From August 15 to August 17, 2025, the 31st China International Medical Equipment Fair and Technical Exchange Conference (China-Hospeq) was held at the Beijing National Convention Center. As one of the key events of the exhibition, on August 16, the "Walking with AI: Health Silk Road AI Empowerment Seminar" was successfully held, hosted by the International Exchange and Cooperation Center of the National Health Commission and co-organized by Beijing Boya Quanjian Intelligent Computing Technology Co., Ltd. (referred to as "Boya Quanjian"), a subsidiary of TEDA BIOMEDICAL Engineering Co., Ltd. (stock code: 08189.HK, hereinafter referred to as "TEDA BIOMEDICAL"), along with Peking University Science Park. Representatives from supporting organizations including the Ningxia Hui Autonomous Region Health Commission, China National Pharmaceutical Industry Information Center, and CITIC Medical Health Industry Group Co., Ltd. gathered at the venue to explore cutting-edge applications and development of AI technology in the medical field.
**Focusing on Resource Optimization Configuration and Anchoring AI + Medical Empowerment Strategic Direction**
The seminar, themed "Walking with AI: Health Silk Road AI Empowerment," aimed to build a new "Health Silk Road" using AI technology, breaking down medical resource barriers and promoting efficient coordination of medical resources across provinces, urban-rural areas, and internationally. This initiative seeks to construct a broader and more efficient medical service network, allowing high-quality medical resources to transcend geographical limitations and benefit a wider population.
The launch ceremony of the "China-Africa Hospital Alliance 'Health Silk Road AI Development' Joint Working Group," the co-construction ceremony of the "Global Children's 'Two Diseases' (Epilepsy and Congenital Heart Disease) Early Diagnosis and Treatment AI Innovation Ecosystem," and the signing ceremony for the "Xihe-1" implementation in Ningxia to jointly build a digital health management platform were held simultaneously. These multiple initiatives injected practical direction into the construction of the "Health Silk Road," providing implementation pathways for breaking medical resource geographical barriers and improving diagnostic accuracy in urban-rural and international regions through technological collaboration and resource integration, contributing Chinese wisdom and public products from AI healthcare to the construction of a global health community.
Meanwhile, the bilingual (Chinese-English) "Xihe-1" medical large model, jointly developed by Peking University Third Hospital, Peking University School of Artificial Intelligence, and Boya Quanjian, became a focal point. The model took nearly 10 years to develop, based on research from 1 million real medical cases, forming 100 billion-level parameters. Relying on dual engines of natural language interaction and traceable knowledge reasoning, it supports multi-scenario applications including disease prediction, diagnosis and personalized treatment, imaging analysis, drug development, clinical trials, and medical education, achieving 98% medical knowledge coverage and over 90% medical knowledge accuracy.
**Building Massive Clinical Data Foundation with Precision as Core to Create Differentiated Competitive Advantages**
"Data is the soul of AI models, and the soul of 'Xihe-1' is rooted in the most rigorous clinical medical data," emphasized Dr. Li Lin, Chairman of Beijing Boya Quanjian Intelligent Computing Technology Co., Ltd., in an interview. In the increasingly crowded AI healthcare track, "Xihe-1" establishes differentiated barriers from the source—using hospital medical records as core training data, replacing the "second-hand information" from the internet that some peers rely on with rigorously screened and annotated clinical cases, imaging, and test results as "gold standards."
Li Lin explained with an example: "Internet data may include health information and user Q&As, but this content lacks clinical validation and may even contain errors. Models trained with internet data are like 'learning to see patients through textbooks,' while 'Xihe-1' is 'following doctors on ward rounds and consultations' on the clinical frontline."
According to reports, the training data for "Xihe-1" all comes from clinical frontlines of 18 medical institutions, covering demographic characteristics, time-series physical and chemical indicators, radiomics, and other multi-modal information. Each piece of data has undergone ethical review and anonymization processing, with the annotation process involving clinical doctors throughout. This data advantage makes "Xihe-1" have an extremely low hallucination (false output) rate, with accuracy reaching over 90%, avoiding fatal errors such as "misdiagnosing common cough as serious lung disease," and gaining advantages in compliance, clearing data compliance obstacles for subsequent commercial implementation.
In the "AI and Future Medical Landscape" session of the conference, Professor Tang Yida, Vice Director of Peking University Health Science Center, Vice President of Peking University Third Hospital, and Chief Scientist of the "Xihe-1" large model, shared practical applications of multimodal large language models in pre-consultation triage in the cardiovascular field. He mentioned that current chest pain and myocardial infarction diseases have pain points including misdiagnosis and missed diagnosis rates exceeding 25%, insufficient grassroots diagnostic capabilities, and delayed guideline implementation. In traditional diagnosis and treatment, manual calculation error rates are high, examinations take over 30 minutes, leading to blind drug selection, and treatment delays at the grassroots level can increase mortality risk by 35%. "Xihe-1" addresses these problems by constructing a R&D system centered on clinical urgent needs.
In the keynote speech "AI and the Future of Medical Research," Li Lin stated: "Boya Quanjian is committed to building an AI + medical ecosystem covering 'data-model-clinical' aspects, spanning multiple fields including omics research, medical imaging, and drug development, promoting rapid transformation of research results." She also mentioned that AI intelligent terminal hardware products such as smart wheelchairs, nutrition management equipment, and guide dog robots will be key expansion directions in the future, better meeting people's health service needs.
In the subsequent roundtable dialogue session, Li Lin engaged in in-depth discussions with representatives including Professor Yao Lan, Dean of the School of Medicine and Health Management at Tongji Medical College, Huazhong University of Science and Technology, and Li Xiuli, Deputy Director of the Haidian District Health Commission in Beijing, around the core topic of "AI Empowering Global Medical Industry." The collision of multiple perspectives sparked wisdom and provided diverse insights for industry development.
**Deeply Cultivating AI + Medical Field, TEDA BIOMEDICAL Empowers Global Medical Digital Intelligence Development**
As a company actively engaged in domestic healthcare industry innovation, TEDA BIOMEDICAL has always been guided by promoting efficient coordination of medical resources and deeply cultivating the AI + medical track. Leveraging technical accumulation in the biotechnology field, the company has formed distinctive layouts across multiple dimensions:
At the clinical application level, focusing on AI medical foundation large model development, promoting specialized disease treatment, medical management efficiency improvement, and hospital research level enhancement through medical data platform construction;
At the resource coordination level, connecting domestic medical institutions and research institutes at all levels, participating in the structured, standardized, and marketized construction of medical big data, forming strong innovation capabilities of "technology R&D - scenario implementation - industry empowerment," and deeply cooperating with Hongxin Electronics (300657, SZ) subsidiary Suihong Huachuang in edge medical diagnosis and treatment integrated machines;
At the service application level, actively exploring "AI + medical" product service models, helping high-quality diagnostic and treatment resources extend to counties and townships, providing more convenient and inclusive medical services for grassroots patients.
On August 11, 2025, TEDA BIOMEDICAL disclosed an announcement showing that its controlling subsidiary Shenzhen Medical Data Precision Technology Co., Ltd. signed a cooperation framework agreement with Shenzhen Institute of Computing Sciences. Both parties will integrate resource advantages and conduct in-depth cooperation in database construction, data quality improvement, and data analysis applications, building a solid data foundation for the iterative optimization of AI medical large models.
"We expect to serve over 100 million patients within three years," Li Lin stated. "More importantly, we will establish a global medical AI collaboration network, working with experts from various countries to improve models together, truly achieving global sharing of medical wisdom."
The convening of this seminar provides clearer direction for the industry's development in the AI + medical field. In the future, TEDA BIOMEDICAL will continue to deepen exploration and practice in the AI + medical field, strengthen cooperation with medical institutions, technology companies, and industry associations, gather cutting-edge digital intelligence forces, actively participate in "Health Silk Road" ecosystem construction, and contribute to the optimization of domestic and international medical resource allocation and the improvement of national health levels.