Scientific agent system for interdisciplinary scientific research task and control method thereof
By leveraging the collaborative work of the intent recognition module, tool retrieval module, multidisciplinary intelligent agent cluster, planning intelligent agent, and code intelligent agent, the shortcomings of existing scientific research intelligent agent systems in structured knowledge representation and multidisciplinary task processing are addressed, achieving stability and automated planning for interdisciplinary scientific research tasks.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INST OF AUTOMATION CHINESE ACAD OF SCI
- Filing Date
- 2026-02-05
- Publication Date
- 2026-06-09
AI Technical Summary
Existing scientific research intelligent agent systems lack structured knowledge representation, struggle to handle complex task chains, and fail to cover the multidisciplinary coupling needs in real-world scientific research, resulting in low system versatility for interdisciplinary scientific research tasks and low automated task planning capabilities.
An intent recognition module performs contextual semantic parsing, a tool retrieval module accurately locates inference tools, a multi-disciplinary intelligent agent cluster collaboratively processes data, a planning intelligent agent performs structured task decomposition and dynamic allocation, a code intelligent agent achieves automated planning, a summary intelligent agent generates reports, a review intelligent agent verifies logical consistency, and a research skills database supports professional processes.
It improves the system's stability in understanding complex scientific research tasks, reduces tool-calling illusions, ensures the rigor and feasibility of scientific research processes, and enhances the versatility and automated planning capabilities of interdisciplinary scientific research tasks.
Smart Images

Figure CN122174859A_ABST