Daekyo announced on the 29th that it successfully completed the ‘2023 AI Training Data Construction Project,’ led by the Ministry of Science and ICT and promoted by the National Information Society Agency (NIA), receiving excellent evaluations.
This project is a government-led initiative aimed at creating and opening AI training data, a core resource of the digital transformation, to foster and normalize the AI ecosystem.
The AI training data constructed by Daekyo through this project includes approximately 300,000 items such as ▲various types of math problems and answers ▲images of learners’ handwritten solution processes.
In July, Daekyo participated as the lead organization of the consortium for this project, building automatic solution data for math subjects and math problem generation data. Utilizing the constructed data, Daekyo developed a total of six AI models and enhanced data quality management and reliability through quality verification by the Telecommunications Technology Association (TTA).
Additionally, Daekyo formed a consortium with specialized companies in various fields such as Definition, Dwitch, Univa, and CrowdWorks to successfully carry out data acquisition and collection, refinement, processing (labeling), AI model development, and quality management.
A Daekyo representative stated, “This project is significant in that it laid the foundation for educational large language models (LLM) by providing high-quality, optimized math problem data to the education industry, which has relatively lacked AI training data. We will continue to actively strive to lead the advancement of AI-based edutech services based on Daekyo’s unique experience and know-how.”
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.


