Agent-based policy generation method, apparatus, device, and medium
By processing multi-source learning data and task description text through an intelligent agent, learning semantic vectors and task context vectors are generated, enabling accurate knowledge matching and real-time decision suggestions. This solves the problem of time-consuming knowledge retrieval and improves task execution efficiency and decision quality.
CN122414232APending Publication Date: 2026-07-17SHENZHEN COOCAA NETWORK TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENZHEN COOCAA NETWORK TECH CO LTD
- Filing Date
- 2026-03-24
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In the process of employees processing unstructured knowledge, knowledge retrieval is time-consuming and semantics cannot be understood in a timely manner, which affects task execution efficiency and decision-making quality.
Method used
The agent acquires multi-source learning data, encodes it to generate learning semantic vectors, combines it with task description text to generate task context vectors, performs similarity matching, generates task decision suggestions, and sends them in real time.
Benefits of technology
It improves the accuracy and real-time performance of strategy generation, ensures timely knowledge retrieval and consistent semantic matching, and enhances task execution efficiency and decision-making quality.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122414232A_ABST
Abstract
本发明公开一种基于智能体的策略生成方法、装置、设备及介质，方法包括：通过智能体获取多源学习数据，并对多源学习数据进行编码处理，得到学习语义向量；通过智能体获取用于描述用户当前执行任务的任务描述文本；通过智能体对任务描述文本进行编码处理，得到任务上下文向量；基于任务上下文向量，智能体对学习语义向量进行相似度匹配，得到目标语义向量；基于目标语义向量，通过智能体生成任务决策建议，并向用户发送任务决策建议。本公开的方法，通过智能体将多源学习数据和任务描述文本编码为向量数据，从而在统一语义空间中对二者进行匹配，得到目标语义向量，继而生成用户所需的任务决策建议，提高了策略生成的精准性与实时性。
Need to check novelty before this filing date? Find Prior Art