人工智能关键技术发展现状与趋势研究  

Research on the Development Status and Trends of Key Technologies in Artificial Intelligence

作　　者：李培源 曾卫民 赵帅 韦安垒 LI Pei-yuan;ZENG Wei-min;ZHAO Shuai;WEI An-lei(China Telecom Cloud Computing Corporation,Beijing 100007,China)

机构地区：[1]天翼云科技有限公司,北京100007

出　　处：《新一代信息技术》2024年第1期36-40,共5页New Generation of Information Technology

摘　　要：在当今全球科技革命和产业变革中,人工智能处于最前沿的领域。本文对人工智能领域的各项关键技术的发展现状和趋势进行综述和探讨。诸如:AI芯片通过制程优化、架构创新提升算力;芯片使能技术优化软件架构,提升异构融合能力;AI开发框架增强兼容性、灵活性和可移植性;大模型在复杂任务处理、多模态信息整合方面取得突破;云智一体推动算力、算法和数据深度融合,促进AI产业生态形成,等。这些技术共同推动AI向更高效、更智能和更广领域应用发展。通过本研究,我们期望能够为政策制定者、科研人员和行业从业者提供有益的参考和启示。In the ongoing global technological revolution and industrial transformation,artificial intelligence(AI)stands at the forefront.This paper reviews and discusses the current status and trends of various key technologies in the field of AI.Specifically,AI chips enhance computing power through process optimization and architectural innovation;Chip-enabling technologies optimize software architectures,thereby improving heterogeneous integration capabilities;AI development frameworks bolster compatibility,flexibility,and portability;Large models have made breakthroughs in handling complex tasks and integrating multimodal information;The integration of cloud computing and AI promotes the deep fusion of computing power,algorithms,and data,facilitating the formation of an AI industrial ecosystem.These technologies collectively propel AI towards more efficient,intelligent,and widespread applications.Through this research,we aspire to provide valuable reference and insights for policymakers,researchers,and industry practitioners.

关 键 词：AI芯片 芯片使能 AI开发框架 大模型 云智一体 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]