气象人工智能通用算法库的设计与实现  

Design and Implementation of General Algorithm Library for Meteorological Artificial Intelligence

作　　者：雷鸣 宫志宏[2] 郭阳 勾志竟 姜罕盛 LEI Ming;GONG Zhihong;GUO Yang;GOU Zhijing;JIANG Hansheng(Tianjin Meteorological Information Center,Tianjin 300074,China;Tianjin Climate Center,Tianjin 300074,China)

机构地区：[1]天津市气象信息中心,天津300074 [2]天津市气候中心,天津300074

出　　处：《微型电脑应用》2024年第5期1-4,共4页Microcomputer Applications

基　　金：国家自然科学基金(41575156);中央级公益性科研院所基本科研业务费专项(IUMKY201605);天津市气象局科研项目(201914ybxm12)。

摘　　要：为了更好地利用AI技术的优势,降低开发周期和难度,针对气象行业的特点,提出一种适合气象行业应用的人工智能通用算法库设计方案。该方案基于标准规范和模块划分,通过系统设计,形成有效的算法资源池,在大大减少开发工作量的同时,有效降低开发门槛,构建积木化算法开发范式。同时,因为系统采用MUSIC和FTP协议进行数据交换,能够方便与全国天擎系统进行对接,形成“数据、算法、产品和展示”的一条龙服务,为AI在气象行业的落地和融合提供了有益的探索。In order to make better use of the advantages of AI technology and reduce the development cycle and difficulty,according to the characteristics of the meteorological industry,a general algorithm library design scheme of artificial intelligence is proposed.The scheme is based on standard specifications and module division.Through system design,an effective algorithm resource pool is formed.While greatly reducing the development workload,it can effectively reduce the development threshold and build a modular algorithm development paradigm.At the same time,because the system uses MUSIC and FTP protocols for data exchange,it can easily connect with the national Tianqing system,forming a one-stop service of“data,algorithms,products and display”,which provides a beneficial exploration for the landing and integration of AI in the meteorological industry.

关 键 词：深度学习 AI算法库 人工智能 MUSIC ModelArts 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]