ISO/AWI 24635-3

Language resource management — Corpus annotation project management — Part 3: Validation Model

Apstrakt

This document specifies validation models and procedures for corpus annotation project management. It establishes quality assurance guidelines for both human and automated annotation, including Large Language Models (LLMs).

It covers:
• cross-validation methods within annotator groups.
• framework for third-party validation.
• framework for Human-AI collaborative validation.
• procedures for Continuous qualification of personnel: Ongoing monitoring and assessment of annotators and validators to maintain quality throughout the project lifecycle.
Excluded:
This document does not cover:
• specific linguistic annotation methodologies or schemes;
• core project management processes (addressed in ISO 24635-1);
• foundational training and initial qualification programs (addressed in ISO 24635-2).