摘要
深度学习增强了人工智能(AI)算法的通用性,近年来催生了AI产业的快速发展,但实践表明AI技术和算法在产业领域的落地应用依然面临极大困难;企业用上并用好AI、学术界和产业界协同以解决算法落地困难等问题,受到广泛关注。本文着眼我国AI产业的健康可持续发展,从企业AI应用落地的实际案例出发,梳理业界现状、剖析发展挑战、探讨根本原因、提出应对策略。企业AI落地应用的复杂性表现在业务需求、数据、算法、基础设施、配套方案等多个维度,应用成熟度取决于数据的准备程度及治理水平。在国家宏观层面,有必要构建更友好的AI产业生态环境,促进AI全产业链协同发展;以更有力的具体举措支持AI产业的技术研发,特别是全栈AI、AI基础平台及工具体系、AI根技术等,提高我国AI核心技术的自主可控能力;鼓励企业积极实施数字化转型,采用AI技术进行智能化升级,形成AI产业技术研发、企业AI落地创新的强耦合及双向循环。
Deep learning has enhanced the versatility of artificial intelligence(AI)algorithms.In the last decade,the AI industry has been spawned and developing rapidly.However,practice shows that the application of AI technology and algorithms in the industrial field faces huge challenges.Approaches need to be explored for enterprises to properly use AI and for the academia and industry to effectively collaborate to facilitate algorithm application.The study focuses on the sustainable development of China’s AI industry,and presents several practical cases of AI application,through which we analyze the current status,challenges,and their root causes pertaining to industrial application of AI and propose corresponding suggestions.The complexity of AI application for enterprises involves multiple dimensions,including business requirements,data,algorithms,infrastructure,and supporting systems.The maturity of AI application depends on the degree of data preparation and the level of data governance.At the national level,a friendly ecology for AI application should be built to promote the coordinated development of the entire industry chain,and specific measures should be taken to support the research and development of AI technologies that focus on full-stack AI,AI basic platform and tool system,and AI root technology,thus to improve the independence of China’s AI core technologies.Moreover,enterprises should be encouraged to actively participate in digital transformation and intelligent upgrading using AI technologies,thereby forming a strong coupling and a two-way cycle between research and application of AI technologies.
作者
徐文伟
肖立志
刘合
XuWenwei;Xiao Lizhi;Liu He(Huawei Technologies Co.,Ltd.,Shenzhen 518129,Guangdong,China;College of Artificial Intelligence,China University of Petroleum,Beijing 102249,China;Research Institute of Petroleum Exploration and Development,Beijing 100083,China)
出处
《中国工程科学》
CSCD
北大核心
2022年第6期173-183,共11页
Strategic Study of CAE
基金
中国工程院咨询项目“中国工程科技未来20年(2021—2040年)创新发展机制与策略研究”(L2124001)
国家自然科学基金项目(72088101)。
关键词
人工智能
企业场景
智能解决方案
落地应用
全栈AI
AI根技术
artificial intelligence(AI)
enterprise scenarios
intelligent solutions
application
full-stack AI
AI root technology
