Technical articles · 2024.05
Data validation stages for generative AI trustworthiness
생성형 AI 신뢰성 확보를 위한 데이터 검증
Overview
The article explains data validation across collection/preprocessing, training, evaluation and deployment, focusing on text-based generative AI. It connects provenance and accuracy checks, duplicate/missing-data handling, bias review, benchmark evaluation and post-deployment monitoring.
Publication and version information
General LLM data-validation article in TTA Journal 213, May/June 2024, pp. 42–49. The authors’ safety discussion concerns matters such as harmful and ethical outputs. The PDF prints the first author as 김민우; the publisher’s index instead lists 김인우.
Version: TTA Journal 213, pp. 42–49