Bid invitation file generation method, system and device and storage medium

Abstract

The invention relates to the technical field of bidding and tendering, and provides a bidding document generation method, system and device and a storage medium, and the method comprises the steps: constructing a bidding knowledge graph through at least one of a policy and regulation library, a bidding document library and an industry standard library in a bidding knowledge base, and obtaining a bidding document generation result based on the project information of a bidding project and the bidding knowledge graph; and performing bid invitation file generation by applying the bid invitation file generation model to obtain the bid invitation file corresponding to the bid invitation project, thereby overcoming the defects of low efficiency, error proneness, poor flexibility and adaptability, knowledge update lagging and insufficient personalization in the existing bid invitation file production. A large model technology and a dynamically updated bid invitation knowledge graph are introduced to generate the bid invitation file, policies, regulations and project characteristics can be dynamically adapted, rapid and accurate bid invitation file generation is realized, the timeliness and reliability of the generated file are ensured, dependence on personal subjective experience is avoided, the flexibility is extremely high, and the application range is extremely wide.

Description

Technical Field

[0001] The present invention relates to the technical field of bidding, and in particular to a bidding document generation method, system, device and storage medium. Background Art

[0002] In the field of bidding, the bidding documents are the core basis of the entire bidding activities. The preparation process is not only complicated but also extremely time-consuming. The content of the bidding documents must strictly comply with the relevant laws and regulations, technical standards and the unique needs of the specific project. Any negligence in any link may have a serious impact on the fairness, impartiality and effectiveness of the bidding activities.

[0003] At present, the production of bidding documents mainly adopts manual writing or template-based semi-automatic generation. The manual writing method requires the tendering personnel to write each part of the bidding document, such as project overview, technical requirements, etc., one by one according to the project requirements. Although this method is flexible, it is inefficient and easily affected by human factors, resulting in incomplete or inaccurate document content, which in turn affects the smooth progress of the bidding activities. The template-based semi-automatic generation method is to fill in information through a preset template to generate a bidding document. Compared with the previous method, this method reduces the workload of manual writing and improves efficiency, but the template has poor versatility. Different projects have diverse requirements, and it is difficult for the template to be fully adapted, resulting in obvious deficiencies in the targeting and adaptability of the generated bidding documents. Summary of the invention

[0004] The present invention provides a bidding document generation method, system, device and storage medium, which are used to solve the problems in the prior art of low efficiency, poor accuracy and consistency of manual writing, and poor flexibility and adaptability of template-based semi-automatic generation, difficulty in coping with complex scenarios, inability to meet the diverse needs of different projects, and inability to dynamically adapt to policy changes and project characteristics. The present invention can efficiently generate bidding documents and ensure the accuracy and consistency of the documents, as well as their alignment with policies, regulations and industry standards.

[0005] The present invention provides a method for generating a bidding document, comprising: Obtain project information of bidding projects; Based on the project information and the bidding knowledge base, a bidding document generation model is applied to generate bidding documents to obtain bidding documents corresponding to the bidding project; The bidding document generation model is obtained by training on the basis of a large language model using sample project information corresponding to a sample bidding project and sample bidding documents corresponding to the sample bidding project; the bidding knowledge base is dynamically updated; The tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; based on the project information and the tender knowledge base, a tender document generation model is applied to generate a tender document, and the tender document corresponding to the tender project is obtained, including: Based on at least one of the policy and regulation library, the tender document library, and the industry standard library, a tender knowledge graph is constructed; the tender knowledge graph is dynamically updated; Based on the project information and the tender knowledge graph, a tender document generation model is applied to generate a tender document, and the tender document corresponding to the tender project is obtained.

[0006] According to a tender document generation method provided by the present invention, constructing a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library includes: Performing entity recognition and relationship extraction based on policy clause information and legal and regulatory information in the policy and regulation library to obtain a policy and regulation triple; Performing entity recognition and relationship extraction based on historical project information and historical tender documents in the tender document library to obtain a tender document triple; Performing entity recognition and relationship extraction based on industry technical information and format standard information in the industry standard library to obtain an industry standard triple; Constructing the tender knowledge graph based on at least one of the policy and regulation triple, the tender document triple, and the industry standard triple.

[0007] According to a tender document generation method provided by the present invention, after generating the tender document corresponding to the tender project by applying a tender document generation model based on the project information and the tender knowledge base, it further includes: Extracting key information from the tender document to obtain the tender key information in the tender document; Based on the tender key information, applying the tender document generation model to perform a review of the tender document to obtain the review result of the tender document; Wherein, the review of the tender document includes at least one of document compliance check, content risk prediction, and clause conflict detection.

[0008] According to a tender document generation method provided by the present invention, after obtaining the review result of the tender document by applying the tender document generation model based on the tender key information, it further includes: In the case where the review result indicates that the review of the tender document fails, determining problem information from the tender document based on the review result; Display the problem information and the review result; Among them, the review result includes the problem type corresponding to the problem information and a modification suggestion.

[0009] According to a tender document generation method provided by the present invention, after the display of the problem information and the review result, it further includes: Receiving a modification instruction fed back by the user based on the displayed problem information and the review result, and updating the tender document based on the modification instruction to obtain the updated tender document; Or, receiving the updated tender document fed back by the user based on the displayed problem information and the review result; Storing the updated tender document, or transmitting the updated tender document to the tender platform.

[0010] According to a tender document generation method provided by the present invention, the generation of the tender document corresponding to the tender project by applying the tender document generation model based on the project information and the tender knowledge base includes: Extracting key information from the project information to obtain project key information; Determining a generation prompt text based on the project key information; Based on the generation prompt text and the tender knowledge base, applying the tender document generation model to generate a tender document to obtain the tender document corresponding to the tender project.

[0011] The present invention also provides a tender document generation system, on which a tender document generation model and a tender knowledge base are deployed; the tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; The tender document generation model is used to construct a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library, and generate a tender document based on the project information of the tender project and the tender knowledge graph to obtain the tender document corresponding to the tender project; the tender knowledge graph is dynamically updated; Among them, the tender document generation model is trained based on the sample project information corresponding to the sample tender project and the sample tender document corresponding to the sample tender project on the basis of a large language model; the tender knowledge base is dynamically updated.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and running on the processor. When the processor executes the computer program, it implements the tender document generation method as described in any one of the above.

[0013] The present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the tender document generation method described in any one of the above.

[0014] The tender document generation method, system, device and storage medium provided by the present invention construct a tender knowledge graph through at least one of the policy and regulation library, tender document library and industry standard library in the tender knowledge base, and based on the project information of the tender project and the tender knowledge graph, apply the tender document generation model to generate tender documents, obtaining the tender documents corresponding to the tender project, overcoming the defects of low efficiency, error-prone, lack of unified standards, poor flexibility and adaptability, lagging knowledge update and insufficient personalization in the traditional manual writing method and the semi-automated generation method based on templates. By introducing large model technology and a dynamically updated tender knowledge graph, and generating tender documents based on project requirements on this basis, it can provide a reliable policy and legal basis for the generation of tender documents, as well as important technical support, thus realizing the rapid generation of tender documents, ensuring the accuracy and consistency of the generated tender documents, while being able to align with the latest policies and regulations and cutting-edge industry standards, deeply associated with project characteristics, and effectively adapted to project requirements, thereby making the generated tender documents more organized, more accurate, more timely and reliable. It not only gets rid of the dependence on personal subjective experience, but also can dynamically adapt to policies, regulations and project characteristics, and can be applied to various industries and various complex scenarios, with extremely strong flexibility and a very wide scope of application. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 is a flowchart showing the tender document generation method provided by the present invention; Figure 2 is one of the example diagrams of the tender documents provided by the present invention; Figure 3 is another example diagram of the tender documents provided by the present invention; Figure 4 is a structural diagram of the tender document generation device provided by the present invention; Figure 5 is a structural diagram of the tender document generation system provided by the present invention; Figure 6 is a structural diagram of the electronic device provided by the present invention. Detailed implementation manners

[0017] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0018] In the field of bidding, the preparation of bidding documents is complex and time-consuming because the content of the documents needs to comply with regulatory requirements, technical standards and project requirements. Currently, the preparation of bidding documents mainly relies on manual writing or semi-automated generation based on templates, but both have obvious defects, which are specifically as follows: First, low efficiency: The manual writing method requires professional personnel to manually sort out a large amount of document content according to project requirements, laws and regulations, and past experience, and write each part of the bidding documents one by one, including project overview, technical requirements, evaluation methods, contract terms, etc. For complex large-scale projects, it will consume a lot of time and energy, and the efficiency is extremely low.

[0019] Second, poor accuracy: The manual writing method is affected by subjective factors and is prone to omissions, errors or unclear expressions, etc., which affect the quality of the bidding documents; in the semi-automated generation method based on templates, due to the poor generality of the templates, it is difficult to adapt to different project requirements, and it is also very easy to cause the generated documents to not match the actual project, thus affecting the accuracy.

[0020] Third, poor consistency: Under the manual writing method, the documents written by different bidders vary greatly in style and content structure, lack a unified standard, and have poor consistency.

[0021] Fourth, poor adaptability: The semi-automated generation method based on templates is difficult to flexibly meet the special requirements of different projects, and it is easy to have the situation that the template does not match the project.

[0022] Fifth, lagging knowledge update: Currently, neither the manual writing method nor the semi-automated generation method based on templates can keep up with the latest policies and regulations and cutting-edge industry standards, and cannot achieve real-time maintenance and dynamic update of relevant knowledge in the field of bidding.

[0023] Sixth, insufficient personalization: Currently, the content of the bidding documents prepared by the manual writing method and the semi-automated generation method based on templates lacks a deep association with the project characteristics, and the clause matching accuracy and personalization degree are insufficient.

[0024] In response to this, the present invention provides a method for generating a tender document, aiming to overcome the defects of traditional solutions, achieving the rapid generation of tender documents, ensuring the accuracy and consistency of the generated tender documents, being applicable to various industries and complex scenarios, meeting the diverse needs of different projects, dynamically adapting to policy changes and industry standards, enhancing the standardization and efficiency of tendering and bidding activities, and providing reliable technical support for public resource trading.

[0025] Figure 1 FIG. is a schematic flowchart of the method for generating a tender document provided by the present invention. As Figure 1 shown, the method includes: Step 110, obtaining project information of a tender project; Step 120, based on the project information and a tender knowledge base, applying a tender document generation model to generate a tender document for the tender project, obtaining the tender document corresponding to the tender project; Among them, the tender document generation model is trained based on a large language model, applying sample project information corresponding to a sample tender project and a sample tender document corresponding to the sample tender project; the tender knowledge base is dynamically updated.

[0026] Specifically, considering the problems of low efficiency, error-proneness, and poor consistency existing in the current manual writing method, as well as the lack of flexibility and adaptability in the semi-automated generation method based on templates, and the problems of lagging knowledge update and insufficient personalization existing in both, in the embodiments of the present invention, it is proposed to introduce large model technology and a knowledge base in the field of tendering and bidding, so as to utilize the powerful capabilities of the large model in multiple fields and knowledge levels, and the dynamically updated tender knowledge base to generate tender documents, so as to improve the generation efficiency and quality of tender documents, obtain accurate and reliable tender documents, and the content of the documents can be deeply associated with project characteristics, adapted to project requirements, aligned with the latest policies, regulations and cutting-edge industry standards, and can be applicable to various industries and various complex scenarios, with extremely strong flexibility and a very wide scope of application.

[0027] In detail, in practical applications, before generating a tender document, it is first necessary to determine the project on which the tender document to be generated depends, that is, the tender project, and the relevant information of this project in the tendering and bidding activities, such as the tender scope, budget amount, technical requirements, evaluation methods, project name, etc. Such information is essentially the specific requirements of this project corresponding to the tender document, and can be collectively referred to as the project information corresponding to the tender document generation task.

[0028] Among them, the tender project can be a project to be tendered in various regions, fields, and industries. For example, it can be a tender project in the medical field, construction field, transportation field, etc., or a tender project in different industries under the same field. For example, tender projects in the pharmaceutical industry and medical device industry under the medical field. It can also be a tender project in the same or different fields in different regions. The embodiments of the present invention do not make specific limitations on this.

[0029] Here, the project information can be directly provided by the user or extracted from a large amount of text descriptions and various project requirements provided by the user. That is, when the user provides various project requirements in a standardized manner, for example, directly inputting the tender scope, budget amount, technical requirements, etc., the project requirements input by the user can be directly used as the project information at this time; on the contrary, when the user describes a large amount of text content in a natural language and cannot directly clarify each project requirement, the text content provided by the user can be refined for key points and key information can be extracted at this time, and the project information can be determined according to the extracted information.

[0030] Further, after obtaining the project information of the tender project, in the embodiments of the present invention, the tender document can be generated according to this project information, and the tender document can be generated according to each project requirement in the project information, so as to obtain the tender document corresponding to the tender project, realize the fast and accurate generation of the tender document, and ensure the reliability and consistency of the generated tender document, as well as align with the latest policies, regulations and cutting-edge industry standards, reduce the dependence on professional personnel, get rid of the influence of subjective factors, and be able to obtain an accurate, reliable and timely tender document, thus providing a strong basis for the normal progress of the bidding and tendering activities and a key support for the orderly development of public resource trading.

[0031] Specifically, considering the problems of low efficiency, poor accuracy and consistency existing in the manual writing method currently used for tender document generation, the lack of flexibility and adaptability in the semi-automated generation method based on templates, as well as the current popularity of large language models (hereinafter referred to as large models) and the significant improvement of large model capabilities, that is, the excellent capabilities of large language models in text understanding, semantic analysis, clause analysis, policy interpretation, and rule reasoning, the embodiments of the present invention propose to introduce large model technology to provide a specific implementation means for the automated generation of tender documents and solve the existing problems at the same time.

[0032] It can be understood that, in view of various existing problems, in the embodiments of the present invention, specifically when generating a tender document, the tender document is generated by using large model technology based on the project information of the tender project, so as to obtain the tender document corresponding to the tender project. That is, the large model is used to generate the tender document according to various specific project requirements, so as to obtain a tender document that is deeply associated with the tender project and meets the project requirements.

[0033] Here, it is worth noting that in order to make the finally generated tender document be deeply associated with the tender project and truly match the project requirements, in the embodiments of the present invention, when generating the tender document, the large model can be used to deeply mine the project characteristics of the tender project; that is, based on the project information, or the project information and other relevant information of the tender project, information mining can be carried out through the large model to deeply mine the project characteristics, such as project scale, technical characteristics, business characteristics, etc., and combine them into the tender document generation process. In this way, the finally generated tender document can not only be deeply associated with the project characteristics of the tender project, but also match the project requirements, achieving multiple improvements in personalization degree, accuracy, adaptability and efficiency.

[0034] Furthermore, considering that although the tender document directly generated based on the large model has obvious improvements in accuracy, consistency, etc., the tender document needs to be adapted to specific policies, regulations and industry standards, and the policies and regulations change over time, and the industry standards are also constantly updated. Therefore, in order to further improve the timeliness, accuracy and reliability of the generated document, in the embodiments of the present invention, it is proposed that a knowledge base in the field of bidding and tendering, that is, a tender knowledge base, can be introduced to generate the tender document based on this, so that the finally generated tender document can be aligned with the latest policies, regulations and cutting-edge industry standards, thus ensuring the timeliness and reliability of the tender document generation. That is, based on the tender knowledge base containing the latest policies, regulations and cutting-edge industry standards, the large model is used to generate the tender document, so as to obtain a tender document with high timeliness, accuracy and personalization degree.

[0035] Among them, the tender knowledge base contains a large amount of information such as sorted and structured laws, regulations, policy clauses, industry standards, historical tender documents, etc., and these information can be obtained by technical means, such as web crawling from authoritative websites (official databases of laws and regulations, websites of industry regulatory agencies, public resource trading platforms, etc.).

[0036] Moreover, it is worth noting that the tender knowledge base here is maintained in real time and updated dynamically, that is, the information in it is constantly changing. Specifically, it can be updated according to real-time policies and regulations, cutting-edge industry standards, historical tender information (project information and tender documents of the latest announced tender projects), etc. In this way, the timeliness and accuracy of the data in the library can be ensured. The real-time policies and regulations, cutting-edge industry standards, historical tender information, etc. here can be obtained regularly / irregularly from authoritative websites. Here, the information in the dynamically updated tender knowledge base can be continuously updated and optimized over time. This dynamic update mechanism can keep the information in the library at the forefront of the industry and in sync with the latest policies and regulations, thus making the tender documents generated based on this more timely, accurate and reliable.

[0037] Specifically, the process of generating tender documents here can be to first build a model for generating tender documents, which can be called a tender document generation model. That is, a tender document generation model can be built on the basis of a large model. Here, a large model refers to a natural language processing model with a large parameter scale, where the number of model parameters and / or the complexity of the model structure exceed a preset threshold, and the model has processed a large amount of text data during training and has the ability to understand and generate natural language. For example, the large model can include the Spark large model, etc.

[0038] Immediately afterwards, this tender document generation model and the tender knowledge base can be applied to generate tender documents, and the tender documents corresponding to the tender project output by the tender document generation model can be obtained. Here, specifically, on the basis of the project information, the tender document generation model can be used to generate documents, so as to automatically analyze and process the project information through the excellent text understanding, semantic analysis and other capabilities of the model, and further combine the laws, regulations, policy terms, industry standards, etc. in the tender knowledge base to generate tender documents. That is, the model interprets the knowledge in the tender knowledge base through its own powerful clause analysis, policy interpretation, and rule reasoning capabilities, and on this basis, combines the understood project information to generate tender documents. In this way, the quality of the generated tender documents can be ensured.

[0039] It should be noted that, compared with the problems of low efficiency, error-proneness, lack of unified standards, poor flexibility and adaptability, lagging knowledge update, and insufficient personalization existing in the traditional methods of manual writing and semi-automated generation based on templates, the embodiments of the present invention can automatically integrate multi-source data and dynamically generate high-quality bidding documents; that is, on the basis of the bidding knowledge base, using the bidding document generation model to generate bidding documents, it can deeply understand the project information, mine the project characteristics of the bidding project, and can deeply analyze the policies and regulations, deeply understand the regulatory provisions, and analyze the technical architecture of industry standards, so as to quickly and accurately generate bidding documents that meet the project requirements, and ensure that the document content is aligned with the latest policies and regulations and cutting-edge industry standards, improving the timeliness, accuracy, reliability, and personalization of the generated bidding documents, and further providing a strong basis for the normal progress of bidding activities and technical support for the orderly development of public resource trading.

[0040] It is worth noting that before using the bidding document generation model to generate bidding documents, in order to improve the model performance and optimize its performance in the bidding document generation task in the bidding field, in the embodiments of the present invention, a large amount of historical data in the bidding field can also be applied, that is, the sample project information corresponding to the sample bidding projects, and the sample bidding documents corresponding to the sample bidding projects, to pre-train the bidding document generation model. Among them, the sample project information corresponding to the sample bidding projects, and the sample bidding documents corresponding to the sample bidding projects can be obtained from public resource trading platforms, filing documents of various industry competent departments, etc.

[0041] Specifically, the training process of the bidding document generation model may include: First, an initial model in the training process is constructed on the basis of a large model. Here, the large model can be directly used as the initial model, or other networks, algorithms, etc. can be built on the basis of the large model to construct a model dedicated to the generation of bidding documents in the bidding field, that is, the initial model. Then, the initial model can be trained by using the sample project information corresponding to the sample bidding projects in the bidding field collected in advance, and the sample bidding documents corresponding to the sample bidding projects, so as to obtain a bidding document generation model for the bidding field. The bidding document generation model trained in this way not only inherits the powerful text understanding, semantic analysis and other capabilities of the original large model, but also has excellent clause analysis, policy interpretation, rule reasoning, risk warning and other capabilities. Based on this, it can quickly and accurately generate bidding documents that meet the project requirements and are adapted to laws, regulations, industry standards, etc., realizing efficient, accurate and reliable bidding document generation, and at the same time improving the consistency and scope of application of bidding document generation.

[0042] The tender document generation method provided by the present invention generates tender documents based on the project information of the tender project and the tender knowledge base, and applies the tender document generation model to generate tender documents corresponding to the tender project. The tender knowledge base is dynamically updated, overcoming the defects of low efficiency, error-prone, lack of unified standards, poor flexibility and adaptability, lagging knowledge update and insufficient personalization in the traditional manual writing method and the semi-automated generation method based on templates. By introducing large model technology and a dynamically updated tender knowledge base, tender documents are generated based on project information on this basis, realizing the rapid generation of tender documents, ensuring the accuracy and consistency of the generated tender documents, and at the same time being able to align with the latest policies and regulations and cutting-edge industry standards, deeply associated with project characteristics, and effectively adapted to project requirements. It not only gets rid of the dependence on personal subjective experience, but also can dynamically adapt to policies and regulations and project characteristics, and can be applied to various industries and various complex scenarios, with extremely strong flexibility and a very wide scope of application.

[0043] Based on the above embodiments, the tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; step 120 includes: Construct a tender knowledge graph based on at least one of a policy and regulation library, a tender document library, and an industry standard library; the tender knowledge graph is dynamically updated; Based on the project information and the tender knowledge graph, apply the tender document generation model to generate tender documents corresponding to the tender project.

[0044] Specifically, the process of generating tender documents corresponding to the tender project by applying the tender document generation model based on the project information and the tender knowledge base may specifically include: Considering that industry standards, policies and regulations in the field of bidding are not static, and in order to adapt to the changing policies and regulations and updated industry standards, in the embodiments of the present invention, it is proposed that tender documents can be generated based on the tender knowledge base. However, considering that each update of policies and regulations and industry standards requires the knowledge base to be updated first, and then information retrieval, association, and matching are carried out based on the updated tender knowledge base to find the corresponding knowledge for tender document generation. This process is too cumbersome and inconvenient. In the embodiments of the present invention, it is proposed that a knowledge graph, that is, a tender knowledge graph, can be constructed according to the tender knowledge base. In this way, when the knowledge is updated each time, the corresponding entities and entity relationships (nodes and edges) in the knowledge graph can be directly adjusted based on the updated content, without having to update the knowledge base again and "pick out" the corresponding information from the complex and numerous information. The knowledge retrieval, association, and knowledge update based on the tender knowledge graph will be more convenient and faster.

[0045] Specifically, here, the knowledge graph can be constructed based on the bidding knowledge base first to obtain the knowledge graph in the bidding field, that is, the bidding knowledge graph. The bidding knowledge base here includes any one or more of the policy and regulation library, bidding document library, and industry standard library. Therefore, when constructing the bidding knowledge graph, it can be based on any one or more of the policy and regulation library, bidding document library, and industry standard library, and use knowledge graph construction technologies such as entity recognition and entity relationship extraction to construct the bidding knowledge graph.

[0046] It should be noted that after the bidding knowledge graph is constructed, every time the policies and regulations, industry standards, historical bidding information, etc. in the bidding knowledge base are updated, the bidding knowledge graph will be updated accordingly. However, the update at this time is only based on the specific updated information each time, such as real-time policies and regulations, cutting-edge industry standards, historical bidding information, etc., to adjust the corresponding nodes and the edges associated with the nodes in the knowledge graph. In this way, the rapid update of the knowledge graph can be realized, which helps to generate bidding documents quickly and accurately.

[0047] Immediately afterwards, the bidding documents can be generated based on this bidding knowledge graph. That is, on the basis of the project information, combined with the bidding knowledge graph, the bidding documents are generated through the bidding document generation model to obtain the bidding documents corresponding to the bidding project. Here, specifically, the project information and the bidding knowledge graph can be used as inputs and input into the bidding document generation model, so that the bidding document generation model can understand and analyze the input information and output the bidding documents accordingly; or the project information of the bidding project can be used as the model input and input into the bidding document generation model, so as to automatically analyze and process the project information through the excellent text understanding, semantic analysis and other capabilities of the model, and further combine the bidding knowledge graph, that is, call the bidding knowledge graph, to combine the relevant policies and regulations, industry standards, etc. in the bidding knowledge graph to generate the bidding documents. That is, the model interprets the bidding knowledge graph through its powerful clause analysis, policy interpretation, and rule reasoning capabilities to determine the relevant knowledge from it, and on this basis, combines the understood project information to generate the bidding documents. In this way, the quality of the generated bidding documents can be guaranteed.

[0048] In addition, it is worth noting that with the changing policies and regulations and the updated industry standards, in addition to the bidding knowledge graph, the model can also be updated; that is, the parameters of the bidding document generation model can be updated according to the real-time policies and regulations, cutting-edge industry standards, historical bidding information, etc. to obtain the bidding document generation model with updated parameters, and this model can be used to generate bidding documents. In this way, the timeliness, accuracy, and reliability of the generated bidding documents can be further guaranteed.

[0049] In the embodiments of the present invention, by constructing a dynamically updated tender knowledge graph and generating tender documents accordingly, it is possible to provide a reliable policy and legal basis for the tender document generation process, as well as important technical support, ensuring that the tender document generation process is always based on the latest policies and regulations and cutting-edge industry standards, so that the generated tender documents are more organized, accurate, timely, and reliable.

[0050] The method provided by the embodiments of the present invention combines a dynamically updated tender knowledge graph on the basis of a tender document generation model to generate tender documents. By making full use of the capabilities of the large model, it is possible to quickly understand and accurately analyze project information. In this process, real-time updated policies, regulations, and industry standards are combined to generate tender documents, which can not only ensure the accuracy and reliability of the generated tender documents, but also further improve the timeliness of the tender documents, enabling them to match the latest and most cutting-edge regulatory standards, adapting to the changing tender environment, meeting the diverse needs of different projects, and improving the standardization and efficiency of the bidding and tendering activities, providing reliable technical support for public resource trading.

[0051] Based on the above embodiments, a tender knowledge graph is constructed based on at least one of a policy and regulation library, a tender document library, and an industry standard library, including: Performing entity recognition and relationship extraction on the policy clause information and legal regulation information in the policy and regulation library to obtain policy and regulation triples; Performing entity recognition and relationship extraction on the historical project information and historical tender documents in the tender document library to obtain tender document triples; Performing entity recognition and relationship extraction on the industry technical information and format standard information in the industry standard library to obtain industry standard triples; Constructing a tender knowledge graph based on at least one of the policy and regulation triples, tender document triples, and industry standard triples.

[0052] Specifically, constructing the tender knowledge graph according to the tender knowledge base as above may specifically include: First, the information in the tender knowledge base can be processed to obtain the entity triples necessary for constructing the knowledge graph, and the entity triples contain entities and the entity relationships between them. Therefore, when obtaining the entity triples, entity recognition technology can be used first to process the information in the tender knowledge base to obtain each entity.

[0053] Here, specifically, entity recognition can be performed on the policies and regulations in the policy and regulation library, that is, policy clause information and laws and regulations information, to extract entities therefrom, such as policy entities (such as systems, rules, regulations, etc.), legal entities (such as codes, laws, etc.); entity recognition can be performed on the historical tender information in the tender document library, that is, historical project information and historical tender documents, to extract entities therefrom, such as demand entities, document entities, etc.; entity recognition can be performed on the industry standards in the industry standard library, that is, industry technical information and format standard information, to extract entities therefrom, such as technical entities, format entities, etc.

[0054] Next, based on the entities identified in the previous step, the relationships between the entities can be extracted to obtain entity relationships. Here, specifically, the entities extracted from the policy clause information and laws and regulations information can be used to perform entity relationship extraction in the policy and regulation library to extract the relationships between entities, so as to obtain the entity relationships in the policy and regulation library, such as the relationships between legal entities, the relationships between policy entities, the relationships between legal entities and policy entities, etc.; the entities extracted from the historical project information and historical tender documents can be used to perform entity relationship extraction in the tender document library to extract the relationships between entities, so as to obtain the entity relationships in the tender document library, such as the relationships between demand entities and document entities, the relationships between document entities, the relationships between demand entities, etc.; the entities extracted from the industry technical information and format standard information can be used to perform entity relationship extraction in the industry standard library to extract the relationships between entities, so as to obtain the entity relationships in the tender document library, such as the relationships between technical entities, the relationships between format entities, the relationships between technical entities and format entities, etc.

[0055] After that, based on the entities and entity relationships extracted from the policy and regulation library, the tender document library, and the industry standard library, the corresponding entity triples can be constructed to obtain the policy and regulation triples, tender document triples, and industry standard triples corresponding to the policy and regulation library, the tender document library, and the industry standard library; further, based on any one or more of these policy and regulation triples, tender document triples, and industry standard triples, a knowledge graph in the bidding field, that is, a tender knowledge graph, can be constructed.

[0056] Based on the above embodiments, in step 120, based on the project information and the tender knowledge base, a tender document generation model is applied to generate a tender document for the tender project. After that, it further includes: Extract the key information from the tender document to obtain the tender key information in the tender document; Based on the key tender information, use the tender document generation model to conduct a review of the tender document and obtain the review result of the tender document; Among them, the review of the tender document includes at least one of document compliance check, content risk prediction, and clause conflict detection.

[0057] Specifically, after generating the tender document for a tender project, to ensure the accuracy of the generated tender document, in the embodiments of the present invention, it can also be reviewed to verify whether it complies with the regulations, whether there are conflicts, whether there are risks, etc., so as to provide guarantee for the normal and orderly progress of subsequent tendering and bidding activities.

[0058] Figure 2 is one of the example diagrams of the tender document provided by the present invention, Figure 3 is the second example diagram of the tender document provided by the present invention. See Figure 2 and Figure 3 It can be seen that the generated tender document is usually relatively complex and has a large amount of content. To improve the review efficiency of the tender document, before reviewing the tender document, the key information of the tender document can be extracted first to extract the key points in the document content of the tender document. For example, important contract clause parts, bid evaluation methods parts, etc., so as to obtain the key information in the tender document, that is, the key tender information. Here, the scope of the key points and key information can be set according to needs.

[0059] After determining the key tender information, in the embodiments of the present invention, the tender document can be reviewed based on this key tender information by using large model technology, so as to accurately identify the non-compliance and abnormalities in the tender document through the excellent text understanding, semantic analysis, logical reasoning and other capabilities of the large model, and give the review result, which can provide a basis for the subsequent modification by tendering personnel, and thus provide key support for the orderly progress of public resource trading.

[0060] Specifically, here, the key tender information can be used as input and input into the tender document generation model, so that the tender document generation model can conduct a review of the tender document based on this. The review of the tender document here can include any one or more of document compliance inspection, content risk prediction, and clause conflict detection. That is, after the key tender information is input into the model, the model can understand and analyze the input information to determine whether there are any non-compliances or abnormalities in the tender document. Specifically, it can detect whether there are non-compliant contents (such as grammar errors, semantic incoherence, incompleteness, sensitive words, writing styles not meeting the specifications and requirements of the tender document, etc.), whether there are risky contents such as discrimination and exclusivity, and whether the contract clauses are contradictory or in conflict in the evaluation methods, contract clauses, etc., so as to obtain the review result of the tender document. The review result here can be passed or not passed, corresponding to whether there are problems in the tender document. When it is confirmed that the tender document review fails, the review result can also include the problem type of the specific problems existing, as well as modification suggestions for the specific problems, such as clause conflicts and risky contents.

[0061] Compared with the current manual review and the review methods based on traditional algorithms and traditional machine learning models, in the embodiments of the present invention, the review of the tender document based on the tender document generation model can deeply analyze the tender document, deeply understand its semantics, parse its structure, and identify grammar, so as to quickly and accurately judge the non-compliances and abnormalities in the tender document, obtain a comprehensive and accurate review result, and thus provide a basis for the subsequent modification by the tender staff in the case of a failed review, and further provide more reliable technical support for the steady and orderly progress of public resource trading.

[0062] The method provided by the embodiments of the present invention extracts the key tender information of the tender document, and based on this key tender information, applies the tender document generation model to conduct a review of the tender document to obtain the review result of the tender document, avoiding the problems of long time consumption and low efficiency of traditional manual review, as well as poor flexibility and suboptimal accuracy of the review method based on traditional machine learning models. By using the large model to conduct the review and detection of the tender document, it can not only get rid of the dependence on personal subjective experience, achieve comprehensive, efficient, and accurate review, improve the standardization and efficiency of the bidding and tendering activities, but also be applicable to various industries and complex scenarios, with extremely strong flexibility and a very wide scope of application.

[0063] Based on the above embodiments, in step 120, based on the key tender information, applying the tender document generation model to conduct a review of the tender document to obtain the review result of the tender document, and then it further includes: In the case where the review result indicates that the tender document review fails, based on the review result, determine the problem information from the tender document; Display the problem information and the review results; Among them, the review results include the problem types corresponding to the problem information and the modification suggestions.

[0064] Specifically, after generating the tender documents through the tender document generation and obtaining the review results of the tender documents, in order to ensure the compliance of the tender documents and promote the normal progress of the bidding activities, in the embodiments of the present invention, the review results can also be displayed so that the user (tendering personnel) can know the review situation of the tender documents and make timely modifications and adjustments in case of problems to make them compliant and problem-free.

[0065] To further urge the user to make modifications and make the modifications more targeted and effective, in the embodiments of the present invention, when the review results are obtained and it is confirmed that there are problems with the tender documents, the content of the tender documents can also be marked with problems to mark the problematic places, so as to obtain the problem information in the tender documents. That is, according to the instructions of the review results, the places with marked problems can be located in the tender documents and the information at that place can be extracted as the problem information. For example, when the review result includes that "the first clause in the contract clause part of the tender document has risks", at this time, the first clause in the contract clause part of the tender document can be marked and the information of this clause can be classified into the problem information.

[0066] After that, the problem information and the review results can be displayed so that the user can not only clearly know the review situation of the tender documents, but also clarify the specific problem content, as well as understand the problem types and the specific modification directions (such as grammar modification, semantic supplement, clause amendment, etc.), so as to make modifications faster, more effectively and more accurately, improve the effectiveness and efficiency of the modifications, and further promote the normal progress of the bidding activities.

[0067] Based on the above embodiments, in step 120, after displaying the problem information and the review results, it further includes: Receiving a modification instruction fed back by the user based on the displayed problem information and review results, and updating the tender documents based on the modification instruction to obtain the updated tender documents; Or, receiving the updated tender documents fed back by the user based on the displayed problem information and review results; Storing the updated tender documents, or transmitting the updated tender documents to the bidding platform.

[0068] Specifically, after the above display of the question information and review results, the user can directly understand the review situation of the tender document, clarify the specific problem content therein, and know the problem types and specific modification suggestions, etc. After knowing the above information, the user can then make targeted modifications to the problematic parts in the tender document. Such modifications can be made online or offline, that is, the user can modify the content of the tender document that has been reviewed online or offline, and then upload the modified document.

[0069] In detail, when it is an online modification, at this time, the modification instruction of the user can be received, and based on this modification instruction, for example, the specific characters, paragraphs, etc. modified by the user, the document content of the tender document can be updated to obtain the updated tender document. When it is an offline modification, after the user has completed the modification and re-uploaded, at this time, the tender document re-uploaded by the user can be received, that is, the document updated offline, that is, the updated tender document.

[0070] After that, the updated tender document, whether online or offline, can be directly stored; that is, the document can be output, for example, the updated tender document can be output in a standard format such as PDF (Portable Document Format), PPT (PowerPoint), doc (DOCument), etc. and stored. Or, it can be directly connected to the tender platform, that is, the updated tender document can be directly uploaded to the tender platform to realize the online release of the tender document.

[0071] Here, it is worth noting that for the updated tender document, in order to further determine its accuracy and standardization, in the embodiments of the present invention, it can also be reviewed to verify whether there are any problems, and its reliability can be further ensured through re-inspection.

[0072] In the embodiments of the present invention, in the field of tendering and bidding with an extremely low error tolerance rate, by means of human-machine collaboration for tender document review and modification, the review ability of the large model and the professional judgment ability of humans can be fully exerted, so as to ensure the accuracy of the tender document content, greatly improve the accuracy, coherence and standardization of the tender document, and optimize the final quality of the document.

[0073] Based on the above embodiments, step 120 includes: Extract key information from the project information to obtain project key information; Determine the generation of prompt text based on the project key information; Based on the generated prompt text and the tender knowledge base, apply the tender document generation model to generate a tender document for the tender project to obtain the tender document corresponding to the tender project.

[0074] Specifically, considering that the model has learned a large number of training samples and language rules during training, it usually has relatively strict requirements for the input information, which need to conform to the language forms, norms, etc. that it can receive and understand. Only in this way can the model understand and analyze, and then output the corresponding tender documents. In the embodiments of the present invention, before applying the tender document generation model to generate tender documents, the information input into the model also needs to be converted into a form that the model can understand, that is, normalized, so as to extract it into structured information, and then obtain a prompt text for prompting the model to generate tender documents, that is, generate a prompt text.

[0075] In detail, due to considerations of the complexity of the user's daily language description, the redundancy of the obtained information, and the generation efficiency, in the embodiments of the present invention, before generating tender documents, the key information of the tender project can be extracted first to extract the key content therein. For example, the scope of tender, budget amount, technical requirements, evaluation methods, project name, etc., so as to obtain the project key information of the tender project.

[0076] After obtaining the project key information, in the embodiments of the present invention, this project key information can be converted into a form that the model can receive and understand, that is, a generation prompt text is formed. Here, specifically, the project key information can be processed to extract it into structured text to obtain the generation prompt text. The generation prompt text here is essentially a kind of data carrying the project key information. For example, it can be a prompt, which is data that the model can receive and understand. After inputting it into the model, the model can automatically generate tender documents according to the input information.

[0077] After that, the generated prompt text can be input into the tender document generation model, so that the tender document generation model can automatically generate tender documents according to the input information and the tender knowledge base, thereby obtaining the tender documents corresponding to the tender project, realizing fast and accurate tender document generation, and ensuring the reliability and consistency of the generated tender documents, as well as aligning with the latest policies, regulations and cutting-edge industry standards, reducing the dependence on professional personnel, getting rid of the influence of subjective factors, providing a strong basis for the normal progress of the bidding activities, and providing key support for the orderly development of public resource trading.

[0078] Specifically, here, the prompt text and the bidding knowledge base can be used as input and input into the bidding document generation model, so that the bidding document generation model can understand and analyze the input information, and generate bidding documents according to specific project requirements, so as to obtain bidding documents that are deeply related to the bidding project and meet the project requirements; or the prompt text can be used as input and input into the bidding document generation model, so that the input information is automatically analyzed and processed through the model's excellent text understanding, semantic analysis and other capabilities, so as to deeply explore the project characteristics, and further combine the relevant policies, regulations, industry standards, etc. in the bidding knowledge base to generate bidding documents, that is, the model interprets the knowledge in the bidding knowledge base through its own powerful clause analysis, policy interpretation, and rule reasoning capabilities to determine relevant knowledge, and on this basis, generates bidding documents in combination with the understood project information, thereby ensuring the quality of the generated bidding documents.

[0079] The bidding document generating device provided by the present invention is described below. The bidding document generating device described below and the bidding document generating method described above can be referred to each other.

[0080] Figure 4 is a schematic diagram of the structure of the bidding document generating device provided by the present invention, such as Figure 4 As shown, the device comprises: An acquisition unit 410 is used to acquire project information of a bidding project; A generating unit 420 is used to generate a bidding document by applying a bidding document generation model based on the project information and a bidding knowledge base to obtain a bidding document corresponding to the bidding project; The bidding document generation model is obtained by training on the basis of a large language model using sample project information corresponding to sample bidding projects and sample bidding documents corresponding to the sample bidding projects; the bidding knowledge base is dynamically updated.

[0081] The tender document generation device provided by the present invention generates tender documents based on the project information of the tender project and the tender knowledge base, and applies the tender document generation model to generate tender documents corresponding to the tender project. The tender knowledge base is dynamically updated, overcoming the defects of low efficiency, error-proneness, lack of unified standards, poor flexibility and adaptability, lagging knowledge update, and insufficient personalization in the traditional manual writing method and the semi-automated generation method based on templates. By introducing large model technology and a dynamically updated tender knowledge base, tender documents are generated based on the project information on this basis, achieving the rapid generation of tender documents, ensuring the accuracy and consistency of the generated tender documents, and at the same time being able to align with the latest policies, regulations and cutting-edge industry standards, deeply correlate with project characteristics, and be effectively adapted to project requirements. It not only gets rid of the dependence on personal subjective experience, but also can dynamically adapt to policies, regulations and project characteristics, and can be applied to various industries and various complex scenarios, with extremely strong flexibility and a very wide scope of application.

[0082] Based on the above embodiments, the tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; The generation unit 420 is used for: Based on at least one of the policy and regulation library, the tender document library, and the industry standard library, construct a tender knowledge graph; the tender knowledge graph is dynamically updated; Based on the project information and the tender knowledge graph, apply the tender document generation model to generate tender documents corresponding to the tender project.

[0083] Based on the above embodiments, the generation unit 420 is used for: Perform entity recognition and relationship extraction based on the policy clause information and legal regulation information in the policy and regulation library to obtain policy and regulation triples; Perform entity recognition and relationship extraction based on the historical project information and historical tender documents in the tender document library to obtain tender document triples; Perform entity recognition and relationship extraction based on the industry technical information and format standard information in the industry standard library to obtain industry standard triples; Construct the tender knowledge graph based on at least one of the policy and regulation triples, the tender document triples, and the industry standard triples.

[0084] Based on the above embodiments, the device further includes an audit unit, which is used for: Extract key information from the tender document to obtain the tender key information in the tender document; Based on the tender key information, apply the tender document generation model to audit the tender document to obtain the audit result of the tender document; Among them, the review of the tender documents includes at least one of document compliance check, content risk prediction, and clause conflict detection.

[0085] Based on the above embodiments, the device further includes a display unit for: In the case where the review result indicates that the review of the tender documents fails, based on the review result, determine problem information from the tender documents; Display the problem information and the review result; Among them, the review result includes the problem type and modification suggestions corresponding to the problem information.

[0086] Based on the above embodiments, the device further includes an update unit for: Receive a modification instruction fed back by the user based on the displayed problem information and the review result, and update the tender documents based on the modification instruction to obtain the updated tender documents; Alternatively, receive the updated tender documents fed back by the user based on the displayed problem information and the review result; Store the updated tender documents, or transmit the updated tender documents to the tender platform.

[0087] Based on the above embodiments, the generation unit 420 is used for: Extract key information from the project information to obtain project key information; Based on the project key information, determine and generate prompt text; Based on the generated prompt text and the tender knowledge base, apply the tender document generation model to generate tender documents, and obtain the tender documents corresponding to the tender project.

[0088] The present invention also provides a tender document generation system, Figure 5 which is a schematic structural diagram of the tender document generation system provided by the present invention, as Figure 5 shown, a tender document generation model and a tender knowledge base are deployed on the system; the tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; The tender document generation model is used to construct a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library, and generate tender documents based on the project information of the tender project and the tender knowledge graph, and obtain the tender documents corresponding to the tender project; the tender knowledge graph is dynamically updated; Among them, the tender document generation model is trained on the basis of a large language model by applying the sample project information corresponding to the sample tender project and the sample tender documents corresponding to the sample tender project; the tender knowledge base is dynamically updated.

[0089] The tender document generation system provided by the present invention is deployed with a tender document generation model and a tender knowledge base. By using at least one of the policy and regulation library, tender document library, and industry standard library in the tender knowledge base, a tender knowledge graph is constructed, and tender documents are generated based on the project information of the tender project and the tender knowledge graph to obtain the tender documents corresponding to the tender project. It overcomes the defects of low efficiency, error-prone, lack of unified standards, poor flexibility and adaptability, lagging knowledge update, and insufficient personalization in the traditional manual writing method and the semi-automated generation method based on templates. The introduction of large model technology and a dynamically updated tender knowledge graph for tender document generation can provide a reliable policy and legal basis for the generation of tender documents, as well as important technical support, thus realizing the rapid generation of tender documents and ensuring the accuracy and consistency of the generated tender documents. At the same time, it can be aligned with the latest policies and regulations and cutting-edge industry standards, deeply associated with project characteristics, and effectively adapted to project requirements, making the generated tender documents more organized, accurate, timely, and reliable. It not only gets rid of the dependence on personal subjective experience but also can dynamically adapt to policies, regulations, and project characteristics, and can be applied to various industries and complex scenarios, with extremely strong flexibility and a very wide range of applications.

[0090] Figure 6 An example of the entity structure diagram of an electronic device is shown as Figure 6 As shown, the electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640. Among them, the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 can call the logical instructions in the memory 630 to execute the tender document generation method, which includes: obtaining the project information of the tender project; based on the project information and the tender knowledge base, applying the tender document generation model to generate tender documents to obtain the tender documents corresponding to the tender project; among them, the tender document generation model is trained on the basis of a large language model by applying the sample project information corresponding to the sample tender project and the sample tender documents corresponding to the sample tender project; the tender knowledge base is dynamically updated.

[0091] In addition, when the logical instructions in the above-mentioned memory 630 can be implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0092] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the tender document generation method provided by the above-mentioned various methods. The method includes: obtaining the project information of a tender project; based on the project information and a tender knowledge base, applying a tender document generation model to generate a tender document, and obtaining the tender document corresponding to the tender project; wherein, the tender document generation model is trained based on a large language model, applying the sample project information corresponding to a sample tender project, and the sample tender document corresponding to the sample tender project; and the tender knowledge base is dynamically updated.

[0093] On another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to execute the tender document generation method provided by the above-mentioned various methods. The method includes: obtaining the project information of a tender project; based on the project information and a tender knowledge base, applying a tender document generation model to generate a tender document, and obtaining the tender document corresponding to the tender project; wherein, the tender document generation model is trained based on a large language model, applying the sample project information corresponding to a sample tender project, and the sample tender document corresponding to the sample tender project; and the tender knowledge base is dynamically updated.

[0094] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0095] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0096] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating a tender document, characterized in that Including: Obtain the project information of the tender project; Based on the project information and the tender knowledge base, apply the tender document generation model to generate a tender document, and obtain the tender document corresponding to the tender project; Among them, the tender document generation model is trained based on a large language model, applying the sample project information corresponding to the sample tender project and the sample tender document corresponding to the sample tender project; the tender knowledge base is dynamically updated; The tender knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; the step of applying the tender document generation model to generate a tender document based on the project information and the tender knowledge base to obtain the tender document corresponding to the tender project includes: Construct a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library; the tender knowledge graph is dynamically updated; Based on the project information and the tender knowledge graph, apply the tender document generation model to generate a tender document, and obtain the tender document corresponding to the tender project.

2. The method for generating a tender document according to claim 1, wherein The step of constructing a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library includes: Perform entity recognition and relationship extraction based on the policy clause information and legal regulation information in the policy and regulation library to obtain policy and regulation triples; Perform entity recognition and relationship extraction based on the historical project information and historical tender documents in the tender document library to obtain tender document triples; Perform entity recognition and relationship extraction based on the industry technical information and format standard information in the industry standard library to obtain industry standard triples; Construct the tender knowledge graph based on at least one of the policy and regulation triples, the tender document triples, and the industry standard triples.

3. The method for generating a tender document according to claim 1 or 2, characterized in that, After the step of applying the tender document generation model to generate a tender document based on the project information and the tender knowledge base to obtain the tender document corresponding to the tender project, it further includes: Extract key information from the tender document to obtain the tender key information in the tender document; Based on the tender key information, apply the tender document generation model to conduct a review of the tender document to obtain the review result of the tender document; Among them, the review of the tender document includes at least one of document compliance check, content risk prediction, and clause conflict detection.

4. The method for generating a tender document according to claim 3, wherein After the step of applying the tender document generation model to conduct a review of the tender document based on the tender key information to obtain the review result of the tender document, it further includes: In the case where the review result indicates that the review of the tender document fails, determine problem information from the tender document based on the review result; Display the problem information and the review result; Among them, the review result includes the problem type and modification suggestions corresponding to the problem information.

5. The method for generating a tender document according to claim 4, wherein After the step of displaying the problem information and the review result, it further includes: Receive a modification instruction fed back by the user based on the displayed problem information and the review result, and update the tender document based on the modification instruction to obtain the updated tender document; Alternatively, receive the updated tender document based on the displayed problem information and the audit result feedback from the user; Store the updated tender document or transmit the updated tender document to the tender platform.

6. The method for generating a tender document according to claim 1 or 2, characterized in that, Based on the project information and the tender document knowledge base, use the tender document generation model to generate a tender document, and obtain the tender document corresponding to the tender project, including: Extract key information from the project information to obtain project key information; Determine a generation prompt text based on the project key information; Based on the generation prompt text and the tender document knowledge base, use the tender document generation model to generate a tender document, and obtain the tender document corresponding to the tender project.

7. A tender document generation system, characterized in that, The tender document generation model and the tender document knowledge base are deployed on the system; the tender document knowledge base includes at least one of a policy and regulation library, a tender document library, and an industry standard library; The tender document generation model is used to construct a tender knowledge graph based on at least one of the policy and regulation library, the tender document library, and the industry standard library, and generate a tender document based on the project information of the tender project and the tender knowledge graph, and obtain the tender document corresponding to the tender project; the tender knowledge graph is dynamically updated; Among them, the tender document generation model is trained based on the sample project information corresponding to the sample tender project and the sample tender document corresponding to the sample tender project on the basis of a large language model; the tender document knowledge base is dynamically updated.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the tender document generation method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the tender document generation method according to any one of claims 1 to 6.

Images