Scholarlink

In Case-Based Reasoning (CBR), a case-base is a structured knowledge collection, often represented as an ontology with a TBox (conceptual structure) and ABox (concrete cases). This work emphasizes the need for a case-base to integrate knowledge from unstructured texts describing hydro-ecosystem restoration experiences, forming the foundation for a CBR framework to address new restoration challenges. However, populating the case-base from unstructured texts presents challenges due to fragmented, inconsistent, and implicit information. Therefore, we propose leveraging large language models (LLMs) to extract knowledge from unstructured texts, where the ontology population is approached as a generative knowledge extraction task. The results demonstrate that LLMs can effectively extract structured knowledge, facilitating the creation of a case-base for future projects.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

LLM-Driven Case-Base Populating for Structuring and Integrating Restoration Experiences

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Fethi Ghazouani,
  • Franco Giustozzi,
  • Florence Le Ber

摘要

In Case-Based Reasoning (CBR), a case-base is a structured knowledge collection, often represented as an ontology with a TBox (conceptual structure) and ABox (concrete cases). This work emphasizes the need for a case-base to integrate knowledge from unstructured texts describing hydro-ecosystem restoration experiences, forming the foundation for a CBR framework to address new restoration challenges. However, populating the case-base from unstructured texts presents challenges due to fragmented, inconsistent, and implicit information. Therefore, we propose leveraging large language models (LLMs) to extract knowledge from unstructured texts, where the ontology population is approached as a generative knowledge extraction task. The results demonstrate that LLMs can effectively extract structured knowledge, facilitating the creation of a case-base for future projects.