Research map / Large language models / Instruction and fine-tuning
SCIR: A Self-Correcting Iterative Refinement Framework for Enhanced Information Extraction Based on Schema
Yushen Fang, Jianjun Li, Mingqian Ding, Chang Liu, Xinchi Zou, Wenqi Yang
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 |
|---|---|---|---|
| wrap OneKE with SCIR | OneKE · FewFC | Micro-F1: 80.11 -> 85.10 | Improved |
| wrap OneKE with SCIR | OneKE · Boson (zh NER) | Micro-F1: 72.61 -> 68.58 | Hurt |
| wrap OneKE with SCIR | OneKE · SKE2020 | Micro-F1: 72.61 -> 73.81 | No clear effect |
| wrap Qwen3-8B with SCIR | Qwen3-8B · CCF Law | Micro-F1: 34.99 -> 65.86 | Improved |
| wrap Qwen3-8B with SCIR | Qwen3-8B · Weibo | Micro-F1: 19.01 -> 34.11 | Improved |
| wrap RUIE with SCIR | RUIE · FewRel | Micro-F1: 49.93 -> 55.02 | Improved |
| wrap RUIE with SCIR | RUIE · CrossNER | Micro-F1: 65.41 -> 65.54 | No clear effect |
| wrap OneKE with SCIR | OneKE · FewFC and Boson | Micro-F1: FewFC 80.11 -> 85.10; Boson 72.61 -> 68.58 | Mixed |
| add redundancy detection to missing-only detection (ablation) | SCIR-Qwen3 · CCF Law | Micro-F1: 63.62 -> 65.86 | Improved |
| add missing detection to redundancy-only detection (ablation) | SCIR-Qwen3 · COAE2016 | Micro-F1: 42.40 -> 46.96 | Improved |
| add dual-path detection to plain extraction (ablation) | SCIR-Qwen3 · WikiEvents | Micro-F1: 8.27 -> 15.61 | Improved |
| train the Qwen3-4B checker on MBSC instead of using it untuned (ablation) | SCIR checker · WikiEvents, Boson, SKE2020 | F1 | Improved |
Closest papers by meaning
- UniversalNER: Targeted Distillation from Large Language Models for Open Named Entity Recognition (ICLR 2024 · similarity 0.87)
- Enhancing Relation Extraction via Supervised Rationale Verification and Feedback (AAAI 2025 · similarity 0.87)
- STAR: Boosting Low-Resource Information Extraction by Structure-to-Text Data Generation with Large Language Models (AAAI 2024 · similarity 0.86)
- HLMEA: Unsupervised Entity Alignment Based on Hybrid Language Models (AAAI 2025 · similarity 0.85)
- DocKS-RAG: Optimizing Document-Level Relation Extraction through LLM-Enhanced Hybrid Prompt Tuning (ICML 2025 · similarity 0.85)
- Small Language Model Makes an Effective Long Text Extractor (AAAI 2025 · 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.
