Valid

GB/T 46284-2025Artificial intelligence - Technical specifications of federated learning (English PDF)

人工智能　联邦学习技术规范

Preview PDF

This is a limited preview

Buy now to download the full PDF (22 pages)

Need quotation for your management?

Get a formal quotation for GB/T 46284-2025 in 30 minutes.

Issued by

SAMR; SAC

Level / Type

National · Recommended

Issue date

October 5, 2025

Implementation date

October 5, 2025

Scope

GB/T 46284-2025 is the English-translated version of 人工智能　联邦学习技术规范.

GB/T 46284-2025 is the Chinese standard "Artificial intelligence - Technical specifications of federated learning". Its scope clause reads: This document establishes the architecture, participants, and workflow of the federated learning system, and specifies its functional requirements and performance.

The requirements describe the corresponding testing methods. This document applies to the design, development, testing, use, and maintenance of federated learning systems, and to conducting horizontal comparisons of federated learning application systems.

Selection. Note. This document only covers the system functionalities and performance requirements for implementing and supporting federated learning methods; it does not provide details on general-purpose computer hardware and software or communication equipment itself. Detailed regulations. Its clauses include terms and definitions.

Found what you need?

Document preview — GB/T 46284-2025

National Standard of the People's Republic of China

ICS: 35.240
Classification: L 60

Issued by: State Administration for Market Regulation; Standardization Administration of the PRC

Preface
1.Scope
2 Normative References
3.Terms and Definitions
4.Abbreviations
5.Overview of Federated Learning Systems
5.1 Federated Learning System Architecture
5.2 Participants in Federated Learning
5.3 Federated Learning Task Process
6.1 Basic Technology Module
6.2 Calculation Module
6.3 Management Module
7.Federated Learning System Performance Requirements
7.1 Federal Data Probing
7.2 Federated Modeling
7.3 Federalized Reasoning
7.4 Federal Data Quality and Incentive Mechanisms
8.Functional Testing Methods for Federated Learning Systems
8.1 Basic Technology Module
8.2 Calculation Module
8.3 Management Module
9.Performance Testing Methods for Federated Learning Systems
9.1 Federal Data Probing
9.2 Federated Modeling
9.3 Federalized Reasoning
9.4 Federal Data Quality and Incentive Mechanisms

1 Scope

This document establishes the architecture, participants, and workflow of the federated learning system, and specifies its functional requirements and performance.

The requirements describe the corresponding testing methods.

This document applies to the design, development, testing, use, and maintenance of federated learning systems, and to conducting horizontal comparisons of federated learning application systems.

Selection.

Note. This document only covers the system functionalities and performance requirements for implementing and supporting federated learning methods; it does not provide details on general-purpose computer hardware and software or communication equipment itself.

Detailed regulations.

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 Terminology for Information Technology and Artificial Intelligence

3 Terms and Definitions

The terms and definitions defined in GB/T 41867 and the following terms and definitions apply to this document.

3.1 Federated learning

Multiple participants interact in a manner that protects privacy, while ensuring that their original data does not leave the trusted domain defined by the data provider.

Methods for collaboratively completing artificial intelligence and machine learning tasks by using inter-data.

3.2

A comprehensive computer and communication system for implementing federated learning methods.

Note. This system includes not only various computer hardware and software (such as algorithms, CPUs, GPUs, etc.) for executing federated learning algorithms, but also supports remote communication between nodes.

The communication protocols and equipment used for data transmission (e.g., network devices and data transmission protocols). This system ensures that each participant can protect the original data locally.

Under the premise of privacy, machine learning tasks can be completed collaboratively by exchanging intermediate model updates (such as gradients or parameters).

3.3

In federated learning methods that divide data into samples, each participant's dataset has the same feature space but different sample spaces.

3.4

In a feature-based federated learning approach, each participant's dataset has the same sample space but different feature spaces.

GB/T 38665.1 | GB/T 38665.2 | GB/T 9813.3 | GB/T 45434.2 | GB/T 11460 | […]

......
This preview omits tables, figures, formulas and parts of the technical clauses. The complete document — 22 pages — is available in the English PDF.

Referenced standards

Normative references

GB/T 41867Information technology—Artificial intelligence—Terminology

Editions of GB/T 46284

Edition	Title	Revision	Status
GB/T 46284-2025	Artificial intelligence - Technical specifications of federated learning	current edition	Current

This page sells the current edition, GB/T 46284-2025. Earlier editions are listed for reference only.

How to Download GB/T 46284-2025

1
Add to cart
Click "Download PDF" on this page and choose "Buy this standard". You can add more standards before checkout.
2
Checkout
Enter your email and billing details. Payment is processed securely by Stripe (cards or bank transfer). Prefer bank details? Request a quotation (min. order $300).
3
Instant delivery (0–9 sec)
Delivery is automatic: within seconds of payment you'll receive an email with a secure download link. The link stays valid for 72 hours.
4
Invoice included
A tax invoice is attached to the confirmation email. Need a custom invoice? Contact us.