Recently, the AIGC Platform, jointly developed by the Informatization Department of our university and Midea Group, has been officially launched and put into operation. Built on the open-source framework Dify and deployed with localized large language model computing power, the platform achieves a closed-loop data processing system within the campus. While ensuring data security and privacy, it provides faculty and students with a high-performance, low-latency intelligent application experience. Integrating agent construction and knowledge base management, the platform is now open to all faculty and students free of charge, marking an important step for our university in empowering education, teaching, scientific research innovation and management services with artificial intelligence.
As a key achievement of deepening industry-education integration between our university and Midea Group, the launch of the AIGC Platform is a concrete practice of collaborative innovation in the field of artificial intelligence between the two parties. Relying on the technological advantages of Midea Group and the development capabilities of the university’s Informatization Department, the platform deeply integrates cutting-edge industrial technologies with educational needs. It not only provides convenient AI tools for faculty and students, but also builds an open and secure intelligent application ecosystem. This cooperation model fully demonstrates our university's active exploration in promoting digital transformation and deepening school-enterprise collaborative education.
Core Functions of the Platform
Shengong Xiaozhi AI Portal
Integrating the independently developed “Shengong Xiaozhi” intelligent service platform, it provides efficient and practical AI services including AI Q&A, personal knowledge base, and smart campus connection.
Intelligent Agent Development
With no programming background required, you can quickly create personalized AI agents through the visual workflow orchestration interface.
For teachers: Develop course Q&A assistants, thesis guidance assistants, experiment simulation assistants, etc.
For students: Create English learning partners, programming debug assistants, postgraduate entrance exam review planners, etc.
Custom Knowledge Base: Inject course materials, thesis data and other resources into the agent to make it more tailored to your needs.
Knowledge Base Management
Upload documents, web links, notes and other content, and the platform will automatically build a searchable knowledge base, supporting intelligent Q&A and content generation.
Teaching scenarios: Import courseware and references into the knowledge base for students to inquire and get accurate AI answers at any time.
Research scenarios: Quickly sort out literature context, extract key information, and assist in thesis writing.
Learning scenarios: Organize error collections and course notes; the AI will help summarize and identify knowledge gaps.
How to Use
Access method: Search for “AIGC Platform Portal” in the campus application center.
Get started: After entering the platform, click “AI Application Development” to build your own AI agent. For detailed operation guidelines, please follow the subsequent information release and training notices.
Basic Guide to Intelligent Agent Development
Click “AI Application Development” on the AIGC portal homepage to enter the agent development platform, where personalized AI agents can be created in the application development module.
Application Development
The platform provides normal dialogue mode, workflow orchestration mode and programmable plug-in development capabilities, supporting the full life cycle management of agents including query, editing and deletion.
Application Type Description
Normal Dialogue Assistant: Develop agents through prompts + knowledge base + plug-ins, easy to use, suitable for simple dialogue assistants.
Programmable Dialogue Assistant: Develop agents through visual orchestration with higher flexibility, adaptable to various agent development scenarios.
Programmable Plug-in: Develop plug-ins through visual orchestration; published plug-ins will be added to the plug-in library and can be referenced by normal dialogue assistants, programmable dialogue assistants and plug-in development.
Workflow Orchestration Operation
Workflow orchestration decomposes complex tasks into multiple steps (nodes) for processing. Each node's input and output can be flexibly configured, and integrated debugging and testing tools are available to realize powerful custom logic. Detailed usage of each node in the workflow can be viewed by clicking the help link on the page to access the Dify online documentation.
Locally Deployed Large Models
Currently, the AIGC Platform has opened 4 large models for free use by all faculty and students. More models will be connected and opened gradually in the future.
Translated by Basic Courses Department
