講演情報

[8B-04]Ontology-Guided Extraction of Knowledge Graphs

*WANG WEIXI1、WANG DALI1、岩井原 瑞穂1 (1. 早稲田大学大学院情報生産システム研究科データ工学研究室)
発表者区分:学生
論文種別:ショートペーパー
インタラクティブ発表:なし

キーワード:

Knowledge Graph Extraction、Large Language Model

Knowledge graphs (KGs) are important for many tasks in data science and AI fields. KGs are usually incomplete, so KG extraction is necessary for acquiring new triples. Ontologies provide schemas of KGs. Large language models (LLMs) are powerful in NLP tasks including KG extraction. In this study, we propose a novel method for ontology-guided LLM-based KG extraction. A subgraph of the ontology will be extracted and appended to the prompt to instruct the LLM to extract a KG matching with it.