Alchera (CEO Young-gyu Hwang) , a vision AI solution specialist, announced that it has been selected as the main organizer of the 'Ultra-Large AI Learning Data Quality Verification' project hosted by the Telecommunications Technology Association (TTA) for the fifth consecutive year.
With this selection, Alchera has established itself as the only company in Korea to perform "semantic accuracy verification" for five consecutive years. This nationally-led project verifies the quality of training data required to advance ultra-large AI models, and aims to increase reliability and usability across the AI industry.
This year's validation focuses on building data for implementing vertical AI based on specialized domain knowledge. Alchera plans to conduct semantic accuracy verification on a total of 31 datasets, including expert domain-based data, Chain of Thought (CoT) data that guides thought processes, and industry-specific data.
In addition to this, Alchera participated as an implementing agency for the 'AI Hub Artificial Intelligence Learning Data Upcycling Project' hosted by the National Intelligence Information Society Agency (NIA), and performed data quality enhancement work such as redesigning existing AI learning data to fit the latest model learning environment and applying CoT-based labeling.
Alchera CEO Young-gyu Hwang stated, “The performance of AI models depends on the quantity and quality of data,” and added, “Through being selected as the lead company for this quality verification, we will contribute to improving the quality of domestic AI training data and, furthermore, take the lead in creating an AI ecosystem with global competitiveness.”
- See more related articles
You must be logged in to post a comment.