CN-121998043-A - Knowledge graph construction based on multiple agents

CN121998043ACN 121998043 ACN121998043 ACN 121998043ACN-121998043-A

Abstract

The disclosure provides a method for multi-agent based knowledge graph construction, comprising obtaining knowledge data, obtaining a demand description for a target knowledge graph, and generating the target knowledge graph based on the knowledge data and the demand description by using a layered multi-agent framework, wherein the layered multi-agent framework comprises a plurality of agents arranged in layers.

Inventors

ZHANG HANSI

Assignees

罗伯特·博世有限公司

Dates

Publication Date: 20260508
Application Date: 20241101

Claims (15)

1. A method for multi-agent based knowledge graph construction, comprising: Obtaining knowledge data; Obtaining a demand description for the target knowledge graph, and The target knowledge-graph is generated based on the knowledge data and the demand description using a hierarchical multi-agent framework, wherein the hierarchical multi-agent framework includes a plurality of agents arranged in a hierarchy.
2. The method of claim 1, wherein, The layered multi-agent framework comprises a management layer, a scheduling layer and a working layer, and The management layer corresponds to a top layer of the layered multi-agent frame, the work layer corresponds to a bottom layer of the layered multi-agent frame, and the scheduling layer corresponds to an intermediate layer between the top layer and the bottom layer of the layered multi-agent frame.
3. The method of claim 2, wherein the generating the target knowledge-graph comprises: and constructing an intelligent agent through the knowledge graph at the management layer, and generating a knowledge graph construction plan.
4. The method of claim 3, wherein the generating a knowledge-graph construction plan comprises: acquiring agents and filling the agents with knowledge data through an ontology at the scheduling layer, generating a plurality of candidate sub-plans, and The knowledge graph construction plan is generated based on the plurality of candidate sub-plans by the knowledge graph construction agent.
5. The method of claim 4, wherein the generating a plurality of candidate sub-plans comprises: Generating at least one ontology acquisition sub-plan based on the demand analysis result by the ontology acquisition agent, and Generating, by the knowledge data population agent, at least one knowledge data population sub-plan based on the at least one ontology acquisition sub-plan and a data type of the knowledge data.
6. The method of claim 5, wherein, The demand analysis results are generated by a demand analysis agent at the working layer based on the demand description, and The demand analysis result comprises at least one of a range, a capability problem and a key concept of the target knowledge graph.
7. The method of claim 5, wherein the generating at least one ontology acquisition sub-plan based on demand analysis results comprises: Accessing an ontology library; identifying whether there is a candidate ontology matching the requirement analysis result in the ontology library, and And generating the at least one ontology acquisition sub-plan according to the identification result.
8. The method of claim 5, wherein the data type of the knowledge data comprises a structured data type or an unstructured data type.
9. The method of claim 3, wherein the generating the target knowledge-graph further comprises constructing an agent from the knowledge-graph: and distributing corresponding tasks to the plurality of agents at the scheduling layer according to the knowledge graph construction plan, wherein the corresponding tasks comprise an entity acquisition task for acquiring the agents at the scheduling layer and a knowledge data filling task for filling the agents with knowledge data at the scheduling layer.
10. The method of claim 9, wherein the ontology acquisition agent is configured to perform one of the following operations for the ontology acquisition task: Under the condition that the ontology acquisition task is associated with the candidate ontology reuse in the ontology library as a target ontology, distributing an ontology reuse subtask to an ontology reuse agent in the working layer; Under the condition that the ontology acquisition task is associated with a development target object on the basis of candidate ontologies in an ontology library, distributing ontology development subtasks to ontology development ontologies in the working layer, distributing ontology reuse subtasks to ontology reuse ontologies in the working layer, and distributing ontology alignment subtasks to ontology alignment ontologies in the working layer; And distributing the ontology development subtasks to the ontology development agents at the working layer under the condition that the ontology acquisition tasks are associated with the development target specimens from scratch.
11. The method of claim 9, wherein the knowledge data population agent is to perform one of the following for the knowledge data population task: In case the knowledge data filling task is associated with performing knowledge data filling for an ontology using a data mapping manner, assigning a knowledge mapping subtask to a knowledge mapping agent at the working layer, or And under the condition that the knowledge data filling task is associated with executing knowledge data filling aiming at the ontology by utilizing a data identification and extraction mode, assigning a named entity identification subtask to a named entity identification agent at the working layer, assigning a relationship extraction subtask to a relationship extraction agent at the working layer, and assigning a data fusion subtask to a data fusion agent at the working layer.
12. The method of claim 1, wherein the layered multi-agent framework is configured to provide at least one of: short-term memory of intermediate data of the construction process of the target knowledge graph; long-term memory of result data of the construction process of the target knowledge graph.
13. An apparatus for multi-agent based knowledge graph construction, comprising: Memory, and A processor coupled to the memory, the processor configured to perform the method of any of claims 1-12.
14. A computer readable medium storing a computer program comprising instructions which, when executed by a processor, cause the processor to be configured to perform the method of any one of claims 1-12.
15. A computer program product comprising computer-executable instructions that, when executed, cause one or more processors to perform the method of any of claims 1-12.

Description

Knowledge graph construction based on multiple agents Technical Field The present disclosure relates generally to the field of computers, and more particularly to methods and apparatus for multi-agent (multi-agent) based knowledge-graph construction. Background The concept of AI agents has been proposed in the field of artificial intelligence (ARTIFICIAL INTELLIGENCE, AI). An AI agent, also referred to simply as an agent, is an intelligent entity driven by an AI model. The agent may typically have capabilities such as perception, decision-making, action, etc. driven by the AI model, thereby enabling the agent to autonomously accomplish a given goal without requiring a human to specify each step of operation. Disclosure of Invention According to one aspect of the disclosure, a method for multi-agent based knowledge-graph construction is provided, comprising obtaining knowledge data, obtaining a demand description for a target knowledge-graph, and generating the target knowledge-graph based on the knowledge data and the demand description using a hierarchical multi-agent framework, wherein the hierarchical multi-agent framework comprises a plurality of agents arranged in a hierarchy. According to another aspect of the present disclosure, there is provided an apparatus for multi-agent based knowledge-graph construction, comprising a memory and a processor. The processor is coupled to the memory and configured to perform the method according to any of the various embodiments of the disclosure. According to yet another aspect of the present disclosure, there is provided a computer-readable medium storing a computer program comprising instructions that, when executed by a processor, cause the processor to be configured to perform a method according to any of the various embodiments of the present disclosure. According to yet another aspect of the present disclosure, there is provided a computer program product comprising computer-executable instructions that, when executed, cause one or more processors to perform the method according to any of the various embodiments of the present disclosure. Drawings Various embodiments of the claimed subject matter will now be described, by way of example, with reference to the accompanying drawings. The use of the same reference symbols in different drawings indicates identical or similar items. Fig. 1 shows a schematic diagram of an overall architecture for building a knowledge-graph using a layered multi-agent framework, according to an example embodiment of the present disclosure. Fig. 2 shows a block diagram of an exemplary architecture of an agent at a management layer according to an exemplary embodiment of the present disclosure. Fig. 3 illustrates a block diagram of an exemplary structure of an agent at a dispatch layer according to one exemplary embodiment of the present disclosure. Fig. 4 shows a block diagram of an exemplary architecture of an agent at a working layer according to an exemplary embodiment of the present disclosure. Fig. 5 shows a schematic diagram of an exemplary structure of a layered multi-agent framework for building a knowledge-graph, according to an example embodiment of the present disclosure. Fig. 6 illustrates a timing diagram of operations for performing knowledge-graph construction using the layered multi-agent framework described in connection with fig. 5, according to an example embodiment of the present disclosure. Fig. 7 illustrates an example flowchart of a method for multi-agent based knowledge-graph construction, according to an example embodiment of this disclosure. Fig. 8 illustrates an exemplary apparatus for multi-agent based knowledge-graph construction in accordance with an exemplary embodiment of the disclosure. Detailed Description In the following description, numerous specific details are set forth in order to provide a thorough understanding of embodiments of the present disclosure. One skilled in the relevant art will recognize, however, that the disclosure may be practiced without one or more of the specific details, or with alternative methods, components, etc. In some instances, well-known structures, operations are not shown or described in detail to avoid unnecessarily obscuring the present disclosure. In view of the advantages of agents, for example, driven by large language models (Large Language Model, LLM), in natural language processing, attempts are currently being made to introduce agent technology into various natural language processing task scenarios in order to increase task processing efficiency. A typical complex natural language processing task scenario may include knowledge graph construction. Knowledge graph is a semantic network that reveals relationships between knowledge entities, provides a rich, machine-understandable semantic model, and can be used to implement knowledge query, knowledge reasoning, etc. Common knowledge-graph construction procedures include stages such as extracting knowledge f