Scholarlink

This paper presents our progress towards creating an end-to-end query generation approach. The resulting pairs of NL and executable queries can be used for setting up KGQA datasets and benchmarking LLMs. The usage of a rigorous theoretical background will ensure that queries are generated with predeterminable complexities.

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

Benchmarking Knowledge Graph Question Answering via Complexity-Aware Queries

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Balázs Mosolygó,
  • Andreas L. Opdahl,
  • Guohui Xiao,
  • Jiaoyan Chen,
  • Ana Ozaki

摘要

This paper presents our progress towards creating an end-to-end query generation approach. The resulting pairs of NL and executable queries can be used for setting up KGQA datasets and benchmarking LLMs. The usage of a rigorous theoretical background will ensure that queries are generated with predeterminable complexities.