算法溯源简述  

An Overview of Algorithm Traceability

作　　者：王亭亭 崔伟群[1] WANG Tingting;CUI Weiqun(National Institute of Metrology,Beijing 100029,China)

机构地区：[1]中国计量科学研究院,北京100029

出　　处：《计量科学与技术》2023年第5期23-30,共8页Metrology Science and Technology

基　　金：中国计量科学研究院基本科研业务费所自主项目(AKYZZ2238)。

摘　　要：数字时代,融入人工智能等新一代信息技术的软件算法,是加快推进制造业进行数字化转型的关键要素之一。基于不同工业革命时期测量仪器和标准在量值传递和溯源中表现的特征,以及测量仪器中软件算法对量值准确可靠的影响,提出算法溯源理念,目的是从计量溯源的角度实现对软件算法在程序实现过程中输出量值的定量评价,从而保证软件算法输出量值的准确和可靠。同时给出了算法溯源的定义、参考标准、方法和溯源结果等的综合性描述,并就算法溯源的研究现状进行了讨论,为进一步开展算法溯源的计量技术应用奠定理论基础。In the digital age,software algorithms imbued with artificial intelligence and other next-generation information technologies serve as crucial factors accelerating the digital transformation of the manufacturing sector.The concept of algorithm traceability has been proposed based on the distinct characteristics manifested by measuring instruments and standards during metrological value transfer and traceability in various stages of the industrial revolution,along with the impact of software algorithms within measuring instruments on the accuracy and reliability of metrological values.The objective of this concept is to quantitatively assess the output values of software algorithms during their execution from a metrological traceability perspective,thereby ensuring their precision and dependability.This paper offers a comprehensive exposition of algorithm traceability,including its definition,reference standards,methodologies,and traceability outcomes.Current research trends in algorithm traceability are discussed,laying a theoretical groundwork for further exploration into metrological applications of algorithm traceability.

关 键 词：计量学 软件算法 算法溯源 标准参考数据 算法标准 

分 类 号：TB91-64[一般工业技术—计量学]