"AI-Driven New Drug Development Faces Investment and Talent Shortages... Collaboration Through Federated Learning Needed"

"Shortage of Talent in New Drug Development and AI Fields... Also Financial Difficulties"
Sharing AI Models Using Federated Learning
Data Should Be Utilized Without Distinction Between Private and Public Sectors

Concerns have been raised that domestic companies developing new drugs using artificial intelligence (AI) are lagging behind neighboring countries in attracting investment and securing talent. As a measure to accelerate AI-driven new drug development, the use of federated learning technology was proposed.

Kim Woo-yeon, Head of the AI New Drug Development Support Center at the Korea Pharmaceutical and Bio-Pharma Manufacturers Association, is giving a lecture at the "Pharmaceutical Bio AI Innovation Forum" held on the 19th at Lotte Hotel, Sogong-dong, Seoul.
[Photo by Korea Pharmaceutical and Bio-Pharma Manufacturers Association]

On the 19th, Kim Woo-yeon, head of the AI New Drug Development Support Center at the Korea Pharmaceutical and Bio-Pharma Manufacturers Association, stated at the 'Pharmaceutical Bio AI Innovation Forum' held at Lotte Hotel in Sogong-dong, Seoul, "To revitalize the new drug development ecosystem, many success cases through step-by-step collaboration must be produced."

According to Director Kim, the United States accounts for more than half of the global AI-driven new drug development market. The North American AI-driven new drug development market is expected to reach $400 million (approximately 530 billion KRW) this year, with global big pharma companies such as Pfizer already widely utilizing AI in new drug development and clinical trial design through technology investments.

In contrast, domestic AI new drug development companies fall short of Chinese companies in terms of investment attraction and human resources. The level of AI new drug development technology in China is evaluated to be similar to that of domestic companies. Director Kim cited domestic company Standigm and China's XtalPi as examples. Standigm, established in 2015, received 87.8 billion KRW in investment up to the pre-IPO (initial public offering) stage, whereas XtalPi, launched in 2014, attracted 533.8 billion KRW in investment. In terms of specialized personnel, Standigm has only 54 employees, while XtalPi has more than 700.

The problems Director Kim points out in the domestic AI new drug development market are also the lack of talent and technology investment. There is a shortage of personnel who understand both new drug development and AI, making it difficult to produce tangible results. The lack of technology investment funds due to difficulties in investment screening based solely on AI technology also hindered progress. Other issues raised include ▲difficulty in accessing data ▲lack of visible outcomes from joint research.

Director Kim emphasized that short-term success cases are necessary to speed up AI-driven new drug development. This means supplying AI technology to pharmaceutical companies with AI demand to accumulate success cases in a short period. He also mentioned the need to establish a technology roadmap to activate joint research and private investment.

He particularly stressed the need for data utilization and cooperation without distinction between private and public sectors through federated learning (FL). Federated learning enables AI models to learn from private and public data stored in a distributed manner across institutions. Instead of directly sharing sensitive medical information, the AI models trained on distributed data are shared for collective use. Director Kim said, "It is necessary to develop technology that can overcome privacy and research security issues and link and utilize public and private data without data leakage," adding, "An AI new drug development platform based on federated learning should be established to create an effective and efficient new drug development research environment."

The Pharmaceutical Bio Association is also promoting the establishment of the 'K-MELLODDY' project to accelerate federated learning-based new drug development. K-MELLODDY benchmarks the European pharmaceutical companies' project 'MELLODDY,' which utilized federated learning-based AI in new drug development. It is a platform that allows multiple institutions to use and collaborate on data without physical data sharing. The association explained that they aim to develop a rapid and efficient AI new drug development methodology based on K-MELLODDY.

Earlier, at the opening speech, Noh Yeon-hong, chairman of the Korea Pharmaceutical and Bio-Pharma Manufacturers Association, said, "The EU-MELLODDY project confirmed that federated learning-based AI models have performance improvement effects compared to standalone AI models," adding, "The pharmaceutical and bio industries will commercialize federated learning technology based on the Korean-style MELLODDY project and build a new drug development data cooperation system to greatly enhance the productivity of new drug development."

Noh Yeon-hong, Chairman of the Korea Pharmaceutical and Bio-Pharma Manufacturers Association, is delivering the opening address at the "Pharmaceutical Bio AI Innovation Forum" held on the 19th at Lotte Hotel, Sogong-dong, Seoul.
[Photo by Korea Pharmaceutical and Bio-Pharma Manufacturers Association]

The event was organized to explore global technology trends related to AI new drug development and strategies for successful new drug development. At the forum attended by academic experts, pharmaceutical companies, and AI company representatives, presentations were made on topics including ▲AI new drug development and big data utilization strategies ▲strategies to activate data sharing and the use of federated learning technology ▲AI-driven new drug development, pharmaceutical companies' innovation and strategic response ▲the importance of collaboration and investment in AI new drug development.

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