Low-dimensional embeddings of nodes in large graphs have proved extremely useful in a variety of prediction tasks, from content recommendation to identifying protein functions. However, most existing approaches require that all nodes in the graph are present during training of the embeddings; these previous approaches are inherently transductive and do not naturally generalize to unseen nodes. Here we present GraphSAGE, a general, inductive framework that leverages node feature information (e.g., text attributes) to efficiently generate node embeddings for previously unseen data. Instead of training individual embeddings for each node, we learn a function that generates embeddings by sampling and aggregating features from a node's local neighborhood. Our algorithm outperforms strong baselines on three inductive node-classification benchmarks: we classify the category of unseen nodes in evolving information graphs based on citation and Reddit post data, and we show that our algorithm generalizes to completely unseen graphs using a multi-graph dataset of protein-protein interactions.
核心贡献 · Key contributions
提出 GraphSAGE，一种利用节点特征对未见数据生成节点嵌入的归纳框架。 Proposes GraphSAGE, an inductive framework for generating node embeddings on unseen data using node features.
引入可训练的聚合函数（均值、LSTM、池化），作用于无序邻居集合。 Introduces trainable aggregator functions (mean, LSTM, pooling) that operate on unordered neighbor sets.
在三个归纳节点分类基准上表现优异，比基线提升高达 51%。 Demonstrates strong performance on three inductive node-classification benchmarks, outperforming baselines by up to 51%.
展示 GraphSAGE 在多图蛋白质相互作用数据集上泛化到完全未见图。 Shows GraphSAGE generalizes to completely unseen graphs in a multi-graph protein-protein interaction dataset.
提供理论分析，将 GraphSAGE 与 Weisfeiler-Lehman 同构测试联系起来，并证明其学习局部图结构的能力。 Provides theoretical analysis linking GraphSAGE to the Weisfeiler-Lehman isomorphism test and its ability to learn local graph structure.
通过采样固定大小的邻居集实现高效运行，在 K=2 且 S1*S2≤500 时保持性能。 Achieves efficient runtime by sampling fixed-size neighbor sets, maintaining performance with K=2 and S1*S2≤500.
局限 · Limitations
侧重于特征丰富的图；在无节点特征的图上性能可能下降。 Focuses on feature-rich graphs; performance may degrade on graphs without node features.
邻居采样引入方差；未探索非均匀采样策略。 Neighborhood sampling introduces variance; non-uniform sampling strategies are not explored.
LSTM 聚合器不具备置换不变性，需要随机排序邻居。 LSTM aggregator is not permutation invariant and requires random ordering of neighbors.
聚类系数逼近的理论分析假设节点特征互异且模型高维。 Theoretical analysis of clustering coefficient approximation assumes distinct node features and high-dimensional model.
评估仅限于节点分类；未测试链接预测等其他任务。 Evaluation limited to node classification; link prediction and other tasks are not tested.