AI/ML

AI-native 6G

Digital twin networks

Definitions

Artificial intelligence (AI): Development of computer systems capable of performing tasks that historically required human intelligence, such as recognizing speech, making decisions, and identifying patterns. AI is an umbrella term that encompasses a wide variety of technologies, including machine learning, deep learning, and natural language processing (NLP).
Machine learning (ML): This is a subfield of AI that uses algorithms trained on data sets to create self-learning models that are capable of predicting outcomes and classifying information without human intervention. It is used today for a wide range of commercial purposes, including suggesting products to consumers based on their past purchases, predicting stock market fluctuations, translating text from one language to another, and much more.
Generative AI (Gen AI): It is a type of AI that has the capability to learn and reapply the properties and patterns of data for a wide range of applications, from creating text, images, and videos in different styles to generating tailored content. It enables machines to perform creative tasks previously thought exclusive to humans.

AI in RAN - standard evolution

The AI-RAN alliance focuses on three key areas, both in 5G and 6G:

AI and RAN: AI and RAN integration to utilize the same infrastructure generating new AI-driven revenue opportunities.
AI on RAN: AI services deployed at network edge enabled by RAN to increase operational efficiency and to offer new services.
AI for RAN: advanced RAN by AI to improve spectral efficiency

In terms of standardization status, here's the summary.

3GPP's progress on AI / ML for 5G started from Release 15 - a new Network Function, NWDAF (Network Data Analytics Function) was introduced and kept enhancing in Rel-16 & Rel-17.In OA&M aspect, 3GPP introduced the MDAF (Management Data Analytics Function) by using AI/ML techniques for data analytics.
In Rel-17, 3GPP started a study item (SI) on AI/ML for NG-RAN with three use cases - (1) Energy saving, (2) Load balancing, (3) Mobility optimization, then moving forward to a work item in Rel-18. Now, In Rel-19, enhancement for AIML for NG-RAN will be the SI followed by WI with two new use cases of AI/ML-based NG-RAN to study including (1) Coverage and capacity optimization (CCO) and (2) Network slicing; as well as to work on the R18 leftovers.
In Rel-18, as the 1st release of 5G-Advanced, 3GPP has studied AI/ML for NR Air Interface with three use cases - (1) CSI feedback enhancement, (2) Beam management, (3) Positioning accuracy enhancement; two of them, AI-based Beam management and AI-enabled Positioning accuracy enhancement will move to the normative phase in Rel-19, while use case of CSI feedback enhancement will continue to study in Rel-19.
Release 19 is the 2nd release of 5G-Advanced in 3GPP, also it serves as a bridge to 6G standardization. The standardization of the Release 19 discussion started from 2023 and is expected to freeze in 2025. The Release 19 timeline is shown in the figures below:

Overview of AI/ML related work in 3GPP

AI models have been in use by vendors and network operators for some time, as an implementation choice, enhancing some conventional methodology in areas such as network management and automation (SON, etc) as well as various processing algorithms at the network and device sides.AI/ML is pervasive and that its reach into 3GPP’s work is increasing – expected to touch virtually every aspect of the system. Although AI/ML models are not standardized, there are significant drivers to encourage standardization around AI/ML in 6G, with AI/ML to be in Rel-21 specifications from day 1. In addition, there is a great potential for specifications to be less ‘watertight’ in the 6G releases, allowing for AI/ML based implementations using the best data-driven parameterization for some functionalities.
The priority in 3GPP groups has been to ensure some basic – but key – advances in the areas of:

Infrastructure/operator control.
AI/ML model performance monitoring.
AI/ML model activation and deactivation.
Air interface extensions specific to AI/ML implementations.
Establishing early standards-based approaches for devices’ data collection, as well as AI/ML model transfer and delivery.
Testing, interoperability and consistent device behaviour.

The agreed principles, driven by operators and resembling some of the principles of the European Commission’s AI Act, are:

Data must be secured and data integrity & confidentiality ensured.
Data privacy, anonymity and user consent respected.
Operators to retain control of standardized data collection transfer process & to manage data transfer to the server for UE-side data collection, without the need for SLA (includes initiating, terminating & managing data transfer).
Operator to have full visibility for standardized data.
Design to be futureproof & extendable.

Page updated

Google Sites

Report abuse