推理的统一框架。在5G-Advanced和6G愿景的推动下，感知、通信和计算之间的紧密耦合得到了进一步加强，这通过集成感知与通信、边缘智能、开放可编程无线接入网（RAN）和非地面/无人机网络等技术，形成了去中心化、部分可观测、时变和资源受限的控制问题。本研究综述了2021-2025年间MADL在分布式感知和无线通信领域的最新进展，强调了相关技术的发展和应用前景。
📄 English Summary
Federated Multi Agent Deep Learning and Neural Networks for Advanced Distributed Sensing in Wireless Networks
Multi-agent deep learning (MADL), which encompasses multi-agent deep reinforcement learning (MADRL), distributed/federated training, and graph-structured neural networks, is emerging as a unifying framework for decision-making and inference in wireless systems where sensing, communication, and computing are intricately linked. The recent advancements in 5G-Advanced and 6G visions further enhance this integration through technologies such as integrated sensing and communication, edge intelligence, open programmable RAN, and non-terrestrial/UAV networking. These developments create decentralized, partially observable, time-varying, and resource-constrained control challenges. This survey synthesizes the state of the art, focusing on research from 2021 to 2025, on MADL for distributed sensing and wireless communications, highlighting the evolution and potential applications of these technologies.