Research map / Graph learning / Knowledge and temporal graphs
Neural Snowball for Few-Shot Relation Learning
Tianyu Gao, Xu Han, Ruobing Xie, Zhiyuan Liu, Fen Lin, Leyu Lin, Maosong Sun
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 |
|---|---|---|---|
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of direct RSN scoring | Few-shot relation learning with CNN · FewRel, 5 seed instances | F1: 36.96 -> 41.30 | Improved |
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of distant supervision | Few-shot relation learning with CNN · FewRel, 10 seed instances | F1: 45.35 -> 49.30 | Improved |
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of direct fine-tuning | Few-shot relation learning with CNN · FewRel, 15 seed instances | F1: 58.46 -> 63.23 | Improved |
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of direct RSN scoring | Few-shot relation learning with BERT · FewRel, 5 seed instances | F1: 44.47 -> 47.26 | Improved |
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of direct fine-tuning | Few-shot relation learning with BERT · FewRel, 10 seed instances | F1: 57.43 -> 61.34 | Improved |
| Use iterative RSN-filtered instance accumulation and classifier fine-tuning instead of direct fine-tuning | Few-shot relation learning with BERT · FewRel, 15 seed instances | F1: 68.43 -> 72.06 | Improved |
| Increase seed instances from 5 to 15 | Fine-tuning, Distant Supervision, and Neural Snowball · FewRel | F1 | Improved |
| Increase seed instances | BREDS and Relational Siamese Network · FewRel | F1 | No clear effect |
| Train with RSN-selected high-confidence instances rather than randomly selected relation instances | Neural Snowball classifier · Chairperson, 5-seed setting | precision | Improved |
| Use RSN-filtered snowball expansion | Neural Snowball classifier · Chairperson, 5-seed setting | recall | Hurt |
Closest papers by meaning
- Few-shot Relation Extraction via Bayesian Meta-learning on Relation Graphs (ICML 2020 · similarity 0.88)
- Hierarchical Relational Learning for Few-Shot Knowledge Graph Completion (ICLR 2023 · similarity 0.88)
- Few-Shot Knowledge Graph Completion (AAAI 2020 · similarity 0.88)
- MoEMeta: Mixture-of-Experts Meta Learning for Few-Shot Relational Learning (NeurIPS 2025 · similarity 0.86)
- Few-Shot, No Problem: Descriptive Continual Relation Extraction (AAAI 2025 · similarity 0.86)
- Synergistic Anchored Contrastive Pre-training for Few-Shot Relation Extraction (AAAI 2024 · similarity 0.85)
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.
