CyBOKClaw: Human-in-the-Loop Framework for Cybersecurity Curriculum Mapping
CyBOKClaw is a human-in-the-loop retrieval system that interprets cybersecurity keywords or phrases (KWoPs) and connects them to the Cyber Security Body of Knowledge (CyBOK). This framework serves as a generator for top-k candidates to be reviewed by experts, incorporating query normalization, curated term expansion, concept-level enhancements, enriched topic descriptions, and domain-sensitive ranking criteria. It tackles the issue of educational KWoPs being often vague, broad, and only loosely related to CyBOK terminology, rendering strict exact matching inadequate. The evaluation employs structural retrieval metrics alongside an expert-driven usefulness metric called ECA-5 (Exact or Closest Acceptable Match at top-5), which assesses whether the returned candidates include at least one mapping deemed exact or acceptable by an expert.
Key facts
- CyBOKClaw is a human-in-the-loop retrieval framework.
- It maps cybersecurity keywords or phrases (KWoPs) to CyBOK.
- The framework is a top-k candidate generator for expert review.
- It uses query normalization, term expansion, concept boosts, topic enrichment, and domain-sensitive ranking.
- Educational KWoPs are often broad, ambiguous, and only approximately aligned with CyBOK.
- Strict exact matching provides only partial practical utility.
- Evaluation uses structural retrieval metrics and ECA-5.
- ECA-5 records whether top-5 candidates contain at least one exact or acceptable match.
Entities
Institutions
- CyBOK
- arXiv