Page 289 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

(continued)

Item Details
Testbeds or Pilot The testbed is used internally, and the use case is part of a broader product
Deployments development effort. 4.3 - 5G
Code Reposito- Not available
ries

2 Use Case Description

2�1 Description

The development of large models has brought new opportunities for industries, but it has also
presented challenges, including a high computing power threshold, increased costs of training
and reasoning, difficulty in selecting base models, high technical complexity in deployment,
and difficulty in scene adaptation. In order to reduce the technical threshold of AI modeling
and application engineering, boost the scale of AI large model application, and promote the
process of AI+ action in various industries, this case establishes a set of one-stop modeling and
application toolchain platform covering the whole process of AI, providing full-stack services
and realizing the efficient sharing of data, arithmetic power, models, tools, and applications.

AI intelligent middle platform, “1+1+3+N” architecture [5]:

• 1 An integrated platform for annotation, training, and inference.
• 2 MaaS model services covering images, text, and videos.
• 3 AI tool platforms (Large Model and Agent Aggregation Platform,Visual AI analysis
application platform, Voice customer service application construction platform).
• N applications in 10+ industries, such as transportation, culture and tourism, city,
government, and enterprise management.

This use case changes the limitations of existing local code development, such as high technical
threshold, low efficiency, poor sharing, etc., and provides AI R&D guidelines and resource
sharing through a one-stop platform.

253