Page 295 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Use Case 2: Empowering cloud network operation with self-

intelligence through the network large model toolchain platform 4.3 - 5G

Organization: China Telecommunication Corporation

Country: China

Contact Person(s):

Primary contact: Yupei Wang, wangyp6@ chinatelecom .cn
Secondary contact: Yungeng Zhang, zhangyg5@ chinatelecom .cn
Any other contacts: Feng Wang, wangfeng6@ chinatelecom .cn

1 Use Case Summary Table

Item Details
Category artificial Intelligent

- Lack of unified AI workflows in telecom operations
- Difficulty integrating diverse multimodal telecom data
- Limited accessibility of intelligent tools for network teams
Problem
Addressed - Single-model limitations for complex operational needs
- High resource costs for model training and deployment
- Model training and inference deployment require large amounts of
arithmetic resources, which is costly
- Unified end-to-end toolchain for data, model, and deployment work-
flows
Key Aspects of - Built-in multimodal model support for telecom-specific data
Solution - Low-code development environment for agent construction
- Intelligent orchestration with knowledge bases and plug-ins
- Scalable arithmetic resources for AI training and inference

Large Language Models (LLMs), Network AI Toolchain, Model-as-a-Service
Technology (MaaS),Intelligent Agents, Multimodal Data Processing, Cloud Network
Keywords Management, AI Infrastructure, Automated Network Operations, AI-driven
Orchestration, Federated Fine-tuning
Private data: we use some sensitive words from
Data Availability https:// github .com/ jkiss/ sensitive -words,
but do some processing in the combination with telecom data features.

Metadata (Type of China Telecom’s massive network data and knowledge corpora.
Data) Text data, Logs, alarms, and configurations

259