Researchers have developed CONSISTRE, a new framework designed to improve document-level relation extraction (DocRE) using large language models (LLMs). The framework addresses the issue of LLMs producing inconsistent or contradictory predictions by incorporating consistency checks. CONSISTRE offers two approaches: one for black-box LLMs that refines predictions through prompting and self-reflection, and another for open-source models that uses knowledge distillation and reinforcement learning to inject consistency knowledge. Experiments on the DocRED dataset demonstrate that both methods enhance the reliability of relation extraction. AI
IMPACT Enhances the reliability of LLM-based relation extraction, potentially improving information retrieval and knowledge graph construction.
RANK_REASON The cluster describes a new framework and methodology presented in an academic paper.
Read on Hugging Face Daily Papers →
- alphaXiv
- CatalyzeX
- CONSISTRE
- DagsHub
- DocRED
- Gotit.pub
- Grpo
- Hugging Face
- Large Language Models
- ScienceCast
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
