FedSDR：带结构噪声检测与重建的联邦图学习
FedSDR: Federated Graph Learning with Structural Noise Detection and Reconstruction
对视觉研究者而言，可复用的是「用谱域结构保真度指标识别被噪声污染的客户端并降权」这一思路，可迁移到多源视觉数据（如跨摄像头、跨传感器）联合训练时对低质量来源的加权聚合，以及用健康全局模型重建局部损坏图结构的做法。
原文依据 Structural Noise-Aware Aggregation (SNAA) introduces a structural fidelity evaluation metric to detect corrupted clients
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论联邦图学习（FGL）已成为在保护数据隐私的前提下分布式训练图神经网络（GNN）的原则性框架。然而，在子图级联邦学习场景中，由数据采集与存储产生的结构噪声会破坏客户端的 GNN 消息传递机制，导致协作冲突。现有方法存在两个关键局限：1）在全局层面，它们无法识别被污染的客户端，造成破坏性的知识不一致；2）在局部层面，全局 GNN 在这些客户端上因结构噪声而表现不佳，限制了它们从联邦协作中获益的能力。为应对这些挑战，我们提出 FedSDR，一个面向高结构噪声场景的基于谱的 FGL 框架。具体而言，结构噪声感知聚合（SNAA）引入结构保真度评估指标来检测被污染的客户端并降低其贡献，从而缓解噪声对全局 GNN 的影响。此外，鲁棒局部结构重建（RLSR）利用健康的全局模型知识来修复局部损坏的图结构。大量实验表明，FedSDR 在结构噪声下的多种场景中均优于最优方法。代码见 https://github.com/Subtleazure/FedSDR。
英文原文摘要 1346 字符
Federated Graph Learning (FGL) has emerged as a principled framework for decentralized training of Graph Neural Networks (GNNs) while preserving data privacy. In subgraph-FL scenarios, however, structural noise arising from data collection and storage can damage the GNN message-passing scheme of clients, leading to conflicts in collaboration. Existing approaches exhibit two critical limitations: 1) Globally, they fail to identify corrupted clients, causing destructive knowledge inconsistencies. 2) Locally, the global GNN performs poorly on these clients due to structural noise, limiting their ability to benefit from federated collaboration. To address these challenges, we propose FedSDR, a spectra-based FGL framework against high-structural-noise scenarios. Specifically, Structural Noise-Aware Aggregation (SNAA) introduces a structural fidelity evaluation metric to detect corrupted clients and reduce their contributions, thereby mitigating the impact of noise on the global GNN. Furthermore, Robust Local Structure Reconstruction (RLSR) leverages the knowledge from the healthy global model to repair locally corrupted graph structures. Extensive experiments demonstrate that FedSDR outperforms state-of-the-art methods across various scenarios under structural noise. The code is available at https://github.com/Subtleazure/FedSDR.