Data analysis method and device based on multi-agent cooperation, and storage medium

By employing a multi-agent collaborative data analysis method, this approach utilizes workflow orchestration agents and a large language model to decompose user requests and generate execution plans. This solves the existing challenges of automatic orchestration and agent collaboration in data analysis tasks, achieving end-to-end automation and efficient data analysis.

CN121073371BActive Publication Date: 2026-07-21TSINGHUA UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
TSINGHUA UNIVERSITY
Filing Date
2025-08-06
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing Data+AI technologies lack sufficient semantic understanding, reasoning, and planning capabilities when dealing with diverse scenarios, making it difficult to achieve automatic orchestration of system workflows and efficient collaboration of intelligent agents, resulting in low adaptability and efficiency of data analysis tasks.

Method used

A multi-agent collaborative data analysis method is adopted. The workflow orchestration agent receives user requests, calls the data exploration agent to obtain metadata, uses a large language model to decompose and orchestrate sub-tasks, combines the capability profiles of task processing units to generate execution plans, and the data engine schedules agents to coordinate execution, thereby achieving end-to-end automated data analysis.

Benefits of technology

It significantly reduces the need for manual intervention, improves the modularity, scalability, and maintainability of data analysis, enhances the efficiency and resource utilization of execution plans, and strengthens the robustness and intelligence of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121073371B_ABST
    Figure CN121073371B_ABST
Patent Text Reader

Abstract

A data analysis method and device based on multi-agent cooperation and a storage medium are disclosed. The method comprises: receiving a natural language request by a workflow arrangement agent, and generating an execution plan comprising a plurality of subtasks and corresponding task processing units by consulting a data exploration agent and a data engine understanding agent; in the execution stage, converting the subtask into an executable instruction by a professional agent matched with the subtask, and calling and coordinating the corresponding task processing unit to execute the instruction by a data engine scheduling agent. Through the hierarchical cooperative architecture, the separation of planning, translation and execution is realized, and the automation level, modularity and scalability of data analysis are significantly improved.
Need to check novelty before this filing date? Find Prior Art