GB/T 46351-2025Artificial intelligence - Multi-algorithm management technical requirements (English PDF)

Foreword

This document complies with the provisions of GB/T 1.1-2020 "Standardization Work Guidelines Part

1.Structure and Drafting Rules of Standardization Documents". Drafting. Please note that some content in this document may involve patents. The issuing organization of this document assumes no responsibility for identifying patents. This document was proposed and is under the jurisdiction of the National Information Technology Standardization Technical Committee (SAC/TC28). This document was drafted by: Hangzhou Hikvision Digital Technology Co., Ltd., China Electronics Technology Standardization Institute, and Sichuan University. West China Hospital, Northwestern Polytechnical University, Harbin Institute of Technology (Shenzhen), Institute of Automation, Chinese Academy of Sciences, Beijing Jiaotong University, Beijing University of Aeronautics and Astronautics Tianjin University, Zhejiang Dahua Technology Co., Ltd., Saifite Engineering Technology Group Co., Ltd., Kunlun Digital Intelligence Technology Co., Ltd., China China National Petroleum Corporation Changqing Oilfield Branch, Guoneng Information Technology Co., Ltd., Beijing Zhongshu Ruizhi Technology Co., Ltd., and Inspur Software Technology Co., Ltd., China Mobile (Hangzhou) Information Technology Co., Ltd., Hangzhou QuChain Technology Co., Ltd., Hangzhou High-tech Zone (Binjiang) Blockchain In collaboration with the Data Security Research Institute, GRG Banking Group Co., Ltd., CETC Big Data Research Institute Co., Ltd., and CETC Jinxin Software Co., Ltd. Company, Wanda Information Co., Ltd., Shanghai Computer Software Technology Development Center, Shanghai Artificial Intelligence Research Institute Co., Ltd., Shenzhen Tencent Computer Systems Co., Ltd., Inspur Electronic Information Industry Co., Ltd., Shenzhen Simou Information Technology Co., Ltd., China Mobile Communications Group Group Limited, China Telecom Corporation Limited Chongqing Branch, and Beijing Ansheng Technology Co., Ltd. The main drafters of this document are. Chen Feng, Qian Xiaodong, Ma Shanshan, Xu Yang, Cao Wenlin, Zhao Junyu, Nie Jiandi, Lin Quanyu, Xu Jiaqi, and Yin Jin. Shen Zhiyue, Wang Zhaojun, Zhang Yanning, Wu Jianlong, Nie Liqiang, Zhu Guibo, Jia Yijun, Tao Renshuai, Ma Yuqing, Kong Weisheng, Li Di, Wang Juan, Wang Xiaohong Liang Qian, Ma Tongsen, Zhang Tianlin, Xie Kailang, Li Zhaochuan, Wang Kechen, Zhang Qing, Pu Guiyang, Wang Zhongwei, Du Jingyi, Wei Zunbo, Hu Wenkui, Cai Huimin Lü Yiliang, Zhang Jingyi, Chen Mingang, Song Haitao, Wang Zikai, Huang Chao, Lu Lu, Liu Shu, Qin Rizhen, Lai Su, Wang Chenzi, Liu Yifan, Wang Song. Artificial Intelligence Multi-Algorithm Management Technology Requirements

1 Scope

GB/T 46351-2025 is the Chinese national standard covering running many models on one platform - the registration and versioning of each algorithm, the resource allocation and scheduling between them, the routing of a request to the right model, the monitoring of each one's performance, and the retirement of the ones that have drifted. At 22,500 words. First edition, already in force. Issued on 5 October 2025, it has been in force since 5 October 2025.

This document establishes the framework for a multi-algorithm management system, specifying the algorithm repository, algorithm scheduling service, algorithm packages, and the multi-algorithm management system interface. Technical requirements. This document applies to the design, research, development, implementation, deployment, acceptance, and evaluation of multi-algorithm management systems.

2 Normative references

The contents of the following documents, through normative references within the text, constitute essential provisions of this document. Dated citations are not included. For references to documents, only the version corresponding to that date applies to this document; for undated references, the latest version (including all amendments) applies. This document.

GB/T 41867-2022 Terminology for Information Technology and Artificial Intelligence

3 Terms and Definitions

The terms and definitions defined in GB/T 41867-2022, as well as the following terms and definitions, apply to this document.

3.1 Multi-algorithm management Manage algorithms from different manufacturers, with different functions, on different platforms, and in different versions.

3.2 algorithm model A model is a model built by training on a dataset using a specific machine learning algorithm, including the model structure and parameters.

3.3 Algorithm task The process by which an application system uses intelligent algorithms to process and analyze data.

Note. Taking images as an example, the algorithm tasks include image recognition, image classification, object detection, and image retrieval.

4.Abbreviations The following abbreviations apply to this document. SoC. System-on-Chip

5.Multi-Algorithm Management System Framework A multi-algorithm management system is a software system capable of algorithm management (add, delete, modify, query) and algorithm scheduling, consisting of an algorithm repository and an algorithm scheduling service.