This page is being moved to a new, faster, and mobile-friendly application! Access the enhanced and centralized experience now on MyWorkspace.
ITU's 160 anniversary

Connecting the world and beyond

  •  

ITU-T work programme

[2022-2024] : [SG16] : [Q5/16]

[Declared patent(s)]

Work item: F.MGCReqs
Subject/title: Requirements and functional architecture of model generalization system for telecommunication intelligent customer service
Status: [Carried to next study period]
Approval process: AAP
Type of work item: Recommendation
Version: New
Equivalent number: -
Timing: -
Liaison: -
Supporting members: China Telecom; Beijing UPT; Zhejiang Lab
Summary: AI models have been used in a variety of industries, bringing new changes to the industry. With the development of telecommunication, there are more and more types of services, and more requirements for intelligent customer service, applying AI technology to telecommunication intelligent customer service has become a trend. But there are still the following problems to be solved in the application of AI in intelligent customer service: (1) The current AI models are oriented to a single scenario and it's difficult to adapt to multiple scenarios; (2)The data collected needs to rely on manual data labeling, and data quality and labeling efficiency cannot be guaranteed; (3) AI models in the customer service mostly deal with the single type of data, which makes it difficult to deal with multimedia data at the same time; (4) Customer service has a large number of scenarios, and AI models need to be generated for different scenarios. As a result, the number of AI models is large, the training cost is high, and the effect of AI models cannot be guaranteed. Therefore, this proposal introduces the requirements and functional framework of a model generalization system for telecommunication intelligent customer service, which supports efficient annotation of massive data, can handle multimedia data. Model generalization system supports rapid adaptation to tasks through fine-tuning, and achieves the sharing and reuse of AI capabilities. The customer service oriented models could be created by analyzing and understanding multiple types of data for comprehensive analysis of multiple scenarios.
Comment: -
Reference(s):
  Historic references:
Contact(s):
Yupei WANG, Editor
Meiling DAI, Editor
Yuying XUE, Editor
Xiaohou SHI, Editor
Yaqi SONG, Editor
ITU-T A.5 justification(s):
Generate A.5 drat TD
-
[Submit new A.5 justification ]
See guidelines for creating & submitting ITU-T A.5 justifications
First registration in the WP: 2023-08-22 12:09:50
Last update: 2024-09-20 10:28:54