Research map / Graph learning / Graph self-supervised learning
Self-Supervised Graph Transformer on Large-Scale Molecular Data
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add GROVER self-supervised pre-training (ablation) | GROVER · 6 classification datasets avg | ROC-AUC: 0.803 -> 0.834 | Helped |
| add GROVER self-supervised pre-training (ablation) | GROVER · ClinTox | ROC-AUC: 0.884 -> 0.944 | Helped |
| add GROVER self-supervised pre-training (ablation) | GROVER · FreeSolv | RMSE: 1.987 -> 1.544 | Helped |
| GROVER large pretraining vs previous SOTA | molecular property prediction · 11 MoleculeNet datasets | relative improvement: 6.1% | Helped |
| scale model from base (48M) to large (100M) (ablation) | GROVER · BACE | ROC-AUC: 0.878 -> 0.894 | Helped |
| scale model from base to large (ablation) | GROVER · QM7 | MAE: 72.5 -> 72.6 | No clear effect |
| GROVER self-supervised tasks instead of Hu et al. pretraining on GIN | GIN · BBBP/SIDER/BACE avg | ROC-AUC: 0.793 -> 0.812 | Helped |
| use GTransformer instead of vanilla Transformer (ablation) | GROVER · 600K toy pretraining set | validation loss | Helped |
| add dynamic message passing (dyMPN) (ablation) | GROVER · 600K toy pretraining set | validation loss | Helped |
| GTransformer backbone instead of GIN/MPNN (ablation) | GROVER · 600K toy pretraining set | validation loss | Helped |
Closest papers by meaning
- Pre-Training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck (AAAI 2025 · similarity 0.87)
- Neural Atoms: Propagating Long-range Interaction in Molecular Graphs through Efficient Communication Channel (ICLR 2024 · similarity 0.87)
- On the Scalability of GNNs for Molecular Graphs (NeurIPS 2024 · similarity 0.87)
- Graph Generative Pre-trained Transformer (ICML 2025 · similarity 0.87)
- Pushing the Limits of All-Atom Geometric Graph Neural Networks: Pre-Training, Scaling, and Zero-Shot Transfer (ICLR 2025 · similarity 0.87)
- Triplet Interaction Improves Graph Transformers: Accurate Molecular Graph Learning with Triplet Graph Transformers (ICML 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.