Trillion Labs Embarks on Development of Specialized AI Model for Medical Science

Participation in the Ministry of Science and ICT’s ‘AI Specialized Foundation Model’ Project

AI startup Trillion Labs has been selected as a development institution for the Ministry of Science and ICT’s ‘AI Specialized Foundation Model Project.’ This project aims to build the world’s first AI model that integrates and learns from full-cycle medical science data. The consortium is led by Lunit, with Trillion Labs and other major domestic AI and biotech companies participating.

The project is titled ‘Development of a Multiscale Specialized Foundation Model for Full-Cycle Medical Science Innovation, from Molecules to Populations.’ This is an unprecedented attempt to integrate vast amounts of life science data-including molecular, protein, omics, pharmaceutical, and clinical data-into a single model. The goal is to create an integrated AI platform that connects all areas of medical science, going beyond simple healthcare-specialized AI.

Trillion Labs CI.

Within the consortium, Trillion Labs is responsible for designing the core engine and learning architecture of the model. Leveraging its experience in developing language models with 7 billion to 70 billion parameters and its proprietary ‘Cross-lingual Document Attention (XLDA)’ technology, the company plans to develop a system capable of efficiently learning from large-scale medical data and making highly precise inferences.

Additionally, through its ‘AI Collaborative Research System (Co-Scientist or Bio-Medical Co-Scientist),’ Trillion Labs will build an intelligent research partner system that enables researchers, physicians, and pharmaceutical companies to formulate and test hypotheses with AI in real time. The company will also implement a ‘Chain of Evidence’ technology that traces evidence by linking guidelines, clinical trials, drug labels, and real-world data, thereby enhancing the reliability and reproducibility of research outcomes.

Trillion Labs will also collaborate with domestic NPU company Rebellions to build an on-premise medical AI infrastructure. This approach will secure sovereignty over medical data while providing a high-performance computing environment that does not rely on external cloud services. It is expected to serve as a foundation for simultaneously addressing medical data security and large-scale computation challenges.

In addition to Trillion Labs, the consortium includes seven companies such as Igen Science, SK Biopharmaceuticals, Kakao Healthcare, Standigm, and Dcircle, as well as research teams from KAIST and Seoul National University, and nine medical institutions including the National Health Insurance Service Ilsan Hospital. The total project budget is 18.2 billion won, with the government supporting GPU infrastructure resources worth 17.4 billion won. The research will proceed in two phases through September next year.

Shin Jae-min, CEO of Trillion Labs, stated, “This project is a turning point for Korea to secure global competitiveness in the fields of medical AI and science,” adding, “We will usher in a new era of medical science where AI collaborates with humans in research and clinical innovation.”

This content was produced with the assistance of AI translation services.