热点聚焦
MORE >- SI CFP: Advanced Technologies for Security, Privacy, Safety, and Regulation of AI-generated Content
- 学术讲座 | 信息与智能大讲堂：Overview of Fluid Antenna System (FAS) and Its Potential for 6G
- 学术讲座 | 信息与智能大讲堂：资源受限下的大容量信息探测与传输方法
- SI CFP: AI-Driven Integration of TN and NTN
- 【征文通知】AI 赋能通信网络专题（2025年08月31日截稿）
- 学术讲座 | 信息与智能大讲堂：多模态场景理解
- SI CFP: AI-RAN: Foundations, Methodologies, and Application
- ICAI2025｜征稿通知正式发布
- 第二十七届中国科协年会 “AI时代网络技术创新”专题征稿通知
快捷入口
投稿中心
审稿中心
编辑办公
主编办公
最新出版
更多 2026年第4卷第1期
-
-
- 摘要：With the rapid advancement of Artificial Intelligence (AI), the Radio Access Network (RAN) is poised to undergo a transformative evolution toward the convergence of AI and RAN functionalities, referred to as the AI-RAN paradigm. AI-RAN integrates high-performance computing resources into RAN infrastructures, thereby enabling the execution of both AI and RAN workloads on the same infrastructure. As a result, it improves resource utilization, reduces energy consumption, and promotes swift AI-related responses. In this paper, we provide a comprehensive overview of AI-RAN, whereby we broadly categorize the discussion into three aspects: AI and RAN, AI for RAN, and AI on RAN. In particular, we begin with AI and RAN, which encompass the hardware architecture, software stack, as well as orchestration of computational and communication resources. We subsequently elaborate on the AI on RAN, examining various approaches to leveraging AI methods to enhance RAN performance. For the topic of AI on RAN, we conduct an in-depth investigation into the schemes that take RAN as a platform to facilitate AI services, where we review distributed learning for multi-cell and multi-vendor RANs, including federated and multi-agent reinforcement learning, highlighting issues of data heterogeneity, control-plane overhead, convergence under mobility, privacy, and adversarial robustness in the RAN ecosystems. We also demonstrate several use cases pertaining to the AI-RAN framework. We conclude by outlining key open issues and research directions.关键词：Artificial Intelligence-native Radio Access Networks (AI-RAN);Distributed learning;Edge inference;Service management and orchestration;AI lifecycle35|17|0更新时间：2026-03-10
- 摘要：The advent of the sixth generation (6G) wireless networks envisions an Artificial Intelligence (AI)-native Radio Access Network (AI-RAN), where Deep Reinforcement Learning (DRL) emerges as a key enabler for intelligent and autonomous network slicing. Despite the demonstrated performance gains of DRL-based solutions in dynamic resource allocation and slice orchestration, their opaque decision-making nature raises critical concerns regarding trust, accountability, and operational deployment. To bridge this gap, Explainable Deep Reinforcement Learning (XDRL) has recently attracted significant attention as a means to enhance transparency, interpretability, and controllability of AI-RAN slicing policies. This survey provides a comprehensive overview of the state of the art in explainable DRL for intelligent network slicing. We first review the fundamental principles of DRL in the context of RAN slicing and identify the unique explainability challenges posed by high-dimensional, multi-slice environments. We then categorize existing XDRL approaches into post-hoc explanation, symbolic abstraction, and human-in-the-loop steering, analyzing their methodologies, strengths, and limitations. Furthermore, we highlight benchmark environments and experimental testbeds that have been employed to evaluate XDRL in realistic network scenarios. Finally, we outline key open challenges, including scalability, generalization across traffic patterns, integration with Large Language Models (LLMs), and alignment with intent-based networking, and discuss promising research directions toward achieving transparent, trustworthy, and human-centric AI-RAN in 6G.关键词：Explainable Deep Reinforcement Learning (XDRL);6G AI-RAN;Intelligent network slicing;Transparency and interpretability;Intent-based networking38|20|0更新时间：2026-03-10
- 摘要：In the evolution from 5G to 6G, network slicing faces challenges due to the coexistence of multiple operators and service providers, where privacy constraints on configuration data and parameters restrict end-to-end slice deployment. This issue is more pronounced in the 6G AI-RAN architecture, especially when cross-domain coordination between the Radio Access Network (RAN) and Core Network (CN) is insufficient, leading to inefficient slice orchestration and management. To address these challenges, we propose three solutions: (1) A non-collaborative cross-domain slicing orchestration architecture for 6G AI-RAN, enabling an adaptive orchestration framework driven by slice metric prediction; (2) a cross-domain Key Performance Indicator (KPI) prediction mechanism combining Graph Convolutional and Attention Networks (GCN-GAT) with a Transformer model to analyze the impact of cross-domain features on RAN slice performance; and (3) an intelligent RAN slice resource orchestration strategy driven by metric prediction. Experiments on real multi-domain datasets demonstrate that the proposed framework outperforms baseline methods by improving end-to-end slice service quality, while significantly enhancing system throughput and reducing both resource overhead and average load.关键词：Radio Access Network (RAN) slicing;Key Performance Indicator (KPI) prediction;Graph Convolutional Network (GCN);Transformer;AI-Radio Access Network (RAN)30|18|0更新时间：2026-03-10
- 摘要：The emergence of 6G native AI network offers new opportunities to support accuracy-critical and latency-sensitive AI inference tasks in Internet-of-Vehicles (IoV) scenarios. However, existing computing offloading schemes often treat the network merely as a communication pipeline or edge computing node, lacking the joint scheduling of the communication resource and the computational resource in a native AI infrastructure. Besides, the deterioration of the Vehicle-to-Infrastructure (V2I) channel is likely to decrease the inference accuracy of AI tasks by decreasing the quality of data offloading. To this end, this paper considers a road lane detection task in IoV scenario and proposes a Quality of AI Service (QoAIS)-guaranteed AI service offloading architecture. Firstly, the functionality between inference accuracy and Signal-to-Noise Ratio (SNR) under different Modulation and Coding Schemes (MCS) is established by the numerical experiments. On this basis, a cross-layer optimization framework is introduced to maximize the number of QoAIS-guaranteed tasks by jointly optimizes MCS selection, uplink bandwidth allocation, and computing resource allocation, simultaneously. A Particle Swarm Optimization (PSO) algorithm is introduced to solve this problem. Numerical results show that the proposed PSO algorithm can significantly improve the number of QoAIS-guaranteed tasks compared with that of the baselines.关键词：6G native-AI network;Internet-of-Vehicles (IoV);AI service offloading;QoAIS guarantee;Resource allocation;Cross-layer optimization33|15|0更新时间：2026-03-10
查看更多
视频讲座
-
00:52:56信息与智能大讲堂第10讲（施巍松教授，美国特拉华大学）：Vehicle Computing: Vision and Challenges
2023-12-04播放量： 37 - 00:49:11
信息与智能大讲堂第11讲（杨鲲教授，南京大学）：机器学习使能的6G内生智能
2023-12-01播放量： 64 - 01:35:18
韩军功教授：浅谈计算机、人工智能方向科技论文撰写技巧（学术研究与论文写作讲座 第11讲）
2023-09-18播放量： 130 - 01:39:24
学术研究与论文写作 第12讲（周一青教授，中国科学院计算所）：浅谈科技论文写作
2023-12-03播放量： 112 - 01:03:05
Prof. Giuseppe Caire: Integrated Sensing and Communications: Information Theory, Communication Theory, and Machine Learning（信息与智能大讲堂 第7讲）
2023-07-24播放量： 32
优秀论文
优秀审稿专家
年度会议集锦
- 地址：陕西省西安市太白南路2号 西安电子科技大学 邮编：710071
- 联系电话：029-88201709 传真：029-88201709 Email：intecom@ynyjxx.cn
- 技术支持由北京北大方正电子有限公司提供 京ICP备09064830号-19
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies
0