Method and system for automatically generating patent document based on large language model
Through methods and systems based on large language models, patent claims and specifications are generated, which solves the problems of high cost of writing patent documents and difficult to guarantee quality, and achieves efficient and low-cost patent document generation.
Patent Information
- Application Number
- CN202510042203.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-10
- Publication Date
- 2025-04-25
AI Technical Summary
The writing of patent documents requires high professional skills, which takes time and costly. Especially in complex technical fields, it is difficult for ordinary users and newcomers to ensure the accuracy of the claims and the details of the instructions.
Using methods and systems based on large language models, we receive patent technical solutions provided by users, generate patent claims and instructions through pre-trained patent large language models, and perform compliance proofreading, content processing and typesetting.
It reduces the labor cost of patent document writing, helps ordinary users and newcomers in the industry to generate high-quality patent documents, especially suitable for patent application scenarios of complex technical solutions.
Smart Images
Figure CN119886071A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of information processing technology, and in particular to a method and system for automatically generating patent documents based on a large language model. Background Art
[0002] With the development of artificial intelligence technology, natural language processing technology has been widely used in various fields, including the generation of patent documents. Patent documents are important documents to protect the rights and interests of inventors, and their quality directly affects the interests of inventors. However, the writing of patent documents requires high professional skills, which is a huge challenge for ordinary users and newcomers in the industry. The writing of patent documents is essentially to convert the description of technical language into the description of rigorous legal documents. The writing process is usually time-consuming and costly. Especially in complex technical fields, it is difficult for ordinary users and newcomers in the industry to ensure the accuracy of the claims and the details of the description. Summary of the invention
[0003] In order to at least to some extent overcome the problem of high manpower cost in writing patent documents in related technologies, the present application provides a method and system for automatically generating patent documents based on a large language model.
[0004] The scheme of this application is as follows:
[0005] According to a first aspect of an embodiment of the present application, a method for automatically generating a patent document based on a large language model is provided, comprising:
[0006] Receive patented technical solutions provided by users;
[0007] Enter the patent technical solution into the pre-trained patent language model to generate patent claims and patent description;
[0008] Conduct compliance check on the generated patent claims and patent specifications;
[0009] Conduct content processing and typesetting on the proofread patent claims and patent specifications;
[0010] Output the processed patent claims and patent specification.
[0011] Preferably, the method further comprises:
[0012] Screening published or authorized patent documents in the patent database as sample data to be processed;
[0013] Performing format cleaning, terminology standardization and noise removal on the sample data to be processed;
[0014] The processed sample data is classified according to the technical field to obtain different types of patent data sets;
[0015] Performing data enhancement on the patent data set, and using the entire patent data set obtained after the data enhancement as a training data set;
[0016] The patent large language model is trained according to the training data set.
[0017] Preferably, data enhancement is performed on the patent dataset, including:
[0018] According to the common template format of patent claims and patent specifications, the patent documents in the patent dataset are reorganized and synthesized according to a fixed structure to generate a new patent dataset;
[0019] The patent datasets in different technical fields are synthesized to generate new patent datasets.
[0020] Preferably, the patent claims are generated, including:
[0021] Analyze the core elements and dependent characteristics of the patent technical solution, determine the number and logical order of the claims, and generate patent claims containing independent claims and dependent claims;
[0022] Fill in the key elements of the independent claim according to the preset independent claim template;
[0023] Check the completeness of the generated independent claims;
[0024] According to the rules for writing dependent claims, provide supplementary explanations for each generated dependent claim;
[0025] Check the consistency of terms in patent claims;
[0026] In the process of generating patent claims, the compliance of the terms in the patent claims is ensured through the built-in term library.
[0027] Preferably, generating a patent specification includes:
[0028] Determine the technical field to which the patent technical solution belongs, call the patent data set in the corresponding technical field, and generate the technical background and technical issues in the patent specification;
[0029] Summarize and screen the technical effects recorded in the patent technical solution, determine the technical effects that meet the technical problem, and generate the invention content and technical effects corresponding to the patent claims based on the patent claims;
[0030] According to the actual implementation results of the patent technical solution, various embodiments in the specific implementation method are generated, and the generated embodiments are sorted by importance;
[0031] Expanding the generated embodiment content by calling the patent data set in the corresponding technical field;
[0032] Generate illustrations based on the contents of the drawings involved in the specification to make the illustrations consistent with the text description;
[0033] Check whether the numbers in the drawings in the specification correspond to the references in the specification;
[0034] Check the consistency of terminology in patent specifications.
[0035] Preferably, the generated patent claims and patent specifications are proofread for compliance, including:
[0036] Compare the generated patent claims and patent specifications with the standard patent format, check whether there are any formal problems with the document symbols, numbering and layout of the generated patent claims and patent specifications, and check whether the use of document symbols complies with the specifications of patent legal documents;
[0037] Perform logic verification on the generated patent claims and patent specifications to determine whether the logic of the patent claims and patent specifications are consistent;
[0038] Perform terminology check on the generated patent claims and patent specifications to determine whether the terms in the patent claims and patent specifications are consistent;
[0039] Adjust the wording of patent claims and patent specifications according to the characteristics of legal language.
[0040] Preferably, the proofread patent claims and patent specification are subjected to content processing and typesetting, including:
[0041] Screen and remove redundant content in patent claims and patent specifications;
[0042] Optimize the sentence structure in patent claims and patent specifications;
[0043] Typeset patent claims and patent specifications based on the standard format of patent documents.
[0044] Preferably, the method further comprises:
[0045] Receiving user feedback on the patent large language model;
[0046] The patent large language model is adjusted according to user feedback information.
[0047] Preferably, the method further comprises:
[0048] The patent large language model is regularly updated and iteratively optimized based on a preset period.
[0049] According to a second aspect of an embodiment of the present application, a system for automatically generating patent documents based on a large language model is provided, comprising:
[0050] A receiving module, used to receive patented technical solutions provided by users;
[0051] A generation module, used to input the patent technical solution into a pre-trained patent large language model to generate patent claims and patent description;
[0052] The proofreading module is used to proofread the generated patent claims and patent specifications for compliance;
[0053] A processing module, used for content processing and typesetting of the proofread patent claims and patent specifications;
[0054] The output module is used to output the processed patent claims and patent specifications.
[0055] The technical solution provided by the present application may include the following beneficial effects: The method for automatically generating patent documents based on a large language model in the present application includes: receiving a patent technical solution provided by a user; entering the patent technical solution into a pre-trained patent large language model to generate patent claims and patent specifications; performing compliance proofreading on the generated patent claims and patent specifications; performing content processing and typesetting on the proofread patent claims and patent specifications; and outputting the processed patent claims and patent specifications.
[0056] In this technical solution, a patent language model for patent document writing is pre-trained. The patent language model can directly generate patent claims and patent specifications based on the patent technical solution provided by the user, and ensure the accuracy of the claims and the detail of the specifications through compliance proofreading, content processing and typesetting processing. It is especially suitable for patent application scenarios with complex technical solutions. This technical solution can help ordinary users directly generate patent documents and reduce the labor cost of patent writing. It can also assist newcomers in the writing work and improve the writing quality of newcomers in the industry.
[0057] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0059] Figure 1 It is a flowchart of a method for automatically generating patent documents based on a large language model provided by an embodiment of the present application;
[0060] Figure 2 It is a flowchart of training a patent big language model in a method for automatically generating patent documents based on a big language model provided by an embodiment of the present application;
[0061] Figure 3 It is a structural diagram of a system for automatically generating patent documents based on a large language model provided by an embodiment of the present application.
[0062] Figure numerals: receiving module-21; generating module-22; checking module-23; processing module-24; output module-25. DETAILED DESCRIPTION
[0063] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.
[0064] Embodiment 1
[0065] Figure 1 is a flowchart of a method for automatically generating patent documents based on a large language model provided by an embodiment of the present application, with reference to Figure 1 , a method for automatically generating patent documents based on a large language model, comprising:
[0066] S11: Receive patented technical solutions provided by users;
[0067] The patent technical solution provided by the user is the "technical briefing document", which includes the user's summary of its technical solution. The patent writing work is to convert the "technical briefing document" provided by the user into a rigorous description of a legal document.
[0068] S12: Input the patent technical solution into the pre-trained patent language model to generate patent claims and patent description;
[0069] Large language models (LLMs) refer to Transformer language models that contain tens of billions (or more) of parameters. These models are trained on large-scale text data, such as GPT-3, PaLM, Galactica, and LLaMA. LLMs have powerful capabilities for understanding natural language and solving complex tasks (through text generation). In this technical solution, the user's patented technical solution is used as the input of the large language model, and the large language model generates the corresponding patent claims and patent specifications.
[0070] S13: Conduct compliance check on the generated patent claims and patent specifications;
[0071] Proofread and generate content to ensure that patent documents comply with legal requirements.
[0072] S14: Content processing and typesetting of the proofread patent claims and patent specification;
[0073] Optimize and format the generated content to ensure that the format of the patent document is compliant and the content is rich.
[0074] S15: Output the processed patent claims and patent specification.
[0075] In this technical solution, a patent language model for patent document writing is pre-trained. The patent language model can directly generate patent claims and patent specifications based on the patent technical solution provided by the user, and ensure the accuracy of the claims and the detail of the specifications through compliance proofreading, content processing and typesetting processing. It is especially suitable for patent application scenarios with complex technical solutions. This technical solution can help ordinary users directly generate patent documents and reduce the labor cost of patent writing. It can also assist newcomers in the writing work and improve the writing quality of newcomers in the industry.
[0076] Embodiment 2
[0077] Figure 2 This is a flow chart of training a patent large language model in a method for automatically generating patent documents based on a large language model provided by an embodiment of the present application, with reference to Figure 2 , the method further comprises:
[0078] S21: Screening published or authorized patent documents in the patent database as sample data to be processed;
[0079] S22: Format cleaning, terminology standardization and denoising of the sample data to be processed;
[0080] S23: classify the processed sample data according to technical fields to obtain different types of patent data sets;
[0081] S24: perform data enhancement on the patent data set, and use the entire patent data set obtained after data enhancement as a training data set;
[0082] S25: Train the patent language model based on the training dataset.
[0083] Publicly available or authorized patent documents can be regarded as high-quality patent documents. In this embodiment, publicly available or authorized patent documents are screened from multiple industry patent databases as sample data to be processed to ensure the quality, authority and diversity of the sample data to be processed.
[0084] The processed sample data is classified by technical field to provide patent data sets specially trained for different types of patent writing. Preferably, in this embodiment, different types of patent data sets are also subdivided based on technical difficulty.
[0085] Format cleaning is to remove redundant symbols, incorrect characters, and irrelevant information to ensure that the data input into the large language model is clean and standardized.
[0086] The purpose of term normalization is to unify terms with similar or identical concepts in different documents, ensuring that large language models can learn consistent word usage habits.
[0087] The purpose of denoising is to remove noise from repeated content and ambiguous descriptions to improve the accuracy of training data.
[0088] It should be noted that data enhancement of the patent dataset includes:
[0089] According to the common template format of patent claims and patent specifications, the patent documents in the patent dataset are reorganized and synthesized according to a fixed structure to generate a new patent dataset;
[0090] The patent datasets in different technical fields are synthesized to generate new patent datasets.
[0091] Data augmentation, also known as data expansion, is a technique to artificially increase the training set by creating a modified copy of the data set using existing data. It is one of the commonly used techniques in deep learning, including making minor changes to the data set or using deep learning to generate new data points. Data augmentation is mainly used to increase the training data set, making the data set as diverse as possible, so that the trained model has stronger generalization ability.
[0092] In this embodiment, there are two data enhancement schemes for patent data sets. One is to reorganize and synthesize the patent documents in the patent data set according to the fixed structure according to the general template format of patent claims and patent specifications to generate a new patent data set and enhance the standardized expression ability of the model. The other is to synthesize the content of patent data sets in different technical fields to generate a new patent data set. By synthesizing different technical solutions and innovative contents, a variety of patent writing samples are generated to ensure that the generation ability of the model covers a variety of technical backgrounds.
[0093] It should be noted that the generation of patent claims includes:
[0094] Analyze the core elements and dependent characteristics of the patent technical solution, determine the number and logical order of the claims, and generate patent claims containing independent claims and dependent claims;
[0095] Fill in the key elements of the independent claim according to the preset independent claim template;
[0096] Check the completeness of the generated independent claims;
[0097] According to the rules for writing dependent claims, provide supplementary explanations for each generated dependent claim;
[0098] Check the consistency of terms in patent claims;
[0099] In the process of generating patent claims, the compliance of the terms in the patent claims is ensured through the built-in term library.
[0100] Patent claim generation is the core of the entire solution, and it is necessary to generate legally effective claims. This technical solution makes full use of the data accumulation of the patent large language model on claim writing specifications to ensure the rigor and accuracy of the generated content.
[0101] The patent large language model trained in this technical solution mainly includes four steps in claim production: claim structure design, independent claim generation, dependent claim generation and terminology compliance check.
[0102] By analyzing the core elements and dependent characteristics of the technical solution, a hierarchical structure containing independent and dependent claims is automatically generated. The number and logical order of the claims are also automatically adjusted according to the complexity and innovation of the technical solution to ensure comprehensive protection.
[0103] According to the preset independent claim template, fill in the key elements of the independent claim to ensure the clarity of the independent claim and the accuracy of the scope of protection.
[0104] Check the completeness of the generated independent claims to ensure that the independent claims fully describe the invention and do not miss any key points.
[0105] According to the rules for writing dependent claims, supplementary explanations are given to each generated dependent claim to further enrich the description of the invention content in a progressive manner.
[0106] In this technical solution, dependent claims are generated according to the detailed requirements of technical characteristics, so that they can supplement the description of secondary or preferred characteristics and enhance the scientificity and professionalism of the document.
[0107] During the generation process of patent claims, the compliance of the terms in the patent claims is ensured through the built-in term base. Finally, the consistency of the terms in the patent claims is checked to ensure that terms that meet industry standards are used during the generation process.
[0108] It should be noted that the generation of patent specifications includes:
[0109] Determine the technical field to which the patent technical solution belongs, call the patent data set in the corresponding technical field, and generate the technical background and technical issues in the patent specification;
[0110] Summarize and screen the technical effects recorded in the patent technical solution, determine the technical effects that meet the technical problem, and generate the invention content and technical effects corresponding to the patent claims based on the patent claims;
[0111] According to the actual implementation results of the patent technical solution, various embodiments in the specific implementation method are generated, and the generated embodiments are sorted by importance;
[0112] Expanding the generated embodiment content by calling the patent data set in the corresponding technical field;
[0113] Generate illustrations based on the contents of the drawings involved in the specification to make the illustrations consistent with the text description;
[0114] Check whether the numbers in the drawings in the specification correspond to the references in the specification;
[0115] Check the consistency of terminology in patent specifications.
[0116] In this technical solution, the technical field to which the patent technical solution belongs is first determined, and then the patent data set of the corresponding technical field is called to generate a description that conforms to the technical background in the patent specification, and automatically analyze and supplement the problems faced by the current technology.
[0117] Based on the patent claims, generate the invention content and technical effects corresponding to the patent claims, the correspondence between the description and the claims, and ensure that the invention content can effectively convey the technological innovation.
[0118] The generated embodiment content is expanded by calling the patent data set of the corresponding technical field to ensure the richness and completeness of the description content.
[0119] Based on the actual implementation results of the patent technical solution, various embodiments in the specific implementation method are generated, and the generated embodiments are sorted by importance to ensure the clarity and repeatability of the description, as well as the readability and scientific nature of the specification.
[0120] Generate illustrations based on the contents of the drawings involved in the specification to make the illustrations consistent with the text description, check whether the numbers in the drawings in the specification correspond to the referenced contents in the specification, and ensure the consistency and accuracy of the document.
[0121] Finally, the terminology consistency in the patent specification is checked to ensure that industry-standard terminology is used during the generation process.
[0122] It should be noted that compliance proofreading of the generated patent claims and patent specifications includes:
[0123] Compare the generated patent claims and patent specifications with the standard patent format, check whether there are any formal problems with the document symbols, numbering and layout of the generated patent claims and patent specifications, and check whether the use of document symbols complies with the specifications of patent legal documents;
[0124] Perform logic verification on the generated patent claims and patent specifications to determine whether the logic of the patent claims and patent specifications are consistent;
[0125] Perform terminology check on the generated patent claims and patent specifications to determine whether the terms in the patent claims and patent specifications are consistent;
[0126] Adjust the wording of patent claims and patent specifications according to the characteristics of legal language.
[0127] In this embodiment, it is first determined whether the terms in the patent claims and the patent specification are consistent, and then it is determined whether the terms used in the patent claims and the patent specification are consistent. This prevents the situation where the terms in the patent claims and the patent specification are consistent, but the terms used in the patent claims and the patent specification are inconsistent.
[0128] Adjust the wording of patent claims and patent specifications according to the characteristics of legal language to make them conform to the legal expressions commonly used in the industry.
[0129] It should be noted that the proofread patent claims and patent specifications are processed for content and layout, including:
[0130] Screen and remove redundant content in patent claims and patent specifications;
[0131] Optimize the sentence structure in patent claims and patent specifications;
[0132] Typeset patent claims and patent specifications based on the standard format of patent documents.
[0133] Screen and remove redundant content from patent claims and patent specifications to ensure concise and informative documents.
[0134] Optimizing the sentence structures in patent claims and patent specifications can make the generated content more in line with natural language expression habits.
[0135] The patent claims and patent specifications are typeset based on the standard format of patent documents to ensure that the overall document is beautiful and compliant. The document content structure and format can also be further adjusted to give the document a good visual effect.
[0136] Embodiment 3
[0137] It should be noted that the method also includes:
[0138] Receive user feedback on the patent large language model;
[0139] Adjust the patent language model based on user feedback.
[0140] This technical solution integrates a feedback mechanism to receive user feedback on the patent large language model, including user opinions and modification suggestions, and uses this as a basis to adjust the patent large language model and further optimize the generated content.
[0141] Preferably, the patent large language model is regularly adjusted based on user feedback to improve the accuracy of the generated effect.
[0142] It should be noted that the method also includes:
[0143] The patent language model is regularly updated and iteratively optimized based on the preset cycle.
[0144] The patent big language model is constantly improving its accuracy and understanding of patent drafting requirements through regular updates and iterative optimization.
[0145] Embodiment 4
[0146] A system for automatically generating patent documents based on a large language model, referring to Figure 3 ,include:
[0147] A receiving module 31 is used to receive a patent technical solution provided by a user;
[0148] A generation module 32 is used to input the patent technical solution into a pre-trained patent large language model to generate patent claims and patent specifications;
[0149] A proofreading module 33, used to proofread the generated patent claims and patent specifications for compliance;
[0150] A processing module 34 is used to perform content processing and typesetting on the proofread patent claims and patent specification;
[0151] The output module 35 is used to output the processed patent claims and patent specification.
[0152] The proofreading module 33 is specifically used to compare the generated patent claims and patent specifications with the standard patent format, check whether there are any formal problems in the document symbols, numbering and layout of the generated patent claims and patent specifications, and check whether the use of document symbols complies with the specifications of patent legal documents; perform logic verification on the generated patent claims and patent specifications to determine whether the logic of the patent claims and patent specifications is consistent; perform terminology verification on the generated patent claims and patent specifications to determine whether the terms in the patent claims and patent specifications are consistent; and adjust the content wording of the patent claims and patent specifications according to the characteristics of the legal language.
[0153] The processing module 34 is specifically used to screen and remove redundant content in the patent claims and patent specifications; optimize the sentence patterns in the patent claims and patent specifications;
[0154] Typeset patent claims and patent specifications based on the standard format of patent documents.
[0155] The system also includes:
[0156] The model training module is used to screen the public or authorized patent documents in the patent database as the sample data to be processed; perform format cleaning, terminology standardization and denoising on the sample data to be processed; classify the processed sample data according to the technical field to obtain different types of patent data sets; perform data enhancement on the patent data set, and use all the patent data sets obtained after data enhancement as the training data set; and train the patent large language model based on the training data set.
[0157] The system also includes:
[0158] The patent claim generation module is used to analyze the core elements and dependent characteristics of the patent technical solution, determine the number and logical order of the claims, and generate patent claims containing independent claims and dependent claims; fill in the key elements of the independent claims according to the preset independent claim template; check the integrity of the generated independent claims; supplement the generated dependent claims according to the writing rules of the dependent claims; check the consistency of the terms in the patent claims; in the process of generating the patent claims, the compliance of the terms in the patent claims is guaranteed by the built-in term library.
[0159] The system also includes:
[0160] The patent specification generation module is used to determine the technical field to which the patent technical solution belongs, call the patent data set of the corresponding technical field, and generate the technical background and technical problems in the patent specification; summarize and screen the technical effects recorded in the patent technical solution, determine the technical effects that meet the technical problems, and generate the invention content and technical effects corresponding to the patent claims based on the patent claims; generate various embodiments in the specific implementation method according to the actual implementation results in the patent technical solution, and sort the generated embodiments by importance; expand the content of the generated embodiment by calling the patent data set of the corresponding technical field; generate illustrations according to the content of the drawings involved in the specification, so that the illustrations are consistent with the text description; check whether the numbers in the drawings in the specification correspond to the referenced content in the specification; and check the consistency of terms in the patent specification.
[0161] The system also includes:
[0162] The update module is used to receive user feedback on the patent large language model; adjust the patent large language model according to the user feedback; and regularly update and iteratively optimize the patent large language model based on a preset period.
[0163] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.
[0164] It should be noted that, in the description of this application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" refers to at least two.
[0165] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.
[0166] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0167] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.
[0168] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0169] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.
[0170] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0171] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.
Claims
1. A method for automatically generating patent documents based on a large language model, characterized in that: include: Receive patented technical solutions provided by users; Enter the patent technical solution into the pre-trained patent language model to generate patent claims and patent description; Conduct compliance check on the generated patent claims and patent specifications; Conduct content processing and typesetting on the proofread patent claims and patent specifications; Output the processed patent claims and patent specification.
2. The method according to claim 1, characterized in that The method further comprises: Screening published or authorized patent documents in the patent database as sample data to be processed; Performing format cleaning, terminology standardization and noise removal on the sample data to be processed; The processed sample data is classified according to the technical field to obtain different types of patent data sets; Performing data enhancement on the patent data set, and using the entire patent data set obtained after the data enhancement as a training data set; The patent large language model is trained according to the training data set.
3. The method according to claim 2, characterized in that Perform data enhancement on the patent dataset, including: According to the common template format of patent claims and patent specifications, the patent documents in the patent dataset are reorganized and synthesized according to a fixed structure to generate a new patent dataset; The patent datasets in different technical fields are synthesized to generate new patent datasets.
4. The method according to claim 1, characterized in that: Generate patent claims, including: Analyze the core elements and dependent characteristics of the patent technical solution, determine the number and logical order of the claims, and generate patent claims containing independent claims and dependent claims; Fill in the key elements of the independent claim according to the preset independent claim template; Check the completeness of the generated independent claims; According to the rules for writing dependent claims, provide supplementary explanations for each generated dependent claim; Check the consistency of terms in patent claims; In the process of generating patent claims, the compliance of the terms in the patent claims is ensured through the built-in term library.
5. The method according to claim 4, characterized in that Generate patent specifications, including: Determine the technical field to which the patent technical solution belongs, call the patent data set in the corresponding technical field, and generate the technical background and technical issues in the patent specification; Summarize and screen the technical effects recorded in the patent technical solution, determine the technical effects that meet the technical problem, and generate the invention content and technical effects corresponding to the patent claims based on the patent claims; According to the actual implementation results of the patent technical solution, various embodiments in the specific implementation method are generated, and the generated embodiments are sorted by importance; Expanding the generated embodiment content by calling the patent data set in the corresponding technical field; Generate illustrations based on the contents of the drawings involved in the specification to make the illustrations consistent with the text description; Check whether the numbers in the drawings in the specification correspond to the references in the specification; Check the consistency of terminology in patent specifications.
6. The method according to claim 1, characterized in that Compliance proofreading of generated patent claims and patent specifications, including: Compare the generated patent claims and patent specifications with the standard patent format, check whether there are any formal problems with the document symbols, numbering and layout of the generated patent claims and patent specifications, and check whether the use of document symbols complies with the specifications of patent legal documents; Perform logic verification on the generated patent claims and patent specifications to determine whether the logic of the patent claims and patent specifications are consistent; Perform terminology check on the generated patent claims and patent specifications to determine whether the terms in the patent claims and patent specifications are consistent; Adjust the wording of patent claims and patent specifications according to the characteristics of legal language.
7. The method according to claim 1, characterized in that Conduct content processing and typesetting on the proofread patent claims and patent specifications, including: Screen and remove redundant content in patent claims and patent specifications; Optimize the sentence structure in patent claims and patent specifications; Typeset patent claims and patent specifications based on the standard format of patent documents.
8. The method according to claim 1, characterized in that The method further comprises: Receiving user feedback on the patent large language model; The patent large language model is adjusted according to user feedback information.
9. The method according to claim 1, characterized in that: The method further comprises: The patent large language model is regularly updated and iteratively optimized based on a preset period.
10. A system for automatically generating patent documents based on a large language model, characterized in that: include: A receiving module, used to receive patented technical solutions provided by users; A generation module, used to input the patent technical solution into a pre-trained patent large language model to generate patent claims and patent description; The proofreading module is used to proofread the generated patent claims and patent specifications for compliance; A processing module, used for content processing and typesetting of the proofread patent claims and patent specifications; The output module is used to output the processed patent claims and patent specifications.