LEC-KG: An LLM-Embedding Collaborative Framework for Domain-Specific Knowledge Graph Construction -- A Case Study on SDGs

Researchers developed LEC-KG, a new framework that combines Large Language Models with Knowledge Graph Embeddings to better extract and structure information from unstructured text. The system was tested on Chinese Sustainable Development Goal reports and showed significant improvements over traditional LLM approaches, particularly for identifying rare relationships in domain-specific content.

• →LEC-KG framework combines LLMs with Knowledge Graph Embeddings for better domain-specific knowledge extraction.
• →The system uses hierarchical relation extraction to address long-tail bias in data distributions.
• →Evidence-guided Chain-of-Thought feedback grounds AI suggestions in actual source text.
• →Testing on Chinese SDG reports showed substantial improvements over baseline LLM methods.
• →The framework enables iterative refinement where both AI components enhance each other's performance.