Amazon Web Services (AWS) announced on the 5th that it will provide cloud infrastructure for the artificial intelligence platform of the federation learning-based new drug development acceleration project 'K-MELLODDY (Machine Learning Ledger for Drug Discovery).'
K-MELLODDY is a large-scale research and development (R&D) initiative involving 38 institutions, including 11 universities, 9 major pharmaceutical companies, 8 AI developers, 7 research institutions, and 3 hospitals in Korea.
According to AWS, K-MELLODDY aims to develop learning-based AI models for predicting the properties of new drug candidates to shorten the preclinical period before human trials. This personal data protection platform can also be applied to other industries such as manufacturing, education, government, and finance.
K-MELLODDY supports domestic hospitals and pharmaceutical companies to independently train AI models with their local data. Additionally, it is shared on a central platform based on AWS, excluding sensitive data. The initiative has adopted the federated learning approach utilized by European global pharmaceutical companies for medical AI development.