• 2025.10.27 (Mon)
  • All articles
  • LOGIN
  • JOIN
Global Economic Times
APEC2025KOREA가이드북
  • Synthesis
  • World
  • Business
  • Industry
  • ICT
  • Distribution Economy
  • Korean Wave News
  • Opinion
  • Arts&Culture
  • Sports
  • People & Life
  • Lee Yeon-sil Column
  • Ko Yong-chul Column
  • Photo News
  • New Book Guide
  • Cherry Garden Story
MENU
 
Home > Synthesis

SNUH Unveils Korea's First Medical Large Language Model, Surpassing Doctor Accuracy in Medical Exams

Desk / Updated : 2025-03-22 21:51:36
  • -
  • +
  • Print

Seoul, South Korea – Seoul National University Hospital (SNUH) has achieved a groundbreaking milestone by developing the first 'Korean Medical Large Language Model (LLM)' in the nation. This innovative model, announced on March 21, 2024, is set to revolutionize the Korean medical landscape by enhancing treatment efficiency and patient safety.

The LLM was trained using an extensive dataset of 38 million clinical texts, including Electronic Medical Records (EMR), Picture Archiving and Communication System (PACS) data, digital pathology, and genomic information. This comprehensive approach ensures the model's proficiency in processing medical data specific to the Korean healthcare system.

"This is a significant technological advancement that will not only streamline medical processes but also elevate the quality of patient care," stated Kim Young-tae, Director of SNUH.   

One of the model's notable achievements is its performance on the Korean Medical Licensing Examination, where it achieved an accuracy rate of 86.2%, surpassing the average accuracy of practicing doctors (79.7%). This feat marks the first time an open-source model has outperformed physicians in this critical assessment.

SNUH plans to integrate this LLM into various hospital operations, including research and administrative tasks. Future developments include expanding the model's capabilities to multimodal AI, incorporating medical imaging and bio-signal data. Projects such as 'HIS.AI' for automated medical record summarization, 'CLAIM.AI' for efficient insurance claims, and 'RESEARCH.AI' for personalized research paper curation are underway.

"By automating routine tasks and providing tailored information, this model will significantly reduce burnout among medical staff, allowing them to focus more on patient interaction," explained Lee Hyung-chul, Vice Director of the Healthcare AI Research Institute at SNUH.

The hospital aims to further refine the LLM's performance for practical use in clinical settings, ensuring that it remains at the forefront of medical technology. This development underscores SNUH's commitment to leveraging cutting-edge solutions for the betterment of healthcare in Korea.

[Copyright (c) Global Economic Times. All Rights Reserved.]

Desk
Desk

Popular articles

  • Korean Gold Rush Overheats as 'Kimchi Premium' Hits Dangerous Levels

  • KOSPI Sensitive to Global Headwinds Ahead of Holiday Break

  • Cocoa Futures Plunge 50% from Peak, Signaling Potential Chocolate Price Relief

I like it
Share
  • Facebook
  • X
  • Kakaotalk
  • LINE
  • BAND
  • NAVER
  • https://globaleconomictimes.kr/article/1065617429970904 Copy URL copied.
Comments >

Comments 0

Weekly Hot Issue

  • South Korea Appoints Special Prosecutor to Probe Alleged Corruption and External Pressure in High-Profile Cases
  • Teen Prodigy Kim Hyeon-seo Makes History at Paganini Competition
  • Badminton Queen An Se-young Conquers French Open for Ninth Title of the Year
  • Suspects Arrested in Audacious Louvre Jewel Heist
  • Former KBS Announcer Kim Jae-won Reveals Truth Behind 'Morning Yard' Exit and Voluntary Retirement
  • K-Pop Group DreamNote Disbands After Seven Years, Agency Announces Contract Termination

Most Viewed

1
Early Winter Chill Grips South Korea as Seoraksan Sees First Snow
2
Gyeongju International Marathon Elevated to 'Elite Label' Status, Welcomes Record 15,000 Runners  
3
South Korean Chip Titans Clash Over Next-Gen HBM4 Memory
4
Deadly Clan Clashes Erupt in Gaza as Israeli Forces Withdraw
5
Global Chip War Intensifies: Micron Woos Korean Engineers with Lucrative Offers, Up to 200 Million KRW Salary
광고문의
임시1
임시3
임시2

Hot Issue

Trump Acknowledges North Korea as a 'Sort of Nuclear Power,' Puts Ball in Kim's Court

ASEAN Summit Opens in Kuala Lumpur, Addressing Trade Tensions and Transnational Crime

Kimcheon Gimbap Festival Becomes a 'Great Rush' as Crowds Swell to 150,000

Rival Parties Pass 70-Plus Public Welfare Bills, Including 'Emergency Room Loitering Prevention Act'

Let’s recycle the old blankets in Jeju Island’s closet instead of incinerating them.

Global Economic Times
korocamia@naver.com
CEO : LEE YEON-SIL
Publisher : KO YONG-CHUL
Registration number : Seoul, A55681
Registration Date : 2024-10-24
Youth Protection Manager: KO YONG-CHUL
Singapore Headquarters
5A Woodlands Road #11-34 The Tennery. S'677728
Korean Branch
Phone : +82(0)10 4724 5264
#304, 6 Nonhyeon-ro 111-gil, Gangnam-gu, Seoul
Copyright © Global Economic Times All Rights Reserved
  • 에이펙2025
  • APEC2025가이드북TV
  • 세종시
Search
Category
  • All articles
  • Synthesis
  • World
  • Business
  • Industry
  • ICT
  • Distribution Economy
  • Korean Wave News
  • Opinion
  • Arts&Culture
  • Sports
  • People & Life
  • Lee Yeon-sil Column
  • Ko Yong-chul Column
  • Photo News
  • New Book Guide
  • Cherry Garden Story
  • Multicultural News
  • Jobs & Workers
  • APEC 2025 KOREA GUIDE