Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

GCC: Graph Contrastive Coding for Graph Neural Network Pre-Training

About

Graph representation learning has emerged as a powerful technique for addressing real-world problems. Various downstream graph learning tasks have benefited from its recent developments, such as node classification, similarity search, and graph classification. However, prior arts on graph representation learning focus on domain specific problems and train a dedicated model for each graph dataset, which is usually non-transferable to out-of-domain data. Inspired by the recent advances in pre-training from natural language processing and computer vision, we design Graph Contrastive Coding (GCC) -- a self-supervised graph neural network pre-training framework -- to capture the universal network topological properties across multiple networks. We design GCC's pre-training task as subgraph instance discrimination in and across networks and leverage contrastive learning to empower graph neural networks to learn the intrinsic and transferable structural representations. We conduct extensive experiments on three graph learning tasks and ten graph datasets. The results show that GCC pre-trained on a collection of diverse datasets can achieve competitive or better performance to its task-specific and trained-from-scratch counterparts. This suggests that the pre-training and fine-tuning paradigm presents great potential for graph representation learning.

Jiezhong Qiu, Qibin Chen, Yuxiao Dong, Jing Zhang, Hongxia Yang, Ming Ding, Kuansan Wang, Jie Tang• 2020

Related benchmarks

TaskDatasetResultRank
Graph ClassificationPROTEINS
Accuracy64.2
1252
Node ClassificationPubmed--
865
Graph ClassificationCOLLAB
Accuracy78.9
469
Graph ClassificationIMDB-M
Accuracy49.4
425
Graph ClassificationIMDB-B
Accuracy72
425
Node ClassificationarXiv
Accuracy26.84
254
Graph ClassificationMUTAG (10-fold cross-validation)
Accuracy74.4
227
Graph ClassificationPROTEINS (10-fold cross-validation)
Accuracy69.5
223
Node ClassificationREDDIT
Accuracy65.21
216
Node ClassificationComputers
Mean Accuracy62.63
169
Showing 10 of 41 rows

Other info

Follow for update