Augmentation-Free Self-Supervised Learning on Graphs
About
Inspired by the recent success of self-supervised methods applied on images, self-supervised learning on graph structured data has seen rapid growth especially centered on augmentation-based contrastive methods. However, we argue that without carefully designed augmentation techniques, augmentations on graphs may behave arbitrarily in that the underlying semantics of graphs can drastically change. As a consequence, the performance of existing augmentation-based methods is highly dependent on the choice of augmentation scheme, i.e., hyperparameters associated with augmentations. In this paper, we propose a novel augmentation-free self-supervised learning framework for graphs, named AFGRL. Specifically, we generate an alternative view of a graph by discovering nodes that share the local structural information and the global semantics with the graph. Extensive experiments towards various node-level tasks, i.e., node classification, clustering, and similarity search on various real-world datasets demonstrate the superiority of AFGRL. The source code for AFGRL is available at https://github.com/Namkyeong/AFGRL.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Molecular property prediction | MoleculeNet BBBP (scaffold) | ROC AUC70.8 | 117 | |
| Molecular property prediction | MoleculeNet SIDER (scaffold) | ROC-AUC0.602 | 97 | |
| Molecular property prediction | MoleculeNet BACE (scaffold) | ROC-AUC79.3 | 87 | |
| Molecular property prediction | MoleculeNet MUV (scaffold) | ROC-AUC0.777 | 68 | |
| Molecular property prediction | TOXCAST (scaffold) | ROC-AUC63.1 | 52 | |
| Toxicity Prediction | Tox21 (scaffold) | AUC0.749 | 46 | |
| Molecular Property Classification | ClinTox (scaffold) | ROC-AUC0.791 | 42 | |
| Molecular Property Classification | HIV (scaffold) | ROC AUC0.76 | 25 | |
| BP Classification | Breast Cancer GRN | Subset Accuracy24 | 11 | |
| Hazard Prediction | Breast Cancer GRN | C-Index0.616 | 11 |