Lunit hosts medical AI hackathon to expand foundation model ecosystem

Lunit, a medical artificial intelligence company, successfully hosted a medical-specialized foundation model hackathon at its headquarters in Gangnam, Seoul. Sponsored by the Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA), the event aimed to expand the ecosystem for medical AI and increase the utility of Lunit’s ‘L2-preview’ foundation model.

Approximately 70 participants, including 15 teams composed of AI developers, clinicians, engineers, and graduate students, competed in tasks involving medical Q&A systems and healthcare chatbot development. The competition included a benchmark task where teams were evaluated on the quality of their answers to medical and health-related queries. OpenAI supported the event by providing ChatGPT Pro subscriptions and API credits.

The AIM team received the Benchmark Award for demonstrating performance comparable to global frontier models in medical science benchmarks. The Frontier Award was given to the ‘Nalyeomhan Wegovy’ team for their high potential for clinical application using specialized medical data and knowledge base search systems.

Lunit plans to use the hackathon as a platform to discover talented individuals and excellent medical AI services. The company intends to provide clinical verification opportunities and technical support through its hospital networks to connect successful results with real-world hospital environments and follow-up R&D.