The reporter learned from the Army Medical University Xinqiao Hospital (hereinafter referred to as "Xinqiao Hospital") that after more than half a year of collaborative research and development, the hospital, together with the School of Computer Science at Chongqing University, has successfully developed an artificial intelligence specialized disease model for pulmonary infectious diseases. This is the first large-scale model of pulmonary infection based on diagnosis and treatment strategies generated in China. At present, the model has been deployed to the respiratory ward medical workstation of Xinqiao Hospital and entered the clinical trial stage. Xu Zhi, director of the Respiratory and Critical Care Medicine Center at Xinqiao Hospital, introduced that pulmonary infection, as a common but highly challenging disease, adds hundreds of millions of new cases and causes over a million deaths worldwide every year. Its complex etiology and diverse manifestations pose great difficulties for clinical diagnosis and treatment. For special populations such as the elderly, pregnant women, and patients with complex conditions, doctors often face significant challenges when formulating treatment plans. This model focuses on the pain points of pulmonary infection diagnosis and treatment, integrating 12893 high-quality clinical cases with authoritative medical guidelines from both domestic and foreign sources containing over 430000 Q&A questions, providing intelligent solutions for the diagnosis and treatment of pulmonary infection diseases. In the process of technical implementation, the team abandoned traditional methods and combined multiple training strategies to enable the model to understand medical problems more accurately and provide personalized diagnosis and treatment recommendations. Chen Xuetao, director of the Information Department of Xinqiao Hospital, introduced that the team has established a three-dimensional evaluation system of "text similarity+medical key point coverage", which has performed well in multiple tests. Compared with advanced general models at home and abroad, this model has significantly improved the accuracy of diagnosing pulmonary infectious diseases and the rationality of recommending treatment plans, with an average performance improvement of about 47%. Especially in terms of key medical information coverage, the text generated by this model is more in line with medical standards, and can more comprehensively capture pathogen types, generate core content such as medication plans and testing recommendations, and better assist doctors in clinical decision-making. Chen Xuetao stated that Xinqiao Hospital will continue to track the performance of the model in a real medical environment, establish a closed-loop feedback mechanism, and continuously optimize the performance of the model. It is reported that Xinqiao Hospital will continue to develop more specialized disease artificial intelligence models and gradually build a specialized intelligent assistance system group covering multiple diseases according to its three-year action plan for artificial intelligence and smart hospital construction. (New Society)