Research map / Graph learning / Graph self-supervised learning
ReCoG: Relational and Compact Context Graph Learning for Few-shot Molecular Property Prediction
Zeyu Wang, Xin Zheng, Yao Lu, Shanqing Yu, Qi Xuan, Shirui Pan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 CPRL and CGIB (ablation) | ReCoG · Tox21 10-shot | ROC-AUC (%): 90.21 -> 93.52 | Improved |
| add CPRL and CGIB (ablation) | ReCoG · Tox21 1-shot | ROC-AUC (%): 86.91 -> 91.01 | Improved |
| add CPRL and CGIB (ablation) | ReCoG · SIDER 1-shot | ROC-AUC (%): 86.93 -> 90.06 | Improved |
| add CPRL (cross-property relation learning) (ablation) | ReCoG · Tox21 1-shot | ROC-AUC (%): 88.18 -> 91.01 | Improved |
| add CPRL (cross-property relation learning) (ablation) | ReCoG · SIDER 10-shot | ROC-AUC (%): 91.60 -> 93.98 | Improved |
| add CGIB (context graph information bottleneck) (ablation) | ReCoG · Tox21 1-shot | ROC-AUC (%): 90.28 -> 91.01 | Improved |
| add CGIB (context graph information bottleneck) (ablation) | ReCoG · SIDER 10-shot | ROC-AUC (%): 91.54 -> 93.98 | Improved |
| use MSE instead of BCE for CPRL relation loss (ablation) | ReCoG · Tox21 1-shot | ROC-AUC (%): 88.79 -> 91.01 | Improved |
| use MSE instead of BCE for CPRL relation loss (ablation) | ReCoG · SIDER 1-shot | ROC-AUC (%): 87.71 -> 90.06 | Improved |
| set Gumbel-Sigmoid temperature t to 1.0 instead of 0.1 (ablation) | ReCoG · Tox21 10-shot | ROC-AUC (%): 89.37 -> 93.52 | Improved |
| raise Gumbel-Sigmoid temperature t from 1.0 to 2.0 (ablation) | ReCoG · Tox21 10-shot | ROC-AUC (%): 93.52 -> 91.47 | Hurt |
| set Gumbel-Sigmoid temperature t to 1.0 instead of 0.1 (ablation) | ReCoG · SIDER 1-shot | ROC-AUC (%): 85.44 -> 90.06 | Improved |
Closest papers by meaning
- Pin-Tuning: Parameter-Efficient In-Context Tuning for Few-Shot Molecular Property Prediction (NeurIPS 2024 · similarity 0.88)
- MIPT: Multilevel Informed Prompt Tuning for Robust Molecular Property Prediction (ICML 2025 · similarity 0.86)
- Association Pattern-enhanced Molecular Representation Learning (AAAI 2025 · similarity 0.86)
- Knowledge Enhanced Representation Learning for Drug Discovery (AAAI 2024 · similarity 0.86)
- Position: Graph Learning Will Lose Relevance Due To Poor Benchmarks (ICML 2025 · similarity 0.86)
- Monroe: A Molecular Foundation Model for In-context Probabilistic Inference (NeurIPS 2026 · 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.
