MearcAll papers

arXiv 1706.02216 · cs.SI cs.LG stat.ML

Inductive Representation Learning on Large Graphs

William L. Hamilton, Rex Ying, Jure Leskovec

Claims were extracted automatically by reducer-1 on 2026-11-12. The paper's license does not allow republishing it, so long claims are shortened to a quotation. Read the full text on arXiv.

No reproduction recordedclm_ef6bc047
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.

Abstract · page 1 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_5308a8f5
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 … (excerpt)

Abstract · page 1 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_b23a4e50
We propose a general framework, called GraphSAGE (SAmple and aggreGatE), for inductive node embedding.

Introduction · page 2 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_9678d09a
Using these benchmarks, we show that our approach is able to effectively generate representations for unseen nodes and outperform relevant baselines by a significant margin: across domains, our supervised approach … (excerpt)

Introduction · page 2 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_510eef12
We also show that the new aggregator architectures we propose provide significant gains (7.4% on average) compared to an aggregator inspired by graph convolutional networks [17].

Introduction · page 2 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_d569d856
3 Algorithm 1: GraphSAGE embedding generation (i.e., forward propagation) algorithm Input : GraphG(V,E); input features{xv,∀v∈V} ; depthK; weight matrices Wk,∀k∈{ 1,...,K}; non-linearityσ; differentiable aggregator functions AGGREGATE k,∀k∈{ 1,...,K}; neighborhood functionN … (excerpt)

Proposed method: GraphSAGE · page 3 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_06a77707
Practically speakingwe found that our approach could achieve high performance with K = 2 andS1·S2≤ 500 (see Section 4.4 for details).

Proposed method: GraphSAGE · page 5 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_526e0e61
We find that GraphSAGE outperforms all the baselines by a significant margin, and the trainable, neural network aggregators provide significant gains compared 7 to the GCN approach.

Inductive learning on evolving graphs: Citation and Reddit data · page 7 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_cc38b0e6
For example, the unsupervised variant GraphSAGE-pool outperforms the concatenation of the DeepWalk embeddings and the raw features by 13.8% on the citation data and 29.1% on the Reddit data, while … (excerpt)

Inductive learning on evolving graphs: Citation and Reddit data · page 8 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_27538542
Again we see that GraphSAGE significantly outperforms the baseline approaches, with the LSTM- and pooling-based aggregators providing substantial gains over the mean- and GCN-based aggregators.6

Generalizing across graphs: Protein-protein interactions · page 8 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_3af469e3
Note that this method is rank-based and essentially tests whether we would expect one particular approach to outperform another in a new experimental setting.

Summary comparison between the different aggregator architectures · page 8 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_4bd83899
However, the gains of the LSTM and pool approaches over the mean-based aggregator are more marginal (T = 1.5,p = 0.03, comparing 6Note that in very recent follow-up work Chen … (excerpt)

Summary comparison between the different aggregator architectures · page 8 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_51264f1f
We refer the reader to their work for the current state-of-the-art numbers on the PPI dataset that are possible using a variant of the GraphSAGE approach.

Summary comparison between the different aggregator architectures · page 8 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_2e48cc6c
The proof of Theorem 1 relies on some properties of the pooling aggregator, which also provides insight into why GraphSAGE-pool outperforms the GCN and mean-based aggregators.

Theoretical analysis · page 9 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_54d75155
We introduced a novel approach that allows embeddings to be efficiently generated for unseen nodes.

Conclusion · page 9 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_71076688
GraphSAGE consistently outperforms state-of-the-art baselines, effectively trades off performance and runtime by sampling node neighborhoods, and our theoretical analysis provides insight into how our approach can learn about local graph … (excerpt)

Conclusion · page 9 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_bc8e0d20
We found that DeepWalk achieved faster wall-clock convergence with a smaller batch size of 64.

Details on the Experimental Setup and Hyperparameter Tuning · page 14 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedres_9fcfd6db
In particular, the Reddit data has fewer edges from the test set to the train set, which help prevent mis-alignment: 96% of the 2005 citation links connect back to the … (excerpt)

Alignment Issues and Orthogonal Invariance for DeepWalk and Related · page 16 · Automated extraction, not reviewed · 0 reviewed reproductions

No reproduction recordedclm_5a486277
We show that there exists a parameter setting for which Algorithm 1 can learn nodes clustering coefficients, which is non-obvious given that it operates by aggregating feature information.

Proof of Theorem 1 · page 16 · Automated extraction, not reviewed · 0 reviewed reproductions

Reproduced one of these claims?

Record your result in the Mearc app on your own machine, then use Publish this result with this paper's id (1706.02216) and the claim id shown next to the claim. Mearc creates a file that holds only what you choose to publish. Add it to the public claim data by pull request; a maintainer other than you reviews it before it counts.