Knowledge graph construction method and system based on large-model multi-agent interaction

Abstract

The invention aims to provide a knowledge graph construction method and system based on large-model multi-agent interaction. The method comprises the following steps: acquiring agricultural data; screening agricultural entity data from the agricultural data; calculating a relationship among the agricultural entity data by adopting an LLM model; the multi-agent interaction module corrects the relationship among the agricultural entity data through role playing; and constructing an agricultural knowledge graph according to the corrected agricultural entity data. According to the method, entity relation extraction and reasoning are enhanced by using a multi-agent interaction and role playing technology, a special model for entity extraction and a large language model (LLM) for relation extraction are combined, and integration is performed through a multi-agent framework.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a knowledge graph construction method and system based on large-model multi-agent interaction. Background Art

[0002] Knowledge graph is a graph-based data structure consisting of nodes and edges. Each node represents an "entity" and each edge is a "relationship" between entities. Knowledge graph is essentially a semantic network. Entities can refer to things in the real world, such as people, places, companies, telephones, animals, etc., and relationships are used to express certain connections between different entities. Knowledge graph is a relational network obtained by connecting all different types of information together. Therefore, knowledge graph provides the ability to analyze problems from the perspective of "relationships".

[0003] The application of knowledge graphs in the agricultural field plays an important role, which is mainly reflected in the following aspects: Agricultural big data analysis: With the advent of the era of agricultural big data, knowledge graphs can help efficiently manage things and their relationships in the real world and promote the development of smart agriculture. Through the application of knowledge graphs, agricultural data analysis can be better carried out to provide a scientific basis for agricultural production. Agricultural knowledge sharing and communication: Knowledge graphs can realize the sharing of agricultural knowledge, simplify the knowledge and experience of agricultural experts, build knowledge graphs, and allow farmers to obtain experience from experts in a wider range of fields, thereby improving the level of agricultural production development. Agricultural pest management and prevention: Knowledge graphs can help establish a knowledge base for agricultural pests and diseases, organize and classify relevant knowledge about pests and diseases, and provide reference and guidance for agricultural practitioners. By establishing a relationship model between crops and pests and diseases, predicting and analyzing the occurrence of pests and diseases, taking corresponding prevention and control measures in advance, and realizing intelligent monitoring of crop pests and diseases. Planting prediction and precision agriculture: Using knowledge graph technology, by analyzing historical meteorological information, soil information, etc., an agricultural planting prediction model is constructed to help farmers and agricultural enterprises formulate the best agricultural production plan for the growth of different crops. This model can be adjusted according to the actual observation data of individual farms, making agricultural production more scientific and refined. Agricultural education and learning: The construction of agricultural meteorology knowledge graphs, such as the knowledge graph of the "Agricultural Meteorology" course, helps learners master the basic theories of agricultural meteorology and its application in their majors, and improves teaching quality and learning effects. The application of knowledge graphs in the agricultural field covers many aspects from basic knowledge learning to practical application, which is of great significance to improving agricultural production efficiency and promoting agricultural modernization.

[0004] ‌Knowledge graphs have become a key tool for organizing and leveraging large amounts of data in various fields. In the agricultural field, knowledge graphs are particularly important for understanding the complex relationships between different entities such as crops, pests, soil types, and weather conditions. Traditional entity relationship extraction methods usually have difficulty handling the diversity and complexity of agricultural data. Summary of the invention

[0005] The purpose of the present invention is to provide a knowledge graph construction method and system based on large-model multi-agent interaction. The method enhances entity relationship extraction and reasoning by using multi-agent interaction and role-playing technology, combines a dedicated model for entity extraction and a large language model (LLM) for relationship extraction, and integrates them through a multi-agent framework.

[0006] A knowledge graph construction method based on large-model multi-agent interaction, comprising: Access to agricultural data; Filtering agricultural entity data from the agricultural data; The LLM model is used to calculate the relationship between the agricultural entity data; The multi-agent interaction module modifies the relationship between the agricultural entity data through role-playing; Construct an agricultural knowledge graph based on the corrected agricultural entity data.

[0007] Preferably, after obtaining the agricultural data, the method further includes preprocessing the agricultural data, specifically: Remove noise and duplicate agricultural data from agricultural data; Format the cleaned agricultural data.

[0008] Preferably, the filtering of agricultural entity data from the agricultural data comprises: The spacy model is used to extract agricultural entities from the preprocessed agricultural data.

[0009] Preferably, the calculating the relationship between the agricultural entity data using the LLM model comprises: The LLM model is used by spacy-llm to identify and extract the relationships between the agricultural entities.

[0010] Preferably, after calculating the relationship between the agricultural entity data using the LLM model, the method further includes constructing a plurality of intelligent agents, specifically: Configure the corresponding number of agents according to the number of entity roles that process agricultural data; The entities that process agricultural data include: agronomists, agricultural market analysts, agricultural merchants, and farmers; The agricultural expert agent is used to verify the information in the knowledge graph based on professional knowledge and make correction suggestions; The agricultural market analyst agent is used to check whether the knowledge graph reflects the current market trends and demands; The agricultural merchant agent is used to verify whether the knowledge graph contains accurate information required by the merchant.

[0011] Preferably, the multi-agent interaction module modifies the relationship between the agricultural entity data through role-playing, including: Assign each agent an entity role to process agricultural data; The agent corrects the relationship between the agricultural entity data in the role-playing, and the agricultural expert agent exchanges information with the market analysis agent to confirm the accuracy of the market-related information in the knowledge graph; The agricultural merchant agent communicates with the agricultural expert agent and provides feedback on the problems and needs in actual use of the graph; Feedback processing agent, used to collect feedback from each role, analyze problems and make preliminary corrections.

[0012] Preferably, after the multi-agent interaction module corrects the relationship between the agricultural entity data through role-playing, it also includes verifying the relationship between the agricultural entity data, specifically: Randomly extracting the relationship between the agricultural entity data as a test sample; The test samples are manually verified.

[0013] Preferably, constructing an agricultural knowledge graph based on the corrected agricultural entity data includes: Update and optimize the agricultural knowledge graph based on the revised agricultural entity data; Evaluate the agricultural knowledge graph; If the evaluation results do not meet the preset requirements, the agricultural entity data will be verified.

[0014] A knowledge graph construction system based on large-model multi-agent interaction, including: A data acquisition module, used to acquire agricultural data; A data screening module, used for screening agricultural entity data from the agricultural data; A data processing module, used for calculating the relationship between the agricultural entity data using an LLM model; A data correction module, used for the multi-agent interaction module to correct the relationship between the agricultural entity data through role-playing; The data fusion module is used to construct an agricultural knowledge graph based on the corrected agricultural entity data.

[0015] An electronic device comprises: a processor and a memory, wherein the memory is used to store computer program code, and the computer program code comprises computer instructions. When the processor executes the computer instructions, the electronic device executes a knowledge graph construction method based on large-model multi-agent interaction.

[0016] The beneficial effects of the present invention are: 1. The present invention uses multiple agents to simulate expert collaboration, thereby forming a more accurate and comprehensive relationship extraction; 2. The present invention enhances the reasoning ability of the agent through role-playing scenarios, so that the agricultural knowledge graph has a deeper insight into complex relationships; 3. The present invention combines a dedicated model for entity extraction and an LLM for relationship extraction, and utilizes the advantages of both to achieve a robust knowledge graph construction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor. Figure 1 This is a flow chart of a knowledge graph construction method based on large-model multi-agent interaction according to the present invention; Figure 2 A schematic diagram of the process of constructing an intelligent agent of the present invention; Figure 3 A schematic diagram of the process of constructing the agricultural knowledge graph of the present invention; Figure 4 A schematic diagram of the hardware structure of an electronic device of the present invention; Figure 5 A schematic diagram of multi-agent role playing and interaction of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0020] In addition, the descriptions of "first", "second", etc. in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0021] Knowledge graphs have become a key tool for organizing and leveraging large amounts of data in various fields. In the agricultural field, knowledge graphs are particularly important for understanding the complex relationships between different entities such as crops, pests, soil types, and weather conditions. Traditional entity relationship extraction methods usually have difficulty handling the diversity and complexity of agricultural data.

[0022] The present invention uses multiple intelligent agents to simulate expert collaboration, thereby forming a more accurate and comprehensive relationship extraction; the present invention enhances the reasoning ability of the intelligent agent through role-playing scenarios, so that the agricultural knowledge graph has a deeper insight into complex relationships; the present invention combines a dedicated model for entity extraction and an LLM for relationship extraction, and utilizes the advantages of both to achieve a robust knowledge graph construction process.

[0023] Example 1 A knowledge graph construction method based on large-scale multi-agent interaction, reference Figure 1 ,include: S100, obtaining agricultural data; The present invention collects agricultural data from multiple sources to ensure comprehensive information coverage.

[0024] S200, screening agricultural entity data from agricultural data; The present invention uses a dedicated model to identify and extract agricultural entities.

[0025] S300, uses the LLM model to calculate the relationship between agricultural entity data; The present invention utilizes LLM to identify and extract the relationships between agricultural entities.

[0026] S400, the multi-agent interaction module modifies the relationship between agricultural entity data through role-playing; In the present invention, each agent analyzes data according to its own role, and conducts interactive discussions to correct information. It can also enhance the ability to reason about complex relationships through scenario simulation and agent participation in role-playing.

[0027] S500, constructs an agricultural knowledge graph based on the corrected agricultural entity data.

[0028] The present invention integrates the verified entities and relationships into the knowledge graph, and continuously updates and maintains them.

[0029] Preferably, after S100, obtaining agricultural data, the method further includes S110, preprocessing the agricultural data, specifically: S111, remove noise and duplicate agricultural data from agricultural data; The functions of data cleaning mainly include improving data quality, discovering data problems, and improving analysis results. ‌Improving data quality: By cleaning data, errors can be corrected, duplicate data can be deleted, missing values ​​can be filled, etc., thereby improving the quality and credibility of the data. ‌This process ensures the accuracy and integrity of the data and provides a solid foundation for subsequent data analysis. ‌Discovering data problems: The data cleaning process helps to discover anomalies, errors or inconsistencies in the data, so that they can be corrected and processed in a timely manner. ‌This helps to solve potential problems before data analysis and avoid misleading conclusions caused by data quality problems. ‌Improving analysis results: Clean and accurate data can significantly improve the accuracy and predictive power of analytical models. ‌Through data cleaning, the reliability and validity of the analysis results can be ensured, and analytical deviations caused by data quality problems can be avoided. ‌Data cleaning is a key step in ensuring data quality. It not only improves the availability and reliability of data, but also provides a solid foundation for data analysis, while also meeting industry-specific regulations and standards requirements. S112, formatting the cleaned agricultural data.

[0030] The role of formatting data is mainly reflected in improving the readability, accessibility, processing efficiency of data, and ensuring the seamless integration and sharing of data between different systems and applications. The purpose of formatting data is to make data easier to process, analyze, store and transmit. Through formatting, data can follow specific rules and conventions, thereby improving the readability and accessibility of data. This process helps to reduce data errors and redundancy and improve data processing efficiency. The importance of data formatting lies in its ability to ensure the seamless integration and sharing of data between different systems and applications, which is crucial to improving work efficiency and communication effectiveness.

[0031] Preferably, S200, screening agricultural entity data from agricultural data includes: The spacy model is used to extract agricultural entities from the preprocessed agricultural data.

[0032] Spacy models are mainly used for natural language processing (NLP) tasks, especially text analysis and understanding. Spacy is an open source Python library that provides a large number of NLP functions, including part-of-speech tagging, named entity recognition, dependency parsing, etc. These functions make Spacy very effective in processing texts in multiple languages, including Chinese. The role of the Spacy model is mainly reflected in the following aspects: Spacy can help developers analyze texts, identify nouns, verbs, adjectives, etc., and provide basic data processing for subsequent NLP tasks. Spacy can identify named entities in texts, such as names of people, places, and organizations, which is very important for tasks such as information extraction and knowledge graph construction. Through dependency parsing, Spacy can analyze the grammatical relationship between words in a sentence to help understand the structure and meaning of the sentence. In addition to English, Spacy also supports multiple languages, including Chinese, which enables it to process text data in different languages.

[0033] In addition, Spacy also supports the use of pre-trained word embedding models, which can capture the semantic relationship between words, so as to perform tasks such as word similarity calculation and cloze test. This is very helpful for understanding the deep meaning of text and performing more advanced NLP applications. In general, the Spacy model makes it easier and more efficient to process and analyze text by providing comprehensive NLP functions and supporting multiple languages. Preferably, S300, calculating the relationship between agricultural entity data by using the LLM model includes: The LLM model is used through spacy-llm to identify and extract the relationships between agricultural entities.

[0034] spacy-llm is a tool that integrates the spaCy natural language processing (NLP) framework with large language models (LLMs, such as OpenAI's GPT-3 or GPT-4). By combining spaCy's efficient processing capabilities with LLM's powerful language understanding capabilities, more accurate and comprehensive natural language processing tasks such as entity recognition and relationship extraction can be achieved. How spacy-llm works: Data input and preprocessing: Raw text data is input into the spaCy pipeline for preprocessing. spaCy is responsible for basic text processing tasks such as word segmentation, part-of-speech tagging, and dependency parsing. Entity recognition: Use the trained spaCy model for entity recognition to extract relevant entities (such as names of people, places, organizations, etc.) from the text. Context generation: After identifying the entity, pass the entity and its surrounding context information to the large language model (LLM). Context information helps LLM better understand the potential relationships between entities. Relation extraction: After receiving the entity and context information, LLM uses its powerful natural language understanding and generation capabilities to identify the relationship between entities. LLM can handle complex language structures and long-distance dependencies, thereby extracting more accurate and in-depth relationships. Result integration: The relationship information extracted by LLM is passed back to spaCy and integrated into the original NLP pipeline. In this way, spaCy can combine the output of LLM with its own processing results to generate a complete knowledge graph containing entities and relationships. Verification and storage: The extracted entity and relationship information needs to be verified to ensure its accuracy and relevance. Verification can be performed through rule verification, manual review, or comparison with existing knowledge bases. The verified entity and relationship information is stored in the knowledge graph for subsequent query and analysis. Through the integration of spacy-llm, the advantages of spaCy and LLM can be fully utilized to improve the accuracy and efficiency of natural language processing tasks, especially in tasks that need to handle complex language structures and long-distance dependencies, such as relationship extraction and knowledge graph construction.

[0035] Preferably, reference Figure 2 , S300, after using the LLM model to calculate the relationship between agricultural entity data, S310, also includes building multiple intelligent agents, specifically: S311, configuring a corresponding number of agents according to the number of entity roles processing agricultural data; The entities that process agricultural data include: agronomists, agricultural market analysts, agricultural merchants, and farmers; S312, the agricultural expert agent is used to verify the information in the knowledge graph based on professional knowledge and make correction suggestions; S313, the agricultural market analyst agent is used to check whether the knowledge graph reflects the current market trends and demands; S314, the agricultural merchant agent is used to verify whether the knowledge graph contains the accurate information required by the merchant.

[0036] The LLM model, or large language model, has a wide range of applications and functions, and plays an important role in education, health care and other fields. The large language model can handle a variety of natural language tasks, such as text classification, question answering, dialogue, etc., and is an important path to artificial intelligence. At present, the large language model uses a Transformer architecture and pre-training objectives similar to the small model. The difference from the small model is the increase in model size, training data and computing resources.

[0037] Preferably, reference Figure 5 ,S400,The multi-agent interaction module modifies the relationship between agricultural entity data through role playing, including: Assign each agent an entity role to process agricultural data; The agents correct the relationship between agricultural entity data in role-playing, and the agricultural expert agents exchange information with the market analysis agents to confirm the accuracy of market-related information in the knowledge graph; The agricultural merchant agent communicates with the agricultural expert agent and provides feedback on the problems and needs in actual use of the graph; Feedback processing agent, used to collect feedback from each role, analyze problems and make preliminary corrections.

[0038] Agent is an important concept in the field of artificial intelligence (AI), which refers to an intelligent entity that can autonomously perceive the environment, make decisions and perform actions. It can be a program, a system or a robot. The core of the agent lies in its artificial intelligence algorithm, including machine learning, deep learning, reinforcement learning, neural network and other technologies. Through these algorithms, the agent can learn from a large amount of data and improve its performance, and continuously optimize its decision-making and behavior. At the same time, the agent also has the following characteristics: Autonomous decision-making: The agent can be given a high-level or even vague goal and take independent actions to achieve the goal. Interaction with the outside world: The agent can freely use different software tools, such as the GPT-4-based agent AutoGPT, which can autonomously search for relevant information on the Internet and automatically write code and manage business according to user needs. Can run indefinitely: The agent allows human operators to "set it and forget it", and is evolvable and can gradually optimize itself through feedback during the work process. The application scenarios of the agent are wide, including but not limited to smart elderly care, cultural and tourism industries, etc. For example, in the tourism industry, agents can generate answers by integrating information from multiple sources, and provide travel services such as hotel and attraction tickets, which greatly saves users' time. Currently, research on agents is developing rapidly, especially agents based on large language models (LLMs), which have demonstrated significant potential in achieving reasoning and planning capabilities comparable to humans. Multi-agent systems provide advanced capabilities by specializing LLMs into different agents with different capabilities, and enabling these diverse agents to interact with each other to effectively simulate complex real-world environments.

[0039] Preferably, in S400, after the multi-agent interaction module corrects the relationship between the agricultural entity data through role-playing, it also includes verifying the relationship between the agricultural entity data, specifically: Randomly extract the relationships between agricultural entity data as test samples; Perform manual verification on the test samples.

[0040] The verification of knowledge graph relationships is mainly carried out through knowledge verification. In the process of knowledge graph generation, knowledge verification plays a vital role, especially when forming knowledge graphs through real-time reasoning. It is no longer possible to screen the quality of triples manually, and knowledge verification needs to be performed through algorithms, which can greatly improve the quality and availability of knowledge graphs. Lightweight machine learning models will not slow down the speed of the entire process too much, and are a feasible direction. In addition, in order to imitate the human brain's ability to store knowledge, the implicit reasoning network (IRN) model can be used to verify entity relationships in knowledge graphs. This model designs a shared memory component to implicitly store knowledge base information, and can implicitly learn multi-step reasoning processes by reading shared memory components without human intervention, simulating the process of the human brain reading memory when making reasoning judgments. This model forms multiple intermediate representations during the prediction process, and uses the RNN controller to determine whether enough information is encoded to generate output predictions, thereby ensuring the accuracy and effectiveness of the relationships in the knowledge graph. ‌‌The verification of knowledge graph relationships not only relies on algorithms for knowledge verification, but also includes the use of machine learning and deep learning technologies to improve the accuracy of tasks such as entity recognition and relationship extraction, thereby ensuring the accuracy and effectiveness of relationships in the knowledge graph.‌

[0041] Preferably, reference Figure 3 , S500, constructing an agricultural knowledge graph based on the corrected agricultural entity data includes: S510, updating and optimizing the agricultural knowledge graph according to the revised agricultural entity data; S520, evaluate the agricultural knowledge graph; S530: If the evaluation result does not meet the preset requirements, the agricultural entity data is verified.

[0042] Knowledge graph evaluation is a process of systematically evaluating and analyzing the quality and performance of knowledge graphs. This process involves many aspects, including but not limited to the completeness, accuracy, consistency, scalability, and user-friendliness of the knowledge graph. The purpose of the evaluation is to ensure that the knowledge graph can effectively support various application scenarios, such as education, scientific research, and enterprise decision support. The methods of knowledge graph evaluation mainly include: Sampling detection: Check whether the knowledge and data in the knowledge graph are accurate through random sampling or priority sampling based on the popularity of entities. Consistency detection: Check the knowledge conflicts in the knowledge graph through preset consistency detection rules to ensure the logical consistency and accuracy of the knowledge. External knowledge comparison evaluation: Use high-quality external knowledge sources as benchmarks to compare and evaluate the quality of the knowledge graph. This method uses manually proofread external knowledge sources for efficient and accurate quality detection. User feedback: Collect user feedback to understand the performance of the knowledge graph in actual applications, as well as user needs and suggestions, so as to make improvements. In addition, the evaluation also includes the analysis of knowledge graph construction technology, such as data acquisition and information extraction, to ensure technical feasibility and efficiency. The results of the evaluation will directly affect the application effect and user experience of the knowledge graph.

[0043] Example 2 A knowledge graph construction system based on large-model multi-agent interaction, including: A data acquisition module, used to acquire agricultural data; A data screening module, used to screen agricultural entity data from agricultural data; A data processing module for calculating the relationship between agricultural entity data using the LLM model; The data correction module is used by the multi-agent interaction module to correct the relationship between agricultural entity data through role-playing; The data fusion module is used to construct an agricultural knowledge graph based on the corrected agricultural entity data.

[0044] The invention can be applied to a variety of agricultural fields, including: Crop management, understanding the relationship between crops, soil types, pests, and weather conditions. Pest control, identifying the best pest control strategy based on historical data and expert knowledge. Soil health, analyzing soil health data to recommend appropriate interventions to increase crop yields. Climate impacts, studying the impact of climate change on agricultural practices and recommending adaptive strategies.

[0045] Example 3 An electronic device includes: a processor and a memory, the memory is used to store computer program code, the computer program code includes computer instructions, when the processor executes the computer instructions, the electronic device executes a knowledge graph construction method based on large-model multi-agent interaction.

[0046] refer to Figure 4 , the electronic device 2 includes a processor 21, a memory 22, an input device 23, and an output device 24. The processor 21, the memory 22, the input device 23, and the output device 24 are coupled via a connector, and the connector includes various interfaces, transmission lines, or buses, etc., which are not limited in the embodiments of the present invention. It should be understood that in various embodiments of the present invention, coupling refers to mutual connection in a specific manner, including direct connection or indirect connection through other devices, for example, through various interfaces, transmission lines, buses, etc.

[0047] The processor 21 may be one or more graphics processing units (GPUs). When the processor 21 is a GPU, the GPU may be a single-core GPU or a multi-core GPU. Optionally, the processor 21 may be a processor group consisting of multiple GPUs, and the multiple processors are coupled to each other via one or more buses. Optionally, the processor may also be other types of processors, etc., which are not limited in the embodiments of the present invention.

[0048] The memory 22 can be used to store computer program instructions and various computer program codes including program codes for executing the scheme of the present invention. Optionally, the memory includes but is not limited to random access memory (RAM), read-only memory (ROM), erasable programmable read only memory (EPROM), or portable read only memory (CD-ROM), which is used for related instructions and data.

[0049] The input device 23 is used to input data and / or signals, and the output device 24 is used to output data and / or signals. The output device 24 and the input device 23 may be independent devices or an integrated device.

[0050] The present invention uses multiple intelligent agents to simulate expert collaboration, thereby forming a more accurate and comprehensive relationship extraction; the present invention enhances the reasoning ability of the intelligent agent through role-playing scenarios, so that the agricultural knowledge graph has a deeper insight into complex relationships; the present invention combines a dedicated model for entity extraction and an LLM for relationship extraction, and utilizes the advantages of both to achieve a robust knowledge graph construction process.

[0051] The foregoing is merely a specific embodiment of the present invention, which enables those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A knowledge graph construction method based on large-model multi-agent interaction, characterized in that: include: Access to agricultural data; Filtering agricultural entity data from the agricultural data; The LLM model is used to calculate the relationship between the agricultural entity data; The multi-agent interaction module modifies the relationship between the agricultural entity data through role playing; Construct an agricultural knowledge graph based on the corrected agricultural entity data.

2. According to claim 1, a knowledge graph construction method based on large model multi-agent interaction is characterized in that: After obtaining the agricultural data, the agricultural data is preprocessed, specifically: Remove noise and duplicate agricultural data from agricultural data; Format the cleaned agricultural data.

3. According to claim 2, a knowledge graph construction method based on large model multi-agent interaction is characterized in that: The filtering of agricultural entity data from the agricultural data comprises: The spacy model is used to extract agricultural entities from the preprocessed agricultural data.

4. According to claim 1, a knowledge graph construction method based on large model multi-agent interaction is characterized in that: The calculation of the relationship between the agricultural entity data using the LLM model includes: The LLM model is used by spacy-llm to identify and extract the relationships between the agricultural entities.

5. According to claim 1, a knowledge graph construction method based on large model multi-agent interaction is characterized in that: After the LLM model is used to calculate the relationship between the agricultural entity data, it also includes constructing multiple intelligent agents, specifically: Configure the corresponding number of agents according to the number of entity roles that process agricultural data; The entities that process agricultural data include: agronomists, agricultural market analysts, agricultural merchants, and farmers; The agricultural expert agent is used to verify the information in the knowledge graph based on professional knowledge and make correction suggestions; The agricultural market analyst agent is used to check whether the knowledge graph reflects the current market trends and demands; The agricultural merchant agent is used to verify whether the knowledge graph contains accurate information required by the merchant.

6. According to claim 1, a knowledge graph construction method based on large model multi-agent interaction is characterized in that: The multi-agent interaction module corrects the relationship between the agricultural entity data through role-playing, including: Assign each agent an entity role to process agricultural data; The agent corrects the relationship between the agricultural entity data in the role-playing, and the agricultural expert agent exchanges information with the market analysis agent to confirm the accuracy of the market-related information in the knowledge graph; The agricultural merchant agent communicates with the agricultural expert agent and provides feedback on the problems and needs in actual use of the graph; Feedback processing agent, used to collect feedback from each role, analyze problems and make preliminary corrections.

7. A method for constructing a knowledge graph based on large-model multi-agent interaction according to claim 6, characterized in that: After the multi-agent interaction module corrects the relationship between the agricultural entity data through role-playing, it also includes verifying the relationship between the agricultural entity data, specifically: Randomly extracting the relationship between the agricultural entity data as a test sample; The test samples are manually verified.

8. The method for constructing a knowledge graph based on large-model multi-agent interaction according to claim 7 is characterized in that: The construction of the agricultural knowledge graph based on the corrected agricultural entity data includes: Update and optimize the agricultural knowledge graph based on the revised agricultural entity data; Evaluate the agricultural knowledge graph; If the evaluation results do not meet the preset requirements, the agricultural entity data will be verified.

9. A knowledge graph construction system based on large-model multi-agent interaction, characterized in that: include: A data acquisition module, used to acquire agricultural data; A data screening module, used for screening agricultural entity data from the agricultural data; A data processing module, used for calculating the relationship between the agricultural entity data using an LLM model; A data correction module, used for the multi-agent interaction module to correct the relationship between the agricultural entity data through role-playing; The data fusion module is used to construct an agricultural knowledge graph based on the corrected agricultural entity data.

10. An electronic device, characterized in that: include: A processor and a memory, wherein the memory is used to store computer program code, wherein the computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device executes a knowledge graph construction method based on large-model multi-agent interaction as described in any one of claims 1 to 8.

Images