发表机构
College of Future Information Technology, Fudan University; Division of Natural and Applied Sciences, Duke Kunshan University; Department of Electrical and Computer Engineering, The University of British Columbia; Ant Group; College of Information Science and Electronic Engineering, Zhejiang University; School of Data Science and Engineering, East China Normal University; College of Intelligent Robotics and Advanced Manufacturing, Fudan University & Fysics AI; Xiaomi EV, Xiaomi Campus; Information and Communications Technology Cluster, Singapore Institute of Technology (SIT); College of Electronic and Information Engineering, Tongji University; School of Computer Science and Informatics, Cardiff University（复旦大学未来信息技术学院; 昆山杜克大学自然科学与应用科学部; 英属哥伦比亚大学电气与计算机工程系; 蚂蚁集团; 浙江大学信息科学与电子工程学院; 华东师范大学数据科学与工程学院; 复旦大学智能机器人与先进制造学院及复肆智能科技（上海）有限公司; 小米汽车、小米园区; 新加坡理工学院信息通信技术集群; 同济大学电子与信息工程学院; 卡迪夫大学计算机科学与信息学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对分布式网络中联邦微调平衡全局知识与局部适应的挑战，提出PFAdapter框架，用分层LoRA分解将适配器参数分离，通过正交正则化和选择性聚合协议减少通信成本，实验证明该框架在多数据集上性能优于基线，为智能AI部署提供有效方案。
AI 中文摘要
智能AI系统通过部署能协作学习并保持数据隐私的自主智能代理重塑通信和网络。在分布式网络环境中，多模态大语言模型是边缘设备的认知引擎，但联邦微调在异构网络条件下平衡全局知识聚合与局部适应方面面临挑战。传统联邦协议依赖统一参数聚合，导致个性化欠佳和通信开销大。为此提出PFAdapter框架，引入分层LoRA分解，将适配器参数分为全局共享和局部私有组件，通过正交正则化防止组件间冗余特征学习，选择性聚合协议仅同步全局共享组件，减少近50%通信成本。在多个数据集上实验表明，PFAdapter性能优于现有基线，在不同边缘智能任务上准确率提高2.4%至4.8%，为资源受限通信网络中的智能AI部署提供了有效解决方案。
英文摘要
Agentic AI systems are reshaping communications and networking by deploying autonomous intelligent agents capable of collaborative learning while maintaining data privacy at network edges. Within distributed network environments, Multimodal Large Language Models (MLLMs) serve as cognitive engines for edge devices, yet federated fine-tuning faces substantial challenges in balancing global knowledge aggregation with local adaptation under heterogeneous network conditions. Conventional federated protocols typically rely on uniform parameter aggregation, which conflates domain-invariant features with client-specific nuances, thereby resulting in suboptimal personalization and excessive communication overhead. To address these challenges, we propose PFAdapter, a communication-efficient framework introducing hierarchical LoRA decomposition to explicitly separate adapter parameters into global-shared and local-private components. Query and key projections are assigned to global synchronization for capturing universal multimodal semantics across the network, while value and output projections remain localized for edge-specific adaptation. Additionally, orthogonality regularization based on the Frobenius norm enforces strict separation between these components, preventing redundant feature learning. Selective aggregation protocols synchronize only global-shared components across the federated network, preserving local expertise and reducing communication costs by nearly 50%. Extensive experiments on VQA-RAD, SLAKE, Hateful Memes, and CrisisMMD datasets demonstrate that PFAdapter consistently outperforms state-of-the-art baselines, achieving accuracy improvements ranging from 2.4% to 4.8% across diverse edge intelligence tasks. Consequently, our framework establishes an efficient solution for agentic AI deployment in resource-constrained communication networks.
CommentsAccepted by IEEE TCCN
