Papers › GraphMix: Improved Training of GNNs for Semi-Supervised Learning

GraphMix: Improved Training of GNNs for Semi-Supervised Learning

25 Sep 2019arXiv:1909.11715archive 2025-07-28

Vikas Verma, Meng Qu, Kenji Kawaguchi, Alex Lamb, Yoshua Bengio, Juho Kannala, Jian Tang

We present GraphMix, a regularization method for Graph Neural Network based semi-supervised object classification, whereby we propose to train a fully-connected network jointly with the graph neural network via parameter sharing and interpolation-based regularization. Further, we provide a theoretical analysis of how GraphMix improves the generalization bounds of the underlying graph neural network, without making any assumptions about the "aggregation" layer or the depth of the graph neural networks. We experimentally validate this analysis by applying GraphMix to various architectures such as Graph Convolutional Networks, Graph Attention Networks and Graph-U-Net. Despite its simplicity, we demonstrate that GraphMix can consistently improve or closely match state-of-the-art performance using even simpler architectures such as Graph Convolutional Networks, across three established graph benchmarks: Cora, Citeseer and Pubmed citation network datasets, as well as three newly proposed datasets: Cora-Full, Co-author-CS and Co-author-Physics.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

vikasverma1077/GraphMix officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Generalization BoundsGraph AttentionGraph Neural NetworkNode Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Node Classification Bitcoin-Alpha GraphMix (GCN) F1-score 0.6534 #1 of 1 Archive leaderboard report
Node Classification Bitcoin-OTC GraphMix (GCN) F1-score 0.6635 #1 of 1 Archive leaderboard report
Node Classification CiteSeer with Public Split: fixed 20 nodes per class GraphMix(GCN) Accuracy 74.52 ± 0.59 #6 of 40 Archive leaderboard report
Node Classification CiteSeer with Public Split: fixed 5 nodes per class GraphMix (GCN) Accuracy 58.55 ± 2.26 #1 of 1 Archive leaderboard report
Node Classification Citeseer random partition GraphMix (GCN) Accuracy 76.45 ± 1.57 #1 of 1 Archive leaderboard report
Node Classification Coauthor CS GraphMix (GCN) Accuracy 91.83 ± 0.51 #21 of 24 Archive leaderboard report
Node Classification Coauthor Physics GraphMix (GCN) Accuracy 94.49 ± 0.84 #13 of 14 Archive leaderboard report
Node Classification Cora Full-supervised GraphMix (GCN) Accuracy 61.8% #9 of 9 Archive leaderboard report
Node Classification Cora random partition GraphMix (GCN) Accuracy 82.07 ± 1.17 #1 of 1 Archive leaderboard report
Node Classification Cora with Public Split: fixed 20 nodes per class GraphMix Accuracy 83.94 ± 0.57 #14 of 36 Archive leaderboard report
Node Classification Cora with Public Split: fixed 20 nodes per class GraphMix (GCN) Accuracy 83.94 ± 0.57 #15 of 36 Archive leaderboard report
Node Classification Cora: fixed 10 node per class GraphMix (GCN) Accuracy 79.3 #2 of 2 Archive leaderboard report
Node Classification Cora: fixed 5 node per class GraphMix (GCN) Accuracy 71.99 ± 6.46 #2 of 2 Archive leaderboard report
Node Classification PubMed with Public Split: fixed 20 nodes per class GraphMix (GCN) Accuracy 80.98 ± 0.55 #9 of 37 Archive leaderboard report
Node Classification PubMed with Public Split: fixed 20 nodes per class GCN(predicted-targets) Accuracy 80.42% #13 of 37 Archive leaderboard report
Node Classification Pubmed random partition GraphMix (GCN) Accuracy 80.72 ± 1.08 #1 of 1 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Graph Convolutional NetworksGraph Neural Network

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections