Jebahyeop Holds Briefing Session on 'Federated Learning-Based New Drug Development Project'

The Korea Pharmaceutical and Bio-Pharma Manufacturers Association K-Melody Project Group announced on the 31st that it will hold a "Federated Learning-Based New Drug Development Acceleration Project Briefing Session" at the Pharmaceutical and Bio-Pharma Association auditorium on the 5th of next month.

[Photo by Korea Pharmaceutical and Bio-Pharma Manufacturers Association]

Federated learning refers to a technology that trains artificial intelligence (AI) at individual institutions without gathering the data held by each institution in one place. It is explained that this method poses almost no risk of information leakage, enabling both the ‘protection’ and ‘utilization’ of sensitive information simultaneously.

This federated learning-based new drug development acceleration project is hosted by the Ministry of Health and Welfare, the Ministry of Science and ICT, the Korea Health Industry Development Institute, and the National Research Foundation of Korea, and is organized by the Pharmaceutical and Bio-Pharma Association and the K-Melody Project Group. The goal is to develop a federated ADMET model (FAM), an ADMET prediction model based on federated learning. ADMET stands for Absorption, Distribution, Metabolism, Excretion, and Toxicity, which are the most critical factors for clinical trial success. The U.S. National Institutes of Health estimates that ADMET analysis accounts for about 22% of the research and development (R&D) costs in new drug development. In particular, in Korea, many cases only reach Phase 1 clinical trials due to technology exports, meaning that ADMET accounts for most of the clinical costs.

The project is broadly divided into ▲platform construction ▲data supply and utilization ▲AI model development. The detailed tasks include ▲platform construction and development (1 task) to build a federated learning-based FAM operation platform ▲data supply and utilization (20 tasks) for pharmaceutical companies, hospitals, and research institutes using FAM ▲AI model development (15 tasks) to develop FAM solutions and application models.

This briefing session is organized to explain the detailed contents of each public project to interested researchers, research institutions, medical institutions, and companies ahead of the full-scale project launch in July, and to gather diverse opinions. The session will include presentations on ▲federated learning platform construction and development ▲establishment of a cooperative system for new drug development using data ▲support for activation of federated learning platform utilization, along with project introduction, strategies, schedules, and participation guidelines, followed by a Q&A session.

Pre-registration for the briefing session can be done through the Pharmaceutical and Bio-Pharma Association website. Additionally, the briefing materials will be posted on the association’s website after the session concludes.