FrailMIND is a Korean national R&D platform that assesses frailty risk with a 9-modality foundation model and delivers personalized interventions across exercise, nutrition, and medication — purpose-built for the world's fastest-aging society.
FrailMIND addresses this gap through medical AI, multimodal learning, and preventive care combined. A 14B foundation model fuses nine data modalities — EMR, imaging, voice, EMG, lifelog, and more — to predict personal frailty trajectories. Lightweight 2B service models then deliver tailored interventions across hospital, community, and home settings.
Funded by ARPA-DEF (2025–2029) and led by Seoul National University Bundang Hospital, the 13-institution consortium will validate the platform across six regions. The program targets HIMSS CCMM certification, national policy alignment, and commercial deployment.
Beyond individual AI models — build a sustainable frailty-care AI platform and scale it nationwide.
Weave individual models and services into a unified PHM-based elderly health management platform. Data linkage → risk stratification → personalized intervention → monitoring.
Validate in 6 pilot regions (metro · mid-size · rural), then scale across primary·secondary·tertiary hospitals + home-based care networks.
KR-CDI/FHIR integration, HIMSS CCMM & TTA/AIIA certification to establish a national standard for frailty care, linked to community integrated care policy.
15 AI services spanning frailty assessment, exercise, nutrition, and medication management.
Hospitals, universities, and industry partners unified under SNUBH leadership, each contributing domain expertise.
Technical architecture, performance metrics, roadmap, and internal documents are available in the Consortium Portal. Sign in with an authorized email.
Sign in to PortalAcademic collaboration · test-bed participation · patient data sharing · policy alignment — all inquiries welcome.