Scientific and technological project document processing method and system, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a science and technology project document processing method and system, electronic equipment and a storage medium. The method comprises the steps of obtaining science and technology project documents; inputting the science and technology project document into a pre-established duplicate checking model, and outputting to obtain a duplicate checking rate of the science and technology project document; wherein the duplicate checking model comprises a science and technology knowledge graph and a large language model. According to the embodiment of the invention, the advantages of the large language model and the knowledge graph are combined, the knowledge graph provides accurate and structured fact knowledge, and the large language model has strong language understanding and generating capability. Through combination of the scientific and technological document duplicate checking method and the scientific and technological document duplicate checking system, semantic features of scientific and technological texts are deeply mined, knowledge in the scientific and technological field can be better captured and utilized, the intelligent level and practicability of scientific and technological document duplicate checking are remarkably improved, the accuracy and efficiency of document processing are improved, and an intelligent and efficient solution is provided for a scientific and technological document duplicate checking task; and the quality and efficiency of science and technology project knowledge engineering management are obviously improved.

Description

Technical Field

[0001] The embodiments of the present disclosure belong to the technical field of scientific and technological project document processing, and specifically relate to a scientific and technological project document processing method, system, electronic device and storage medium. Background Art

[0002] In the knowledge engineering management of scientific and technological projects, there are often phenomena such as multiple applications or repeated applications for project approval, and low efficiency in the utilization of results. This has become one of the prominent problems in the field of scientific and technological project management. It will not only cause a waste of the group's scientific and technological resources, but also hinder the self-reliance and self-improvement of scientific and technological innovation and worsen the scientific research and innovation environment.

[0003] Regarding the methods of checking for plagiarism in scientific and technological projects, traditional text checking methods mainly include checking algorithms based on string matching and checking algorithms based on semantic analysis. Checking algorithms based on string matching, such as the Rabin-Karp algorithm and the KMP algorithm, determine whether there is plagiarism or duplication by comparing the similarity of two text strings; this type of algorithm is simple to implement, but cannot effectively identify plagiarism in the form of synonym replacement, word order adjustment, and other transformations. Checking algorithms based on semantic analysis, such as latent semantic analysis (LSA) and topic models, determine similarity by extracting the semantic features of the text; this type of algorithm can overcome the shortcomings of string matching algorithms to a certain extent, but it is still difficult to accurately identify complex semantic changes.

[0004] At present, with the rapid development of artificial intelligence technology, some studies have proposed text duplication detection models based on convolutional neural networks (CNN) and long short-term memory networks (LSTM), which extract deep semantic features of text by building deep neural network structures, thereby improving the accuracy of duplication detection. Other studies have proposed duplication detection methods based on attention mechanisms and pre-trained language models, which use attention mechanisms to capture key information in texts and use language models pre-trained on large-scale corpora to enhance the generalization ability of text representation.

[0005] The above methods have improved the effect of text duplication detection to a certain extent, but there are still some shortcomings. First, the existing deep learning duplication detection models are mainly based on classic network structures such as CNN and LSTM, which have limited modeling capabilities and are difficult to fully mine the semantic information of the text. Second, most existing methods regard the duplication detection task as a binary classification problem, that is, judging whether two documents are plagiarized, while ignoring the quantitative evaluation of the degree of plagiarism. Summary of the invention

[0006] The embodiments of the present disclosure aim to solve at least one of the technical problems existing in the prior art, and provide a scientific and technological project document processing method, system, electronic device and storage medium.

[0007] One aspect of the present disclosure provides a method for processing scientific and technological project documents.

[0008] The law includes:

[0009] Obtain scientific and technological project documents;

[0010] The scientific and technological project document is input into a pre-established duplicate checking model, and the duplicate checking rate of the scientific and technological project document is obtained as an output; wherein the duplicate checking model includes a scientific and technological knowledge graph and a large language model.

[0011] Furthermore, the duplicate checking model is pre-established through the following steps:

[0012] Obtain information in the field of science and technology;

[0013] Construct a scientific and technological knowledge graph based on the information in the scientific and technological field;

[0014] The scientific and technological knowledge graph and the large language model are integrated to obtain the duplicate checking model.

[0015] Furthermore, the acquisition of information in the field of science and technology includes:

[0016] The text data in the field of science and technology are obtained from the Internet through crawler technology; wherein the text data in the field of science and technology includes one or more items of papers, patents, and project reports.

[0017] Furthermore, constructing a scientific and technological knowledge graph based on the scientific and technological field information includes:

[0018] Extract entities and relationships from the text data in the scientific and technological field through natural language processing technology to form a preliminary scientific and technological knowledge graph;

[0019] Verify and supplement the preliminary scientific and technological knowledge graph to obtain a scientific and technological knowledge graph.

[0020] Furthermore, the integration of the scientific and technological knowledge graph and the large language model includes:

[0021] Allocating a shielding probability of each of the entities through a knowledge structure; and / or,

[0022] The input text and knowledge text are encoded and fused separately.

[0023] Furthermore, the technology project document is input into a pre-established duplicate checking model, including

[0024] Retrieving entities and relationships related to the scientific and technological project document from the scientific and technological knowledge graph to obtain a scientific and technological knowledge subgraph;

[0025] The scientific and technological project document and the scientific and technological knowledge subgraph are input into a pre-established duplicate checking model.

[0026] Another aspect of the present disclosure provides a scientific and technological project document processing system, the system comprising:

[0027] Acquisition module, used to obtain scientific and technological project documents;

[0028] The duplicate checking module is used to input the scientific and technological project document into a pre-established duplicate checking model, and output the duplicate checking rate of the scientific and technological project document; wherein the duplicate checking model includes a scientific and technological knowledge graph.

[0029] Furthermore, the system further comprises a model module; the model module is used for:

[0030] Obtain information in the field of science and technology;

[0031] Construct a scientific and technological knowledge graph based on the information in the scientific and technological field;

[0032] The scientific and technological knowledge graph and the large language model are integrated to obtain the duplicate checking model.

[0033] Another aspect of the present disclosure provides an electronic device, comprising:

[0034] at least one processor; and,

[0035] The memory communicatively connected to the at least one processor is used to store one or more programs, and when the one or more programs are executed by the at least one processor, the at least one processor can implement the scientific and technological project document processing method described above.

[0036] Another aspect of the present disclosure provides a computer-readable storage medium storing a computer program, which implements the above-mentioned scientific and technological project document processing method when executed by a processor.

[0037] A method, system, electronic device and storage medium for processing scientific and technological project documents in the disclosed embodiments combine the advantages of large language models (LLM) and knowledge graphs (KG). The knowledge graph provides accurate and structured factual knowledge, while the large language model has powerful language understanding and generation capabilities. Through the combination of the two, the semantic features of scientific and technological texts are deeply excavated, and the knowledge in the field of science and technology can be better captured and utilized, which significantly improves the intelligence level and practicality of scientific and technological document duplicate checking, improves the accuracy and efficiency of document processing, and provides an intelligent and efficient solution for scientific and technological document duplicate checking tasks, which significantly improves the quality and efficiency of scientific and technological project knowledge engineering management. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 A schematic diagram of a process of a scientific and technological project document processing method according to an embodiment of the present disclosure;

[0039] Figure 2 A schematic diagram of a process for establishing a duplicate checking model according to another embodiment of the present disclosure;

[0040] Figure 3 A schematic diagram of the fusion process of the knowledge graph and the large prediction model according to another embodiment of the present disclosure;

[0041] Figure 4 A schematic diagram of the structure of a scientific and technological project document processing system according to another embodiment of the present disclosure;

[0042] Figure 5 The figure is a schematic diagram of the structure of an electronic device according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present disclosure.

[0044] In addition, the described features, structures or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the present disclosure.

[0045] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0046] It should be understood that although the terms first, second, third, etc. may be used in this disclosure to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another component. Therefore, the first component discussed below can be referred to as the second component without departing from the teachings of the concepts of the present disclosure. As used in this disclosure, the term "and / or" includes any one of the associated listed items and all combinations of one or more.

[0047] Those skilled in the art will appreciate that the drawings are merely schematic diagrams of example embodiments, and the modules or processes in the drawings are not necessarily necessary for implementing the present disclosure, and therefore cannot be used to limit the protection scope of the present disclosure.

[0048] like Figure 1 As shown, an embodiment of the present disclosure provides a method for processing scientific and technological project documents, the method comprising:

[0049] Step S1, obtaining scientific and technological project documents.

[0050] Specifically, the scientific and technological project documents of this embodiment may include scientific and technological project applications, papers, patents, etc. to be checked for plagiarism.

[0051] Step S2: input the scientific and technological project document into a pre-established duplicate checking model, and output the duplicate checking rate of the scientific and technological project document; wherein the duplicate checking model includes a scientific and technological knowledge graph and a large language model.

[0052] Specifically, Figure 2 As shown, the duplicate checking model of this embodiment is pre-established through the following steps:

[0053] Step S21, obtaining information in the field of science and technology.

[0054] Specifically, crawler technology is used to obtain text data in the field of science and technology from the Internet, such as papers, patents, project reports, etc., and integrate information from different sources, including human experts, encyclopedias, and databases.

[0055] Step S22: construct a scientific and technological knowledge graph based on the scientific and technological field information.

[0056] Specifically, through natural language processing techniques such as named entity recognition and relationship extraction, entities and relationships are extracted from text data to construct a preliminary knowledge graph. The knowledge graph stores a large number of facts in the form of triples (i.e., head entity, relationship, tail entity), provides accurate explicit knowledge, and has symbolic reasoning capabilities to produce interpretable results. Then, through manual verification and supplementation, the knowledge graph is further improved and expanded to form a high-quality knowledge base in the field of science and technology.

[0057] Step S23, integrating the scientific and technological knowledge graph and the large language model to obtain the duplicate checking model.

[0058] Specifically, in the pre-training stage, the Knowledge Graph (KG) information is injected into the Large Language Model (LLM) in a variety of ways. One method is to expose more knowledge entities in the pre-training target, and to assign masking probabilities by utilizing the knowledge structure, giving important entities a higher masking probability. Another method is to introduce an additional fusion module to encode the input sentence and the text of the knowledge separately and fuse them. In addition, trainable multi-layer perception can be added through adapters to achieve the fusion of language and factual knowledge. The fusion process diagram of the knowledge graph and the large prediction model is shown in the figure. Figure 3 As shown in the figure, knowledge graphs are integrated into the pre-training and reasoning stages of large language models to enhance the model's understanding and application of professional knowledge in scientific and technological literature. By introducing structured information of knowledge graphs during model training, the model can use entity relationships in the knowledge graph to improve its ability to grasp professional terms, concept associations, and domain knowledge, thus solving the shortcomings of traditional large language models in processing knowledge in specific fields.

[0059] Using a large language model as a sequence-to-sequence generator, we can directly generate the query triples to the tail entity text in scientific and technological document processing. This method uses the relationship-guided demonstration technology and the entity-aware hierarchical decoding strategy to not only improve the generation efficiency, but also ensure the accuracy and domain relevance of the generated content. It is particularly suitable for knowledge graph completion tasks in scientific and technological documents.

[0060] Exemplarily, in order to cope with changes and updates in knowledge, a method of dynamically injecting knowledge into the reasoning process can also be adopted. By using a dual-tower architecture, one module processes text input and the other module processes related knowledge graph input. On this basis, an interactive mechanism between text and knowledge is introduced, first encoding the input knowledge graph and then enhancing the input text representation. In order to cope with the dynamic changes of the knowledge graph, a dynamic knowledge fusion strategy is adopted in the reasoning stage. For each input scientific and technological document, the entities and relationships related to it are retrieved from the knowledge graph to form a subgraph. Then, the subgraph is input into the model together with the document content, and the interactive fusion of text and knowledge is realized through methods such as the bidirectional attention mechanism. In this way, the model can use the latest document-related knowledge for reasoning and prediction.

[0061] The above-mentioned set of dynamic knowledge fusion algorithms in this embodiment enables the large language model to absorb the latest scientific and technological literature information in real time, automatically update the internal knowledge base, and ensure the sensitivity of the model to the latest scientific research progress. This mechanism effectively solves the challenges of knowledge graph construction and maintenance, allowing the model to continue to evolve and adapt to the rapid development of the science and technology field.

[0062] The enhanced large-scale language model established through the above steps, i.e., the duplicate checking model described in this embodiment, forms a two-way reasoning system, in which the large-scale language model and the knowledge graph support each other and jointly promote the understanding and generation of scientific and technological documents. On the one hand, the large-scale language model uses its powerful language understanding and generation capabilities to assist the knowledge graph in inferring new facts and representing unknown knowledge; on the other hand, the knowledge graph provides solid entity and relationship support for the large-scale language model, enhances the reasoning ability and interpretability of the model, and ensures that the processed scientific and technological documents are both accurate and rich in deep information. It is applied to tasks such as knowledge duplicate checking of scientific and technological projects and transformation of scientific and technological innovation results. In the submission and review process of scientific and technological documents such as project applications, the duplicate checking model is used to quickly compare whether the documents meet the requirements and improve the efficiency of duplicate checking.

[0063] A method for processing scientific and technological project documents in an embodiment of the present disclosure combines the advantages of a large language model (LLM) and a knowledge graph (KG). The knowledge graph provides accurate and structured factual knowledge, while the large language model has powerful language understanding and generation capabilities. Through the combination of the two, the semantic features of scientific and technological texts are deeply excavated, and the knowledge in the field of science and technology can be better captured and utilized, which significantly improves the intelligence level and practicality of scientific and technological document duplicate checking, improves the accuracy and efficiency of document processing, and provides an intelligent and efficient solution for scientific and technological document duplicate checking tasks, significantly improving the quality and efficiency of scientific and technological project knowledge engineering management. At the same time, the mechanism of dynamic knowledge injection and reasoning enables the model to adapt to constantly changing and updated knowledge, improving the generalization ability and practical application value of the model.

[0064] like Figure 4 Another embodiment of the present disclosure provides a scientific and technological project document processing system, the system comprising:

[0065] The acquisition module 410 is used to acquire scientific and technological project documents;

[0066] The duplicate checking module 420 is used to input the scientific and technological project document into a pre-established duplicate checking model, and output the duplicate checking rate of the scientific and technological project document; wherein the duplicate checking model includes a scientific and technological knowledge graph.

[0067] Exemplarily, the system further includes a model module 430; the model module 430 is used to:

[0068] Obtain information in the field of science and technology;

[0069] Construct a scientific and technological knowledge graph based on the information in the scientific and technological field;

[0070] The scientific and technological knowledge graph and the large language model are integrated to obtain the duplicate checking model.

[0071] Specifically, a scientific and technological project document processing system of an embodiment of the present disclosure is used to implement the scientific and technological project document processing method described in the above embodiment. The specific implementation process has been described in detail in the previous embodiment and will not be repeated here.

[0072] like Figure 5 As shown, another embodiment of the present disclosure provides an electronic device, including:

[0073] At least one processor 501; and a memory 502 communicatively connected to the at least one processor 501, for storing one or more programs, which, when executed by the at least one processor 501, enable the at least one processor 501 to implement the scientific and technological project document processing method described above.

[0074] The memory 502 and the processor 501 are connected in a bus manner, and the bus may include any number of interconnected buses and bridges, and the bus connects various circuits of one or more processors 501 and the memory 502 together. The bus can also connect various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be one element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices on a transmission medium. The data processed by the processor 501 is transmitted on a wireless medium via an antenna, and further, the antenna also receives data and transmits the data to the processor 501.

[0075] The processor 501 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management and other control functions. The memory 502 can be used to store data used by the processor 501 when performing operations.

[0076] Yet another embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, wherein the computer program implements the above-mentioned scientific and technological project document processing method when executed by a processor.

[0077] The computer-readable storage medium may be included in the system or electronic device of the present disclosure, or may exist independently.

[0078] Computer-readable storage media may be any tangible media that contains or stores a program, which may be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device. More specific examples include, but are not limited to, an electrical connection with one or more wires, a portable computer disk, a hard disk, optical fiber, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0079] The computer-readable storage medium may also include a data signal propagated in baseband or as part of a carrier wave, in which the computer-readable program code is carried. Specific examples include but are not limited to electromagnetic signals, optical signals, or any suitable combination thereof.

[0080] It is to be understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of the present disclosure, but the present disclosure is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and substance of the present disclosure, and these modifications and improvements are also considered to be within the scope of protection of the present disclosure.

Claims

1. A method for processing scientific and technological project documents, characterized in that: The method comprises: Obtain scientific and technological project documents; The scientific and technological project document is input into a pre-established duplicate checking model, and the duplicate checking rate of the scientific and technological project document is obtained as an output; wherein the duplicate checking model includes a scientific and technological knowledge graph and a large language model.

2. The method according to claim 1, characterized in that The duplicate checking model is pre-established through the following steps: Obtain information in the field of science and technology; Construct a scientific and technological knowledge graph based on the information in the scientific and technological field; The scientific and technological knowledge graph and the large language model are integrated to obtain the duplicate checking model.

3. The method according to claim 2, characterized in that The acquisition of information in the field of science and technology includes: The text data in the field of science and technology is obtained from the Internet through crawler technology; wherein the text data in the field of science and technology includes one or more items of papers, patents, and project reports.

4. The method according to claim 3, characterized in that The constructing of a scientific and technological knowledge graph according to the scientific and technological field information includes: Extract entities and relationships from the text data in the scientific and technological field through natural language processing technology to form a preliminary scientific and technological knowledge graph; Verify and supplement the preliminary scientific and technological knowledge graph to obtain a scientific and technological knowledge graph.

5. The method according to claim 4, characterized in that The integration of the scientific and technological knowledge graph and the large language model includes: Allocating a shielding probability of each of the entities through a knowledge structure; and / or, The input text and knowledge text are encoded and fused separately.

6. The method according to any one of claims 1 to 5, characterized in that: The step of inputting the scientific and technological project document into a pre-established duplicate checking model includes: Retrieving entities and relationships related to the scientific and technological project document from the scientific and technological knowledge graph to obtain a scientific and technological knowledge subgraph; The scientific and technological project document and the scientific and technological knowledge subgraph are input into a pre-established duplicate checking model.

7. A scientific and technological project document processing system, characterized in that: The system comprises: Acquisition module, used to obtain scientific and technological project documents; The duplicate checking module is used to input the scientific and technological project document into a pre-established duplicate checking model, and output the duplicate checking rate of the scientific and technological project document; wherein the duplicate checking model includes a scientific and technological knowledge graph.

8. The system according to claim 7, characterized in that The system further comprises a model module; the model module is used for: Obtain information in the field of science and technology; Construct a scientific and technological knowledge graph based on the information in the scientific and technological field; The scientific and technological knowledge graph and the large language model are integrated to obtain the duplicate checking model.

9. An electronic device, characterized in that: include: at least one processor; as well as, The memory communicatively connected to the at least one processor is used to store one or more programs, and when the one or more programs are executed by the at least one processor, the at least one processor can implement the scientific and technological project document processing method described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the scientific and technological project document processing method described in any one of claims 1 to 6 is implemented.

Images