FedGraphNN
收藏资源简介:
FedGraphNN是一个开放的联邦学习基准系统,专注于图神经网络(GNN)。该数据集包含36个来自分子、蛋白质、知识图谱、推荐系统、引文网络和社交网络等7个领域的数据集。数据集的创建旨在通过联邦学习保护数据隐私,同时解决大规模图数据集的训练挑战。FedGraphNN支持多种流行的GNN模型和联邦学习算法,通过高效的系统支持,促进了跨领域的研究和应用。
FedGraphNN is an open federated learning benchmark system focused on graph neural networks (GNNs). This dataset suite includes 36 datasets spanning 7 domains, namely molecular science, proteomics, knowledge graphs, recommendation systems, citation networks and social networks. The dataset suite was developed to protect data privacy via federated learning while addressing the training challenges of large-scale graph datasets. FedGraphNN supports a wide range of popular GNN models and federated learning algorithms, and its efficient system infrastructure promotes cross-domain research and applications.

- 1FedGraphNN: A Federated Learning System and Benchmark for Graph Neural Networks南加州大学维特比工程学院 · 2021年



