Method and device for evaluating and analyzing quality of patent technology disclosure book
By using language models to evaluate and complete the technical briefing, the problem of inefficiency in the writing and review of patent technical briefings is solved, and a faster and more efficient patent application process is achieved.
Patent Information
- Application Number
- CN202510042218.1
- 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
In the prior art, the writing and review process of patent technical briefing is inefficient, resulting in low communication and modification efficiency, which may lead to the rejection of patent applications.
By evaluating the technical briefing book with the first pre-trained language model, identifying the missing necessary elements, and using the pre-set second language model to generate and complete these elements to form a complete target technical briefing book.
It greatly shortens the review time of technical briefing, improves work efficiency, reduces the repeated communication and modification process between applicants and patent agents, and speeds up the patent application process.
Smart Images
Figure CN119886113A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field related to text analysis, and specifically to a method and device for evaluating and analyzing the quality of patent technology briefing documents. Background Art
[0002] In the context of the rapid development of modern science and technology, patent applications have become an important means for enterprises and individuals to protect their innovative achievements. However, the patent application process is complex and demanding, especially the writing of technical disclosure documents, which requires detailed descriptions of the technical background, technical solutions, technical effects, and drawings of the invention. For non-professional patent applicants, understanding and writing a technical disclosure document that meets the requirements is a very challenging task.
[0003] In the traditional patent application process, applicants usually need to communicate with patent agents several times to ensure the completeness and accuracy of the technical briefing document. However, this process is not only time-consuming and inefficient, but may also result in the rejection of the application by the patent office due to inaccurate or incomplete information transmission. In addition, patent agents often need to invest a lot of time and energy in reviewing and revising the technical briefing document to ensure that the document meets the requirements of patent law. Summary of the invention
[0004] In view of this, the embodiments of the present application are committed to providing a method and device for evaluating and analyzing the quality of patent technology briefing documents, so as to solve the problem of low efficiency in communication and modification between inventors and agents in the prior art.
[0005] This application provides a method for evaluating and analyzing the quality of patent technology briefing documents, including:
[0006] Obtain technical briefing documents;
[0007] Input the technical briefing document into the pre-trained first language model to obtain a technical briefing document evaluation result;
[0008] The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result;
[0009] Based on the evaluation results, the missing necessary elements in the technical briefing document are generated and completed through the preset second language model to obtain the target technical briefing document.
[0010] In some embodiments, the generation and completion of the missing necessary elements in the technical briefing book includes:
[0011] Initial generation and completion of the missing necessary elements in the technical disclosure document to obtain a complete technical disclosure document;
[0012] Send the complete technical briefing to the inventor for review and obtain the review results;
[0013] If the audit result indicates that the audit is passed, the complete technical briefing document will be determined as the target briefing document.
[0014] In some embodiments, it also includes:
[0015] Obtaining a modified document provided by the inventor; wherein the modified document is obtained by the inventor modifying the complete technical disclosure document after reviewing the complete technical disclosure document and deeming the document unsatisfactory;
[0016] Determine that the modified document is the target technical briefing document.
[0017] In some embodiments, the second largest language model is retrained based on the modified document.
[0018] In some embodiments, the first large language model is a Yi-VL-6B model.
[0019] In some embodiments, it also includes:
[0020] Send the target technology briefing to the agent;
[0021] Obtaining feedback from the agent on revisions;
[0022] Based on the modification suggestions, training samples are generated to train the first language model and the second language model.
[0023] In some embodiments, the evaluation result includes: a prompt word;
[0024] The prompt word is used to guide the second largest language model to complete the necessary elements.
[0025] This application also provides a patented technology briefing document quality assessment and analysis device, including:
[0026] Acquisition module, used to obtain technical briefing documents;
[0027] An input module, used for inputting the technical briefing document into the pre-trained first language model to obtain an evaluation result of the technical briefing document;
[0028] The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result;
[0029] The completion module is used to generate and complete the missing necessary elements in the technical briefing book based on the evaluation results through a preset second language model to obtain a target technical briefing book.
[0030] The present application also provides an electronic device, including:
[0031] A processor, and a memory for storing a program executable by the processor;
[0032] The processor is used to implement the above-mentioned patent technology briefing document quality assessment and analysis method by running the program in the memory.
[0033] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the processor executes the patent technology briefing document quality assessment and analysis method as described above.
[0034] The present application provides a method for evaluating and analyzing the quality of patent technology briefing documents. First, the technology briefing document is obtained; the technology briefing document is input into the pre-trained first language model to obtain the evaluation result of the technology briefing document; wherein the first language model is used to evaluate whether the technology briefing document lacks necessary elements and obtain the evaluation result; based on the evaluation result, the missing necessary elements in the technology briefing document are generated and supplemented by the preset second language model to obtain the target technology briefing document. In this way, by inputting the technology briefing document into the pre-trained first language model, the evaluation result can be quickly obtained to determine whether the technology briefing document lacks necessary elements. Compared with the traditional method of manual review by patent agents word by word, the review time is greatly shortened and the work efficiency is improved. After evaluating the missing necessary elements, the missing parts are generated and supplemented by using the preset second language model, and the target technology briefing document can be quickly obtained. This process reduces the repeated communication and modification between the applicant and the patent agent, speeds up the entire patent application process, and enables the applicant to complete the patent application faster and improve efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0036] Figure 1 It is a flow chart of a method for evaluating and analyzing the quality of a patent technology briefing document provided in one embodiment of the present application.
[0037] Figure 2 It is a flowchart of a method provided by an embodiment of the present application.
[0038] Figure 3 It is a structural schematic diagram of a patent technology briefing document quality assessment and analysis device provided in one embodiment of the present application.
[0039] Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION
[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0041] Figure 1 1 is a flow chart of a method for evaluating and analyzing the quality of a patent technology briefing document provided in one embodiment of the present application. Figure 1 , Figure 2 As shown, the method includes the following contents.
[0042] Step S110, obtaining a technical briefing document;
[0043] Specifically, in this step, the patent applicant or relevant personnel submits technical documents related to the invention to the system set up based on the solution provided by this application. These documents may contain the applicant's preliminary description of the invention, technical implementation ideas, possible drawings, etc., but in the initial stage, their format and completeness may vary.
[0044] Specifically, the submitted technical briefing document can be in a variety of formats, such as common text formats (such as .doc, .txt, etc.), which is convenient for the system to read the text information; if it contains drawings, it may be in picture format (such as .jpg, .png, etc.). In terms of content, the applicant may elaborate on the technical solution in detail, but the description of the background technology, technical effects, etc. may be insufficient or non-standard. For example, some applicants may simply mention the function of the invention without in-depth analysis of its advantages and innovations over the prior art.
[0045] Step S120, inputting the technical briefing document into the pre-trained first language model to obtain a technical briefing document evaluation result;
[0046] The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result;
[0047] Specifically, first load the pre-trained first language model. This model has been pre-trained on a large amount of standard technical briefing sample data and has the ability to understand the structure and content of the technical briefing. Input the pre-processed technical briefing into the first language model. The model will conduct an in-depth analysis of the document and identify the necessary elements therein, such as background technology, technical solutions, technical effects, etc. The model evaluates whether the necessary elements are missing in the technical briefing based on preset standards and rules. For example, check whether there is a lack of detailed description of the background technology, clear explanation of the technical solution, etc. The model outputs the evaluation results of the technical briefing, clearly indicating which necessary elements are missing or incomplete. The evaluation results may be presented in the form of a report, listing the specific information of the missing elements, providing a basis for subsequent generation and completion.
[0048] Step S130, based on the evaluation result, the missing necessary elements in the technical briefing document are generated and completed through a preset second language model to obtain a target technical briefing document.
[0049] Specifically, first load the preset second largest language model, which is specially used to generate and complete the missing content in the technical briefing document.
[0050] The evaluation results and the existing technical briefing document content are input into the second largest language model. Based on the evaluation results and contextual information, the model generates content with missing necessary elements and completes it in the technical briefing document. The generated content should conform to the context and logic of the technical briefing document to ensure the coherence and consistency of the overall document.
[0051] Integrate the generated and completed missing elements into the original technical disclosure document to form a complete target technical disclosure document. The target technical disclosure document should meet the standards and requirements of patent applications and have complete necessary elements and high-quality descriptions.
[0052] Through the above steps, this method can effectively evaluate and analyze the quality of patent technology briefing documents, generate complete and compliant technology briefing documents, and improve the efficiency and quality of patent applications.
[0053] In some embodiments, the generation and completion of the missing necessary elements in the technical briefing book includes:
[0054] Initial generation and completion of the missing necessary elements in the technical disclosure document to obtain a complete technical disclosure document;
[0055] Send the complete technical briefing to the inventor for review and obtain the review results;
[0056] If the audit result indicates that the audit is passed, the complete technical briefing document will be determined as the target briefing document.
[0057] The following is a detailed description of the steps for generating and completing the missing necessary elements in the technical disclosure document in this embodiment:
[0058] First, based on the evaluation results of the first language model, clearly point out the necessary elements missing from the technical briefing document. For example, there may be a lack of detailed step descriptions of the technical solution or specific data support for the technical effect.
[0059] Determine the prompt words. It should be noted that, in some embodiments, the evaluation results include the missing necessary elements determined by the first largest language model and the prompt words generated by the first largest language model. In other embodiments, the prompt words can be generated by the second largest language model based on the evaluation results. The specific generation of prompt words (Prompt includes: for each missing necessary element, generate a corresponding prompt word. The prompt words should be specific and clear, and can guide the language model to generate content that meets the requirements. For example, if the step description of the technical solution is missing, the prompt word can be "Please describe in detail the various steps of the technical solution of the present invention and their interrelationships."
[0060] Input the prompt words and the existing technical disclosure document content into the second language model. Based on the prompt words and context information, the model generates preliminary content with missing necessary elements and completes it in the technical disclosure document. The generated content should be as accurate and complete as possible, and conform to the context and logic of the technical disclosure document.
[0061] Integrate the generated and completed missing elements into the original technical disclosure document to form a complete preliminary technical disclosure document. At this point, the technical disclosure document should have all the necessary elements, but may need further optimization and improvement.
[0062] The complete preliminary technical disclosure document is sent to the inventor through the system platform for review. The inventor is the creator of the patented technology and has the most in-depth understanding of the technical details and innovations.
[0063] Provide inventors with convenient review tools and interfaces, so that they can efficiently review various parts of the technical briefing document and mark, modify or make comments on the content. For example, a function can be provided to highlight the missing elements to be completed, so that inventors can focus on them.
[0064] During the review process, the inventor may find that the generated and completed content has certain problems, such as inaccurate descriptions, insufficient data, etc. The inventor can record these problems and modification opinions and submit them to the patent application system through the system.
[0065] The system receives the review results submitted by the inventor, which may include passing the review, requiring modification or supplementation, etc.
[0066] If the review result indicates that the review has passed, that is, the inventor believes that the content of the generated and completed technical briefing document is accurate and complete and meets the requirements of the patent application, then the complete technical briefing document will be determined as the target briefing document.
[0067] The technical briefing document determined as the target briefing document will be output for use in the subsequent patent application process. The target briefing document should have a high-quality description that can clearly display the technical solution, technical effects and other key information of the invention to increase the success rate of the patent application.
[0068] Through the steps of this embodiment, it is possible to ensure that the contents of the generated and completed technical briefing document are reviewed and approved by the inventor, thereby improving the quality and reliability of the technical briefing document and providing strong support for patent applications.
[0069] In some embodiments, it also includes:
[0070] Obtain a modified document provided by the inventor; wherein the modified document is obtained by the inventor modifying the complete technical briefing document after reviewing the complete technical briefing document and determining that the complete technical briefing document fails the review; determine that the modified document is the target technical briefing document.
[0071] The following is a detailed description of the steps in this embodiment to obtain the modified document provided by the inventor and determine it as the target technology briefing document:
[0072] First, we obtain the revised documents provided by the inventor. After reviewing the complete technical disclosure document, if the inventor finds that there are still areas that do not meet the requirements or need to be improved, the review will be considered as failed. For example, it may be found that some technical details are not described clearly, the data is not accurate, or there are logical errors.
[0073] The inventor shall modify the complete technical disclosure document based on the problems found during the review process. The modifications may include supplementing missing technical details, correcting erroneous data, optimizing language expression, etc., to improve the quality and accuracy of the technical disclosure document.
[0074] The inventor submits the revised technical disclosure document back to the patent application system through the system platform. At this time, the revised document already contains the inventor's further improvement and perfection of the technical disclosure document, which is closer to the ideal state that meets the patent application requirements.
[0075] The patent application system receives the revised documents submitted by the inventor, records and stores them. The system needs to ensure that the received documents are complete, accurate and ready for subsequent processing.
[0076] Confirm that the modified document is the target technical briefing document
[0077] Furthermore, the system conducts preliminary verification of the modified document provided by the inventor to check whether it meets the basic format requirements and the structural specifications of the technical disclosure document. At the same time, the system can also conduct a simple analysis of the modified content to confirm whether it solves the problems that failed the previous review. If the verification passes, the system will determine the modified document as the target technical disclosure document. This means that the document has been reviewed and modified by the inventor, meets the standards and requirements of the patent application, and can be used as the final patent technical disclosure document submitted.
[0078] The modified document determined as the target briefing document will be output for use in the subsequent patent application process. The target briefing document should have a high-quality description that can clearly and accurately display the key information such as the technical solution and technical effect of the invention, providing strong support for the success of the patent application.
[0079] The system records the inventor's modification process and content of the technical briefing, including the comparison of versions before and after the modification, the specific content of the modification, etc. These records can be used as an important reference for subsequent patent application review, patent protection and other links.
[0080] Through the steps of this embodiment, the professional knowledge and experience of the inventor can be fully respected and utilized to ensure that the quality of the technical briefing document is improved to the greatest extent, thereby increasing the success rate of patent applications and the validity of patents.
[0081] In some embodiments, the method further includes: retraining the second largest language model based on the modified document.
[0082] The following is a detailed description of the steps for retraining the second largest language model based on the modified document in this embodiment:
[0083] First, collect the modified document data, and extract the specific content of the inventor's modifications to the technical briefing from the modified document submitted by the inventor. This includes supplementary technical details, revised data, optimized language expressions, and other parts. The extracted modified content will become an important source of data for retraining the model. Annotate the extracted modified content, and clearly indicate which parts are supplementary, which are revised, and which are optimized. Annotation can help the model better understand different types of modifications during the training process, so as to more accurately learn how to generate and complete missing elements in the technical briefing. Integrate the data in the modified document with the dataset previously used to train the second largest language model. The integrated dataset contains more practical application scenarios and high-quality modification examples, which can provide richer learning resources for the model.
[0084] Then retrain the second language model: According to the characteristics of the modified document and the current performance of the model, select the appropriate training strategy. For example, incremental training can be used to further train the original model using the data of the modified document to improve the adaptability and accuracy of the model. During the retraining process, the parameters of the model may need to be adjusted to better adapt to the new data and task requirements. For example, parameters such as the learning rate and optimization algorithm can be adjusted so that the model can learn the information in the modified document faster. Monitor the training process of the model in real time, and pay attention to the changes in indicators such as the loss function value and accuracy of the model. Through monitoring, problems in the training process, such as overfitting and underfitting, can be discovered in time, and corresponding measures can be taken to adjust them. After the training is completed, the model is verified using a part of the modified document data that did not participate in the training to evaluate the performance of the model in generating and completing the missing elements in the technical briefing. Verification indicators can include the accuracy, coherence, and consistency of the generated content.
[0085] Update model application: Deploy the retrained second largest language model to the patent technical disclosure document quality assessment and analysis system to replace the original model. The updated model can better understand and generate the content in the technical disclosure document, and improve the system's generation and completion quality of missing elements.
[0086] Continuously optimize the model: In the actual application process, continue to collect user feedback and new modified document data, and continuously optimize and iterate the model. Through continuous optimization, the model will be able to better adapt to various technical fields and patent application scenarios, and provide users with higher quality services. The retrained model will significantly improve the overall effect of the system on the quality assessment and analysis of technical briefing documents. The generated and completed missing elements will be more accurate and meet the requirements, thereby improving the success rate of patent applications and the quality of patents. By retraining the second largest language model based on the modified documents, the model can continuously learn and adapt to changes in demand in actual applications, improve its ability to generate and complete missing elements in technical briefing documents, and provide stronger support for patent applications.
[0087] The process of training the first and second language models includes the following key steps:
[0088] The first language model training includes:
[0089] Collect a large number of technical briefing documents: Collect a large number of technical briefing documents from patent databases, patent agencies, etc. These documents should cover different technical fields and patent types to ensure that the model can learn a wide range of patent knowledge and language expressions.
[0090] Data cleaning and preprocessing: Clean and preprocess the collected technical briefing book samples, including removing irrelevant information (such as headers and footers, formatting tags, etc.), correcting typos, unifying terminology and formats, etc. At the same time, perform operations such as word segmentation and part-of-speech tagging to prepare for model training.
[0091] Select the appropriate model architecture: According to the task requirements and data characteristics, select the appropriate pre-trained language model architecture, such as Transformer, BERT, Yi-VL-6B model, etc. These models have powerful natural language processing capabilities and can understand and generate complex text content. Based on the selected model architecture, build the specific structure of the model, including determining the number of layers of the model, the number of hidden units, the configuration of the attention mechanism, etc. At the same time, design the input and output interfaces of the model so that it can receive the technical briefing text and output the evaluation results.
[0092] Use a large number of technical briefing samples to pre-train the model. In the pre-training stage, the model mainly learns the basic features and language rules of the text, such as the co-occurrence relationship of words, the structure of sentences, etc. Through pre-training, the model can initially have the ability to understand the text of the technical briefing. On the basis of pre-training, supervised training is carried out. Prepare annotated training data, and the annotation content includes the existence or non-existence of each necessary element in the technical briefing (such as background technology, technical solutions, etc.). By learning the annotated data, the model further improves its ability to recognize the structure and necessary elements of the technical briefing. During the training process, the parameters and hyperparameters of the model, such as the learning rate, regularization coefficient, etc., are continuously optimized. At the same time, according to the performance of the model on the validation set, the model structure is fine-tuned to improve the accuracy and generalization ability of the model.
[0093] Select appropriate evaluation indicators, such as accuracy, recall, F1 score, etc., to measure the model's performance in identifying the necessary elements of the technical briefing document. Use an independent validation dataset to validate the model and evaluate its performance in different technical fields and patent types. Ensure that the model has good generalization ability and stability.
[0094] The second language model training includes: The overall process is similar to the first language model training. When generating training samples and selecting models, you need to make adjustments based on your actual needs.
[0095] In some embodiments, the method for evaluating and analyzing the quality of a technical briefing document is characterized by further comprising:
[0096] Send the target technology briefing to the agent;
[0097] Obtaining feedback from the agent on revisions;
[0098] Based on the modification suggestions, training samples are generated to train the first language model and the second language model.
[0099] The following is a detailed description of the steps of the quality assessment and analysis method of the technical disclosure document in this embodiment:
[0100] First, send the target technology briefing to the agent: select a suitable patent agent based on the specific requirements of the patent application and the content of the target technology briefing. Patent agents should have rich experience and professional knowledge in patent applications, and be able to conduct in-depth analysis and evaluation of technical briefings. Send the target technology briefing after quality assessment and generation completion to the selected patent agent through the system platform. During the sending process, ensure the integrity and correct format of the briefing to facilitate the agent's subsequent review and modification. While sending the briefing, you can provide the agent with relevant background information, such as the technical field of the invention, innovation points, application strategy, etc. This information helps the agent better understand the content of the briefing and the purpose of the application, so as to provide more accurate feedback.
[0101] Obtain feedback and revision opinions from the agent: After receiving the target technology briefing document, the patent agent will conduct a detailed review of it. From the perspective of patent application, the agent will evaluate whether the structure, content, language expression and other aspects of the briefing document meet the requirements of patent law and application standards.
[0102] During the review process, the agent will record the modification opinions on the briefing document. These opinions may include further refinement of the technical solution, supplementation of background technology, optimization of language expression, etc., aiming to improve the quality of the briefing document and the success rate of patent applications.
[0103] The agent will submit the recorded modification opinions back to the patent application system through the system platform. The submitted feedback should be clear and specific to facilitate subsequent processing and application.
[0104] Generate training samples based on modification opinions to train the model: Extract key information from the agent's feedback, such as the specific content that needs to be modified, the direction and requirements of the modification, etc. This key information will become the basis for generating training samples. Generate corresponding training samples based on the extracted key information. The training samples should include the original technical briefing content, the agent's modification opinions, and the ideal content after the modification. For example, if the agent suggests refining the technical solution, the technical solution content before and after the refinement can be used as a pair of training samples. Annotate the generated training samples to clearly indicate which parts are original content, which are modification opinions, and which are ideal content after the modification. Annotation can help the model better understand the intention and direction of the modification during the training process.
[0105] The generated training samples are integrated with the datasets previously used to train the first and second language models. The integrated datasets contain more practical application scenarios and high-quality modification examples, which can provide richer learning resources for the model.
[0106] The first and second language models are trained using the integrated data set. The models will learn how to accurately evaluate and generate completions for the technical briefing document based on the modification opinions, thereby improving the performance and application effect of the models. Through training, the first language model can better identify the areas in the technical briefing document that need to be modified and improved, and improve the accuracy and comprehensiveness of the evaluation. Through training, the second language model training can more accurately generate completion content that conforms to the modification opinions, improving the quality and compliance of the generated content.
[0107] Through the steps of this embodiment, the professional knowledge and experience of patent agents can be fully utilized to continuously optimize and improve the models in the technical briefing document quality assessment and analysis methods, so that they can better adapt to actual patent application needs and improve the success rate of patent applications and the quality of patents.
[0108] Exemplary Devices
[0109] The device embodiments of the present application can be used to execute the method embodiments of the present application. For details not disclosed in the device embodiments of the present application, please refer to the method embodiments of the present application.
[0110] Figure 3 The block diagram of the device for evaluating and analyzing the quality of patent technology briefing documents provided by one embodiment of the present application is shown. Figure 3 As shown, the device comprises:
[0111] An acquisition module 31 is used to acquire a technical briefing document;
[0112] An input module 32 is used to input the technical briefing document into the pre-trained first language model to obtain a technical briefing document evaluation result;
[0113] The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result;
[0114] The completion module 33 is used to generate and complete the missing necessary elements in the technical briefing document based on the evaluation result and through a preset second language model to obtain a target technical briefing document.
[0115] In some embodiments, the generation and completion of the missing necessary elements in the technical briefing book includes:
[0116] Initial generation and completion of the missing necessary elements in the technical disclosure document to obtain a complete technical disclosure document;
[0117] Send the complete technical briefing to the inventor for review and obtain the review results;
[0118] If the audit result indicates that the audit is passed, the complete technical briefing document will be determined as the target briefing document.
[0119] In some embodiments, it also includes:
[0120] Obtaining a modified document provided by the inventor; wherein the modified document is obtained by the inventor modifying the complete technical disclosure document after reviewing the complete technical disclosure document and deeming the document unsatisfactory;
[0121] Determine that the modified document is the target technical briefing document.
[0122] In some embodiments, the second largest language model is retrained based on the modified document.
[0123] In some embodiments, the first large language model is a Yi-VL-6B model.
[0124] In some embodiments, it also includes:
[0125] Send the target technology briefing to the agent;
[0126] Obtaining feedback from the agent on revisions;
[0127] Based on the modification suggestions, training samples are generated to train the first language model and the second language model.
[0128] In some embodiments, the evaluation result includes: a prompt word;
[0129] The prompt word is used to guide the second largest language model to complete the necessary elements.
[0130] Exemplary Electronic Devices
[0131] Below, reference Figure 4 To describe an electronic device according to an embodiment of the present application. Figure 4 A block diagram of an electronic device according to an embodiment of the present application is illustrated.
[0132] like Figure 4 As shown, electronic device 400 includes one or more processors 410 and memory 420 .
[0133] The processor 410 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 400 to perform desired functions.
[0134] The memory 420 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 410 may run the program instructions to implement the patent technology briefing quality assessment and analysis methods of the various embodiments of the present application described above and / or other desired functions. Various contents such as category correspondences may also be stored in the computer-readable storage medium.
[0135] In one example, the electronic device 400 may further include: an input device 430 and an output device 440 , and these components are interconnected via a bus system and / or other forms of connection mechanisms (not shown).
[0136] In addition, the input device 430 may also include, for example, a keyboard, a mouse, an interface, etc. The output device 440 may output various information to the outside, including analysis results, etc. The output device 440 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.
[0137] Of course, to simplify, Figure 4 Only some of the components in the electronic device related to the present application are shown, and components such as a bus, an input / output interface, etc. are omitted. In addition, the electronic device may further include any other appropriate components according to specific application conditions.
[0138] Exemplary computer program products and computer-readable storage media
[0139] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the patent technology briefing document quality assessment and analysis method according to various embodiments of the present application described in the above-mentioned "Exemplary Method" section of this specification.
[0140] The computer program product may be written in any combination of one or more programming languages to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.
[0141] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, causes the processor to execute the steps of the method for evaluating and analyzing the quality of patent technology briefing documents according to various embodiments of the present application described in the above “Exemplary Method” section of this specification.
[0142] The computer readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0143] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.
Claims
1. A method for evaluating and analyzing the quality of patent technology briefing documents, characterized in that: include: Obtain technical briefing documents; Input the technical briefing document into the pre-trained first language model to obtain a technical briefing document evaluation result; The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result; Based on the evaluation results, the missing necessary elements in the technical briefing document are generated and completed through the preset second language model to obtain the target technical briefing document.
2. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 1 is characterized in that: Generate and complete the missing necessary elements in the technical disclosure document, including: Initial generation and completion of the missing necessary elements in the technical disclosure document to obtain a complete technical disclosure document; Send the complete technical briefing to the inventor for review and obtain the review results; If the audit result indicates that the audit is passed, the complete technical briefing document will be determined as the target briefing document.
3. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 2 is characterized in that: Also includes: Obtaining a modified document provided by the inventor; wherein the modified document is obtained by the inventor modifying the complete technical disclosure document after reviewing the complete technical disclosure document and deeming the document unsatisfactory; Determine that the modified document is the target technical briefing document.
4. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 3 is characterized in that: The second largest language model is retrained based on the modified document.
5. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 1 is characterized in that: The first largest language model is the Yi-VL-6B model.
6. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 2 is characterized in that: Also includes: Send the target technology briefing to the agent; Obtaining feedback from the agent on revisions; Based on the modification suggestions, training samples are generated to train the first language model and the second language model.
7. The method for evaluating and analyzing the quality of patent technology briefing documents according to claim 1 is characterized in that: The evaluation result includes: prompt words; The prompt word is used to guide the second largest language model to complete the necessary elements.
8. A patented technology briefing document quality assessment and analysis device, characterized in that: include: Acquisition module, used to obtain technical briefing documents; An input module, used for inputting the technical briefing document into the pre-trained first language model to obtain an evaluation result of the technical briefing document; The first language model is used to evaluate whether the technical briefing document lacks necessary elements to obtain an evaluation result; The completion module is used to generate and complete the missing necessary elements in the technical briefing book based on the evaluation results through a preset second language model to obtain a target technical briefing book.
9. An electronic device, characterized in that: include: A processor, and a memory for storing a program executable by the processor; The processor is used to implement the patent technology briefing document quality assessment and analysis method as described in any one of claims 1 to 7 by running the program in the memory.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which, when executed by a processor, enables the processor to execute the patent technology briefing document quality assessment and analysis method as described in any one of claims 1 to 7.