Page 218 - AI for Good-Innovate for Impact Final Report 2024
P. 218
AI for Good-Innovate for Impact
Use case – 51: A New Mode and Practice of Human-Robot
Interaction Operation and Maintenance Based on Network Large
Language Model (AI Chat Operations, AIChatOps)
Country: China
Organization: China Telecom Jiangsu Branch
Contact Person: Wenhan Rong
+86 15301582482
rongwh@ chinatelecom .cn
51�1� Use case summary table
Domain Telecommunications
The problem to be Traditional network maintenance lacks real-time scheduling based
addressed on intent, relying on manual operations or limited automation.
Key aspects of the Using Large Language Model (LLM) for intent recognition, mobile
solution data query, and execution. Implementing ChatOps for human-robot
interaction in network maintenance.
Technology keywords Large Language Model (LLM), AI Chat Operations, ChatOps,
Network Maintenance Automation
Data availability Data is privately available.
Metadata (type of Textual data, Numerical data, Categorical data, Image data, and
data) Video data
Model Training and Fine-tuning LLM for task scheduling and parameter completion.
fine-tuning Implementing multi-agent systems for execution efficiency.
Testbeds or pilot Not publicly available
deployments
51�2� Use case description
51�2�1� Description
Existing problem: Traditional network maintenance mainly relies on manual operation and
equipment inspection or network management systems to carry out specific and inherent
automated operations, which cannot realize real-time scheduling based on intent.
Solution: (1) Develop the ability to quickly query and execute data on mobile devices. (2)
Based on the intention recognition ability of the Large Language Model、LLM、, retrieve the
most similar automation capabilities and extract the parameters based on the natural language
instructions input by the user. (3) Based on the network LLM intelligent agent technology,
202
