近年来，孟买极端降雨频率与强度显著上升，而传统全球预报系统因分辨率不足难以捕捉局地天气特征。为此，印度理工学院孟买分校与马里兰大学合 作，开发了基于卷积神经网络与迁移学习的预测模型，实现了对极端降雨事件的提前预报。 每年 6 月至 9 月，印度孟买进入季风季节。近年来，孟买极端降雨事件频发，平均降雨量较 2019 年前增加了近 40%。这座拥有 1800 万人口的沿海城市 常因暴雨陷入混乱：天气预警滞后往往导致停工停学甚至严重洪灾，季风带来的重大灾害对更精准的本地化天气预测提出了迫切需求。 然而，在热带季风气候下，标准全球天气模型约 25 平方公里的分辨率难以捕捉局部天气系统的细微差异，地形的复杂性也加剧了洪水的空间不确定性。 因此，对季风预测此前仅限于宏观趋势。 为填补城市洪水风险预判对缺口，印度理工学院孟买分校与马里兰大学研究团队合作开发了基于卷积神经网络（CNN）和迁移学习（CNN-TL）的超本地 预测模型，实现了提前数天预测大部分强降雨事件。根据 Science 最新报道，目前，孟买方面已经考虑将其纳入官方预警体系，标志着南亚城市洪水预报 进入了超本地化的新阶段。 相关研究成果以「Hyperlocal ...

Core Viewpoint - The "2025 Global Smart Education Conference" emphasizes the integration of artificial intelligence in education, highlighting the need for educational transformation in the digital age [1][2][3]. Group 1: Educational Transformation - The conference focuses on the theme of "Human-Machine Collaboration in Creating a New Educational Ecosystem," indicating a shift towards integrating AI in educational practices [1]. - The Chinese education system is undergoing profound changes due to the deep integration of AI, which is seen as a strategic matter affecting the success of building a strong education nation [2]. - The mission and direction of education remain unchanged despite the evolving educational forms and tools brought about by AI [3]. Group 2: AI's Role in Education - AI technology is enhancing the educational experience for special needs students, improving their engagement and confidence in learning environments [2]. - The relationship between education and AI is reciprocal, where educational scenarios contribute to the iterative development of AI, leading to improved data quality and teaching standards [3]. - The conference aims to foster international collaboration in developing safe and controllable AI models for education, ensuring data sovereignty and cultural value transmission [3].