Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Few-shot Relational Reasoning via Connection Subgraph Pretraining
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 |
|---|---|---|---|
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · NELL transductive | MRR: 0.490 -> 0.577 | Helped |
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · FB15K-237 transductive | MRR: 0.805 -> 0.781 | Hurt |
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · ConceptNet transductive | MRR: 0.577 -> 0.606 | Helped |
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · NELL inductive | MRR: 0.353 -> 0.511 | Helped |
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · FB15K-237 inductive | MRR: 0.453 -> 0.624 | Helped |
| Use CSR-GNN instead of the strongest listed baseline | Few-shot KG completion · ConceptNet inductive | MRR: 0.402 -> 0.611 | Helped |
| Add entity embeddings (ablation) | CSR-GNN on NELL · NELL | MRR: 0.511 -> 0.577 | Helped |
| Add finetuning (ablation) | CSR-GNN on NELL without entity embeddings · NELL | MRR: 0.466 -> 0.511 | Helped |
| Add hypothesis proposal (ablation) | CSR-GNN on NELL without entity embeddings or finetuning · NELL | MRR: 0.441 -> 0.466 | Helped |
| Add evidence proposal (ablation) | CSR-GNN on NELL without entity embeddings or finetuning · NELL | MRR: 0.436 -> 0.466 | Helped |
| Replace the original NELL test tasks with randomly sampled test relations | CSR versus FSRL · NELL distribution-shift test | MRR: 0.590 vs 0.540; 0.578 vs 0.459 | Mixed |
Closest papers by meaning
- Towards Graph Foundation Models: Training on Knowledge Graphs Enables Transferability to General Graphs (NeurIPS 2025 · similarity 0.88)
- Context-aware Inductive Knowledge Graph Completion with Latent Type Constraints and Subgraph Reasoning (AAAI 2025 · similarity 0.88)
- MoEMeta: Mixture-of-Experts Meta Learning for Few-Shot Relational Learning (NeurIPS 2025 · similarity 0.87)
- Replacing Paths with Connection-Biased Attention for Knowledge Graph Completion (AAAI 2025 · similarity 0.86)
- Towards Global-Topology Relation Graph for Inductive Knowledge Graph Completion (AAAI 2025 · similarity 0.86)
- Towards Foundation Models for Knowledge Graph Reasoning (ICLR 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.