摘要
人工智能正在驱动通信网络架构和技术演进升级,并呈现双向赋能的特征。在NetworkforAI方向,需构建支撑人工智能全场景的新型网络架构,通过增加训练网络带宽与降低传输成本提升AI算力效能,同时强化推理网络对海量智能体的泛在接入与端边云协同能力;在AIforNetwork方向,需推进算力与运力的硬件协同,实现AI能力在网络功能架构中的内生融合。通过算网融合关键技术突破、智能资源编排与分布式协同机制,促进网络系统向高阶自智演进,最终形成AI与通信网络双向驱动的良性发展范式。
AI accelerates the evolution of communication architectures and technologies,demonstrating two-way empowerment.To enable Network for AI,developing AI-native network architectures supporting full-scenario applications is critical.Optimizing training network bandwidth and transmission costs enhances AI computational efficiency,while ubiquitous access for intelligent agents and strengthened end-edge-cloud collaboration in inference networks ensure scalable intelligence deployment.To enable AI for Network,Advancing AI-Network integration requires hardware co-optimization of computing and transport capacities,embedding AI natively within network architectures.Breakthroughs in computing-networking convergence,intelligent resource orchestration,and distributed coordination mechanisms will enable self-evolving network intelligence,establishing a sustainable co-evolution cycle between AI and communication systems.
作者
张健
张嗣宏
Zhang Jian;Zhang Sihong(ZTE Communications Co.,Ltd.,Nanjing 210012,China)
出处
《邮电设计技术》
2025年第8期19-23,共5页
Designing Techniques of Posts and Telecommunications
关键词
人工智能
通信网络
AI智能体
Artificial intelligent
Communication network
AI agent
