SK Biopharmaceuticals builds massive AI dataset to speed up drug discovery

SK Biopharmaceuticals, a Korean pharmaceutical company, has constructed an extensive artificial intelligence (AI) dataset comprising over 7 million chemical entries to expedite the process of identifying potential new treatments. This initiative aligns with the Ministry of Science and ICT’s initiative to develop AI foundation models tailored to specific industries.

Foundation models are large AI systems trained on vast data volumes that can be repurposed for various applications. SK Biopharmaceuticals is part of a consortium of medical and health care entities, including Lunit, Trillion Labs, Kakao Healthcare, and other organizations, universities, and hospitals. The consortium created a foundational model tailored to the medical sector, subsequently developing specialized versions for three key areas: drug development, encompassing chemical compounds and proteins; omics, which involves the extensive study of genes and other biological molecules; and clinical research.

SK Biopharmaceuticals assembled the drug discovery dataset based on the actual requirements of researchers in laboratory settings. The dataset encompasses 66 distinct drug development tasks, such as predicting various outcomes.