Seoul, South Korea – EuclidSoft, a leading AI data solutions provider, has successfully completed a data construction project as part of the "Ultra-Large AI Diffusion Ecosystem Creation Project," solidifying its position as a pioneer in AI learning data.

In collaboration with Saramgwasup Media Group and the MLP Research Team at Seoul National University of Science and Technology, EuclidSoft formed a consortium to construct visual data-based question-and-answer data. This endeavor has culminated in the release of "Blossom," a Korean-specialized visual-language model.

The newly constructed dataset is designed to enhance AI's contextual understanding and reasoning abilities by focusing on structured documents containing visual elements such as images, tables, and graphs. This type of data has the potential to revolutionize various applications, including document generation, question-answering systems, and search services. Upon completion, the dataset will be made publicly available through AI Hub.

Over the past four years, EuclidSoft has built a robust reputation in data processing and refinement by constructing more than 50 million AI learning data points. The company's proprietary crowdsourcing platform, LabelOn, has been instrumental in streamlining data management and processing.

"We are excited to leverage our expertise to provide high-quality data and unlock new possibilities within the ultra-large AI ecosystem," said Ju-han Park, CEO of EuclidSoft. "Blossom represents a significant milestone in our journey to advance AI research and development."