Taiwanese Medical Centers Embrace NVIDIA AI for Advanced Research

Taiwan’s leading medical centers, the National Health Research Institute (NHRI) and Chang Gung Memorial Hospital (CGMH), are adopting NVIDIA’s accelerated computing and generative AI technologies to significantly advance biomedical research and healthcare for patients.

These institutions are leveraging AI for a broad range of applications, from medical imaging and patient care enhancement to clinical workflow optimization and drug discovery research. Dr. Hung-Yi Chiou, director of the Institute of Population Health Sciences (IPHS) at NHRI, stated that AI’s ability to quickly and accurately analyze vast amounts of data is pivotal for developing personalized medicine strategies and early intervention methods.

Transforming Healthcare with AI

Dr. Wen-Jin Cherng of CGMH highlighted the promising future of AI in healthcare, noting that it will enable more accurate diagnoses, better predictive treatment plans, and quicker patient recovery times. AI's integration into healthcare decision-making processes allows for more efficient and cost-effective operations.

The transformative capabilities of the NVIDIA Blackwell platform are being harnessed by these medical centers to integrate advanced AI into their practices, enhancing patient care and streamlining clinical workflows. Dr. Cherng emphasized that the computational power provided by NVIDIA's platform is essential for expanding language model services across all hospitals under CGMH’s network, thereby improving professional support and patient care.

AI-Driven Biomedical Research

NHRI, a leading medical research institution in Taiwan, is incorporating NVIDIA accelerated computing into its IT infrastructure to drive AI-driven healthcare innovations. This collaboration extends to developing large language models tailored for Taiwan’s healthcare needs, addressing unique challenges such as the complexity of traditional Chinese medical records and precise genomic interpretations specific to Taiwan’s population.

NHRI currently utilizes six NVIDIA DGX A100 systems for cloud and data center services, focusing on biomedical model training and genomic analysis. By leveraging NVIDIA’s technology, NHRI is working on projects aimed at predicting the risk of chronic diseases like diabetes and cardiovascular conditions by analyzing genetic and environmental parameters, which was previously unattainable due to computational constraints.

Comprehensive AI Applications at CGMH

CGMH, a major healthcare provider in Taiwan with a network of 10 hospitals and over 11,000 inpatient beds, uses a diverse array of NVIDIA hardware, including H100 and A100 Tensor Core GPUs, for medical imaging development and deployment. The hospital’s AI center, led by Dr. Chang-Fu Kuo, is focused on expanding AI services across its network to support various medical disciplines and diverse patient populations.

CGMH’s initiatives include:

• Clinical Decision Support System: Ensures patient data confidentiality and privacy while assisting clinicians with access to up-to-date data and guidelines.
• Patient Interaction System: Uses robots to answer patient queries about medications and conditions, reducing the burden on medical staff.
• Medical Imaging: Enhances radiology and imaging tasks with AI, one of the most mature AI technologies in CGMH’s system.
• Precision Medicine: Manages large-scale genomic data to transform sequences into readable medical reports for doctors.
• Expansion of AI Services: Extends language model services to all CGMH hospitals, leveraging the computational power of the Blackwell platform.

Other AI applications at CGMH include early detection of colorectal cancer, autoimmune disease screening, and kidney disease prediction.

The adoption of NVIDIA’s accelerated computing by NHRI and CGMH underscores the critical role of AI and advanced computing in enhancing medical research and healthcare delivery. These advancements position Taiwan at the forefront of biomedical innovation, improving patient outcomes and advancing scientific understanding.

For the original source, visit NVIDIA Blog.