11月7日，2026年“妈祖”全球数智创新应用挑战活动成果发布专场在上海举行。在来自近30个国家和地区的参赛队伍中，由我校中外学生组成的参赛团队凭借“面向沙特的MAZU多灾种早期预警智能体算法研发”项目脱颖而出，荣获“全球创新伙伴”（Global Innovation Partners）奖，并获中国气象局颁发的获奖奖杯。
图1成果发布专场现场
本项目紧扣中国气象局MAZU系列早期预警系统“多灾种、预警、零差距、普惠”的核心定位，由我校人工智能团队联合国家卫星气象中心国际用户服务中心共同完成。团队针对沙特阿拉伯热带沙漠气候下突发性暴雨、极端高温、干旱等高频高影响气象灾害，开展面向本地化应用的轻量化、智能化、可嵌入MAZU系统的多灾种早期预警智能体算法研发，推动中国智能预警技术“走出去”。
在技术路线上，团队依托全球大气再分析、地表实况观测、海表温度、卫星云图等多源气象数据资源，创新性地融合大语言模型与气象知识图谱，赋予智能体多模态气象数据解析与逻辑推理能力，显著提升了预警模型对沙特独特地理气候特征的精准性与适应性。项目成果形成了可快速部署、可持续迭代的智能预警技术原型，既满足MAZU系统的模块化嵌入要求，又可面向沙特气象、应急、农业、港航等部门提供实用化预警服务，具有良好的应用落地价值与商业化推广潜力。
图2团队荣获的“全球创新伙伴”奖杯
值得一提的是，本次参赛团队是一支真正的跨国协作队伍：由李瑞瑞担任团队负责人，成员包括来自我校阿塞拜疆籍本科生的Turgut Sofuyev、沙特籍研究生Dari Abdullah Alrwoaily，以及国家卫星气象中心首席专家张艳及田林、李若琳共6人。团队成员跨越不同专业与文化背景，在项目研发中分工协作、优势互补，既展现了我校国际学生的培养质量与科研实力，也体现了学校开放办学的国际化特色，是我校“以赛促学、以赛促创”国际化人才培养模式的生动实践。
此次获奖是我校在国际气象科技合作与创新应用领域取得的重要成果。未来，团队将继续深化智能体算法研发与本地化适配，推动成果在共建“一带一路”国家的落地应用，为服务国家气象技术装备与服务出海战略、助力全球防灾减灾贡献青春力量。
图3 2026年“妈祖”全球数智创新应用挑战活动成果发布专场
Our Chinese Foreign Student Team Wins the “Global Innovation Partners” Award at the 2026 MAZU Global Intelligence Innovation Application Challenge
Project “R&D of MAZU Multi Hazard Early Warning Agent Algorithm for Saudi Arabia” Earns International Recognition
On November 7, the results announcement event of the 2026 MAZU Global Intelligence Innovation Application Challenge was held in Shanghai. Amid competing teams from nearly 30 countries and regions, our university’s Chinese foreign student team stood out with its projectR&D of MAZU Multi Hazard Early Warning Agent Algorithm for Saudi Arabia. The team was honoured with theGlobal Innovation Partners Award and received the trophy presented by the China Meteorological Administration.
Figure 1 Scene of the Results Announcement Event
Aligned with the core positioning of “multi hazard, early warning, zero gap and inclusive services” for the MAZU early warning system series developed by the China Meteorological Administration, this project is jointly completed by our university’s artificial intelligence team and the International User Service Center of the National Satellite Meteorological Center. Targeting high frequency and high impact meteorological hazards under Saudi Arabia’s tropical desert climate, including sudden rainstorms, extreme high temperatures and droughts, the team has developed a lightweight, intelligent multi hazard early warning agent algorithm that can be embedded into the MAZU system for localised application, facilitating the global outreach of China’s intelligent early warning technologies.
In terms of technical approaches, drawing on multi source meteorological datasets such as global atmospheric reanalysis data, in situ surface observations, sea surface temperature records and satellite cloud images, the team innovatively integrates large language models with meteorological knowledge graphs. This empowers the agent with capabilities for multi modal meteorological data interpretation and logical reasoning, greatly improving the accuracy and adaptability of the early warning model for Saudi Arabia’s unique geographical and climatic conditions. The project delivers a rapidly deployable and iterable technical prototype for intelligent early warning. It not only meets the modular embedding requirements of the MAZU system, but also provides practical early warning services for Saudi Arabia’s meteorological, emergency response, agricultural, port shipping and other sectors, boasting sound value for practical implementation and commercial promotion potential.
Figure 2 The “Global Innovation Partners” Trophy Awarded to the Team
Notably, this competing team represents genuine cross border collaboration. Led by Li Ruirui, the six member group includes Turgut Sofuyev, an Azerbaijani undergraduate, Dari Abdullah Alrwoaily, a Saudi postgraduate from our university, alongside chief expert Zhang Yan, Tian Lin and Li Ruolin from the National Satellite Meteorological Center. Boasting diverse majors and cultural backgrounds, team members engaged in well organised work and complementary strengths throughout research and development. Their achievement demonstrates the quality of international student training and research capacity at our university, embodies the international vision of the university’s open door education philosophy, and serves as a vivid practice of our talent training model that “promotes learning and innovation through competitions”.
This award marks a major milestone for our university in international meteorological technology cooperation and innovative application. Going forward, the team will further advance R&D and localised adaptation of the agent algorithm, promote the implementation of project outcomes in Belt and Road partner countries, and contribute youthful strength to China’s strategy for exporting meteorological technologies and services as well as global disaster risk reduction.
Figure 3 Promotional material for the Results Announcement Event of the 2026 MAZU Global Intelligence Innovation Application Challenge
审核:李瑞瑞 姜世超 李齐方