Scholarlink

Job Descriptions are a common currency for hiring talent. Whilst they are ubiquitous, they are also unstructured, difficult to analyse, and have no common format. In this paper, we present a novel approach to generate knowledge graphs from job descriptions by utilising a bespoke Large Language Model to extract and reconcile concepts to our ground truth taxonomy of HR entities.

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

Rapid Graph Generation from Job Descriptions: Combining Taxonomies and LLMs

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Kaan Karakeben,
  • Henri Egle Sorotos,
  • Alisa Milchevskaya,
  • Ahmad Assaf

摘要

Job Descriptions are a common currency for hiring talent. Whilst they are ubiquitous, they are also unstructured, difficult to analyse, and have no common format. In this paper, we present a novel approach to generate knowledge graphs from job descriptions by utilising a bespoke Large Language Model to extract and reconcile concepts to our ground truth taxonomy of HR entities.