Specification creation support system

The specification drafting support system addresses the limitations of existing systems by using a language model to generate and refine specifications, ensuring high-quality, patentable, and easily enforceable specifications with a broader scope, thus improving efficiency and effectiveness in specification creation.

JP2026012108APending Publication Date: 2026-01-23SEMICON ENERGY LAB CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2025113945
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2024-07-12
Filing Date
2025-07-04
Publication Date
2026-01-23

AI Technical Summary

Technical Problem

Existing specification preparation systems, such as those described in Patent Document 1, do not consider the criteria for obtaining high-quality specifications, including ease of obtaining a patent, ease of enforcing rights, and broad scope of rights, and they do not address reducing the time required for specification creation.

Method used

A specification drafting support system that utilizes a language model to generate and refine specifications by accepting an invention summary, dividing the specification into chapters, retrieving and revising chapters based on additional information, and comparing versions to select the most suitable content for inclusion in the final specification, ensuring high-quality and efficient specification creation.

Benefits of technology

Enables the creation of high-quality specifications that are more likely to be patentable, easier to enforce rights, and have a broader scope of rights, while reducing the time required for preparation, even for users unfamiliar with language models.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026012108000001_ABST
    Figure 2026012108000001_ABST
Patent Text Reader

Abstract

To provide a specification creation support system using a language model.SOLUTION: Receiving an outline of an invention, generating a first prompt including content for creating a first specification corresponding to the outline of the invention, acquiring the first specification from a language model as a response to the first prompt, displaying the first specification, segmenting the first specification into chapters, and generating a second prompt including content for acquiring information related to a segmented first chapter from a second component and content for creating a second chapter by revising the first chapter according to the information; The specification creation support system has a function of acquiring a second chapter as a response to the second prompt from the language model, a function of displaying the first chapter and the second chapter, a function of comparing the first chapter and the second chapter and selecting one suitable for the invention outline, and a function of creating a second specification including the selected chapter.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] The present invention relates to a specification preparation support system that utilizes a language model, typically a large language model (LLM).

[0002] The above technical field is one embodiment of the present invention, and the present invention is not limited to the above technical field. Other embodiments of the present invention can include, for example, a semiconductor device, a display device, a light-emitting device, a power storage device, a memory device, an electronic device, a lighting device, an input device (for example, a touch sensor), an input / output device (for example, a touch panel), a driving method thereof, or a manufacturing method thereof. [Background technology]

[0003] Although engineers have a good understanding of the technology related to inventions, it is difficult to prepare a highly complete patent application specification (referred to as a specification in this specification etc.) without practical experience. In light of the above, Patent Document 1 proposes a preparation support system that enables even engineers who are unfamiliar with specifications to easily prepare specification documents.

[0004] A technique called Retrieval-Augmented Generation (RAG) has attracted attention as a way to improve the accuracy of LLM answers. This technique involves preparing a database of related information in advance, and when a user asks a question, searching the database for information highly relevant to the question, adding the search results to the question and inputting them into LLM, thereby obtaining a highly accurate answer. Non-Patent Document 1 discloses a mechanism related to RAG, and shows that in language generation tasks, RAG enables language generation based on specific and diverse facts. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2013-80279 [Non-patent literature]

[0006] [Non-Patent Document 1] Patrick Lewis et.al.,”Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks”, 12 Apr 2021 Summary of the Invention [Problem to be solved by the invention]

[0007] Generally, the criteria that a high-quality specification should meet include the ease of obtaining a patent, the ease of enforcing rights, and the broad scope of rights. However, the system of Patent Document 1 does not consider any of the above criteria, so there is room for improvement in creating high-quality specifications. Furthermore, Patent Document 1 does not consider reducing the time required to create specifications.

[0008] The present invention has been made in consideration of the above description, and an object of one aspect of the present invention is to provide a novel specification writing support system and an operation method thereof. Another object of one aspect of the present invention is to provide a specification writing support system and an operation method thereof that are capable of obtaining high-quality specifications.

[0009] The present invention does not necessarily solve all of these problems. Furthermore, the description of these problems does not preclude the existence of other problems of the present invention. For example, problems other than these can be extracted from the description of the specification, drawings, and claims. [Means for solving the problem]

[0010] In view of the above problems, one aspect of the present invention is a specification drafting support system having a first component, the first component having the following functions: accepting an invention summary; generating a first prompt including content for creating a first specification corresponding to the invention summary; retrieving the first specification from a language model as a response to the first prompt; displaying the first specification; dividing the first specification into chapters; generating a second prompt including content for retrieving information related to the divided first chapters from a second component and content for revising the first chapter in accordance with the information to create a second chapter; retrieving the second chapter from a language model as a response to the second prompt; displaying the first chapter and the second chapter; comparing the first chapter and the second chapter and selecting one that is more suitable for the invention summary; and creating a second specification including the selected chapter.

[0011] Another aspect of the present invention is a specification drafting support system including a first component having the following functions: accepting an invention abstract; generating a first prompt including content for creating a first specification corresponding to the invention abstract; retrieving the first specification from a language model as a response to the first prompt; displaying the first specification; dividing the first specification into chapters; generating a second prompt including content for retrieving a plurality of pieces of information related to the divided first chapters from a second component, content for revising the first chapter in accordance with the plurality of pieces of information to create a second chapter, and content for creating a third chapter in addition to the second chapter in accordance with the plurality of pieces of information; retrieving a plurality of pieces of information, the second chapter, and the third chapter, as a response to the second prompt from a language model; displaying the plurality of pieces of information, the second chapter, and the third chapter; comparing the second chapter and the third chapter and selecting one or both of them that are suitable for the invention abstract; and creating a second specification including the selected chapters.

[0012] In the present invention, when both chapters suitable for the invention summary are selected, the first component is a specification drafting support system having a function of generating a third prompt including content for describing the invention-specifying matters included in the second chapter as independent claims in a third specification, and content for describing the invention-specifying matters included in the third chapter as dependent claims in the third specification, with the dependent claims citing the independent claims, and having a function of obtaining the third specification from a language model as an answer to the third prompt.

[0013] Another aspect of the present invention is a specification drafting support system having a first component having the following functions: a function for accepting an invention summary and first experimental data; a function for generating a first prompt including content for creating a first specification corresponding to the invention summary and content for describing the first experimental data in a first example of the first specification; a function for retrieving the first specification from a language model in response to the first prompt; a function for displaying the first specification; a function for dividing the first specification into chapters; a function for generating a second prompt including content for retrieving information related to the first example from a second component and content for revising the first example in accordance with the first experimental data to create a second example; a function for retrieving the second example from a language model in response to the second prompt; a function for displaying the first example and the second example; a function for comparing the first example and the second example and selecting one that best suits the invention summary; and a function for creating a second specification including the selected example.

[0014] Another aspect of the present invention includes a first component, the first component having a function of receiving an outline of the invention and first experimental data, a function of generating a first prompt including content for creating a first specification corresponding to the outline of the invention and content for describing the first experimental data in a first embodiment of the first specification, a function of obtaining the first specification from a language model as a response to the first prompt, a function of displaying the first specification, a function of dividing the first specification into chapters, a function of obtaining a plurality of pieces of information related to the first embodiment from a second component, and a function of generating a first prompt including content for describing the first experimental data in a first embodiment of the first specification. a function to generate a second prompt including content to revise the first example to create a second example and content to create a third example in addition to the second example according to a plurality of pieces of information; a function to obtain the plurality of pieces of information, the second example and the third example, from a language model as a response to the second prompt; a function to display the plurality of pieces of information, the second example and the third example; a function to compare the second example and the third example and select one or both that are suitable for the invention summary; and a function to create a second specification including the selected examples.

[0015] In the present invention, when both embodiments suitable for the invention summary are selected, the first component is a specification drafting support system having a function of generating a third prompt including content for describing the invention-specifying matters included in the second embodiment as independent claims in a third specification, and content for describing the invention-specifying matters included in the third embodiment as dependent claims in the third specification, with the dependent claims citing the independent claims, and having a function of obtaining the third specification from a language model as an answer to the third prompt.

[0016] In the present invention, the invention summary is a specification drafting support system that includes a description of the problem and a description of the effect. [Effects of the Invention]

[0017] According to one aspect of the present invention, it is possible to provide a novel specification writing support system and an operating method thereof. Also, according to another aspect of the present invention, it is possible to provide a specification writing support system that is capable of obtaining high-quality specifications.

[0018] The present invention does not necessarily have to have all of these effects. Furthermore, the description of these effects does not preclude the existence of other effects of the present invention. For example, other effects can be extracted from the description of the specification, drawings, and claims. [Brief explanation of the drawings]

[0019] [Figure 1] 1(A) and 1(B) are diagrams showing examples of screens of the specification preparation support system of the present invention. [Figure 2] 2(A) and 2(B) are diagrams showing examples of screens of the specification preparation support system of the present invention. [Figure 3] 3(A) and 3(B) are diagrams showing examples of screens of the specification preparation support system of the present invention. [Figure 4] FIG. 4 is a diagram showing an example of a specification preparation support system according to the present invention. [Figure 5] FIG. 5 is a diagram showing an example of a specification preparation support system according to the present invention. [Figure 6] FIG. 6 is a diagram showing an example of a specification preparation support system according to the present invention. [Figure 7] FIG. 7 is a flowchart showing an example of the operation method of the specification preparation support system of the present invention. [Figure 8] FIG. 8 is a flowchart showing an example of the operation method of the specification preparation support system of the present invention. [Figure 9] FIG. 9 is a flowchart showing an example of the operation method of the specification preparation support system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0020] The following description will discuss embodiments of the present invention with reference to the accompanying drawings. However, it will be readily understood by those skilled in the art that the present invention can be modified in various ways without departing from the spirit and scope of the present invention. Therefore, the present invention should not be construed as being limited to the following description of the embodiments.

[0021] In the drawings, the position, size, range, etc. of each component may not accurately represent the actual components, and therefore the position, size, range, etc. of each component are not necessarily limited to the position, size, range, etc. disclosed in the drawings.

[0022] In this specification, the words "first" and "second" are used for the convenience of understanding the technical content or to identify each component. Therefore, the words "first" and "second" do not limit the number of each component. Furthermore, the words "first" and "second" do not limit the order of each component. Furthermore, the words "first" and "second" or identifying symbols used in this specification may not match the words or identifying symbols in the claims.

[0023] In this specification, a language model is based on the Transformer architecture and is incrementally trained to become a conversational (also called dialog-based) model. In other words, a conversational generative model is a lower-level concept of a language model. A language model is also generally called a large-scale language model (LLM).

[0024] In this specification, a document refers to a written expression of human intent using characters or symbols. A document may be divided into multiple chapters or multiple paragraphs. A chapter may also consist of multiple paragraphs.

[0025] (Embodiment) In this embodiment, an example of a specification preparation support system will be described.

[0026] <Screen and transition of specification writing support system> The screens and screen transitions for using the specification drafting support system will be explained using Figures 1(A) to 3(B). The figures shown in Figures 1(A) to 3(B) are examples of screens displayed on a user's information terminal. Note that the screen transitions can be realized by information processing such as the first information processing device 10 described below, but are not limited to this. The user is expected to be an engineer who proposes an invention, or a person in a company's intellectual property department, but is not limited to these.

[0027] As shown in FIG. 1(A), a screen 200 for inputting conditions for receiving specification preparation support is displayed on a user's information terminal. Since the specification preparation support system has a function for accepting an invention summary, it is preferable that the screen 200 has an invention summary input field 201. The invention summary input field 201 has a configuration field 100, in which the configuration of the invention can be input. The invention summary input field 201 has a problem field 101, in which the problem of the invention can be input. The invention summary input field 201 has an effect field 102, in which the effect of the invention can be input. Furthermore, since the problem and the effect are two sides of the same coin, a specification may be configured such that inputting information into one field eliminates the need to input information into the other field.

[0028] Note that each of the above-mentioned fields can be read as an input area, and the position of each field on the screen 200 is not limited to that shown in the figure. The invention summary input field 201 may also have other input fields. Examples of other input fields include a field for inputting prior art documents, a field for inputting the inventor, a field for inputting drawings, and a field for inputting experimental data. In particular, it is preferable if the specification drafting support system has a function for accepting experimental data (referred to as first experimental data for distinction), since this allows examples based on the experimental data to be included in the specification. Note that experimental data includes experimental reports.

[0029] Input into each field can be done by either entering text into a text box or by uploading a file, or both. When uploading a file, if the file does not contain text data, the specification drafting support system preferably has the function of extracting text data from the file and saving that text data. Furthermore, the specification drafting support system preferably has the function of preparing chunk data corresponding to each file. Chunk data is text data that groups together connected parts of document data.

[0030] Providing multiple fields according to purpose, as in screen 200 shown in Figure 1(A), is preferable because it reduces user input errors. However, when an inexperienced user enters data, they may not fully understand the items in each field. Therefore, it is useful to provide a field in which data can be freely entered without specifying items, instead of each field. Depending on the user's level of proficiency, the specification preparation support system may have a function to switch between displaying each field or a field in which data can be freely entered.

[0031] The invention summary input field 201 has multiple buttons such as Save, Cancel, and Create. The Save button 105 is a button for saving the input content, and the Cancel button 106 is a button for canceling the input content. The Save button 105 can be applied to any of the screens 200 described below. The Cancel button 106 can be applied to any of the screens 200 described below.

[0032] The specification drafting support system preferably has a generate button 107 on screen 200 to generate a prompt (sometimes referred to as a first prompt for distinction) containing content for creating a specification corresponding to the invention summary (sometimes referred to as a first specification for distinction) using information entered in invention summary input field 201. If the specification drafting support system has the function of creating a first prompt (sometimes referred to as automatic generation) without the user having to create the first prompt, it becomes possible to obtain a highly accurate answer regardless of the user's level of proficiency. Note that the function of automatically generating the first prompt includes a function in which the specification drafting support system edits an existing prompt (a past prompt used in the system) according to the invention summary entered by the user. Preferably, a checkbox 109 is provided adjacent to the generate button 107 so that the user can select whether to automatically generate a first prompt.

[0033] The specification drafting support system may include in the first prompt content describing the first experimental data in an example of the first specification (referred to as the first example for the sake of distinction), and the first prompt allows the first example based on the first experimental data to be included in the first specification.

[0034] Such a specification writing support system enables even a user who is unfamiliar with language model processing to obtain a highly accurate answer. However, in the specification writing support system, the user may create the first prompt, and by unchecking automatic generation and pressing generate button 107, a field for the user to enter the first prompt may be displayed on screen 200.

[0035] Furthermore, since the specification drafting support system retrieves the first specification from the language model as a response to the first prompt, it is preferable that, by pressing the generate button 107, the specification drafting support system automatically generates the first prompt based on the invention summary entered by the user, then provides the first prompt to the language model, and executes a series of processes up to retrieving the first specification from the language model as a response. Such generate button 107 can be applied to any of the screens 200 described below.

[0036] The specification preparation support system can use BERT (Bidirectional Encoder Representations from Transformers) and T5 (Text-to-Text Transfer Transformer) as language models. When an LLM is used as the language model, the LLM can be GPT-3 (registered trademark), GPT-3.5, GPT-4 (registered trademark), LaMDA (Language Model for Dialogue Applications), PaLM (Pathways Language Model), Llama2, Llama3, or the like.

[0037] The LLM has appropriate values ​​for the number of input tokens and the number of output tokens depending on the learning process, algorithm, or hardware constraints. In this specification, the number of input tokens and the number of output tokens are collectively referred to as the number of tokens that the LLM can process at one time. An appropriate number of tokens may be set in the specification writing support system. For example, a prompt used in the specification writing support system may specify the number of output tokens.

[0038] Providing multiple input fields and multiple buttons as shown in FIG. 1(A) is preferable because it reduces input errors and operation errors by the user.

[0039] <Screen transition 1> Pressing the generate button 107 in the invention summary input field 201 transitions the screen 200 from FIG. 1(A) to FIG. 1(B). Since the specification drafting support system, having acquired the first specification from the language model, has the function of displaying the first specification, it is preferable that the screen 200 have a specification display field 202. Specifically, as shown in FIG. 1(B), the first specification, which is the response from the language model, is displayed in the specification display field 202, allowing the user to confirm the first specification. Note that in the first specification, content other than the information entered in the invention summary input field 201 is within the scope of the language model's prior learning. In this case, it is preferable that the invention summary input field 201 and the specification display field 202 are displayed so that they do not overlap. When reviewing the first specification in the specification display field 202, the user can confirm whether the content entered in the invention summary input field 201 was appropriate. Furthermore, the user can confirm the first working example included in the first specification.

[0040] Furthermore, if the document relating to the first specification does not fit in the specification display field 202, it is preferable that the specification display field 202 has a scroll bar. The scroll bar can be applied to any of the fields described below.

[0041] The specification writing support system has a function to divide the first specification into appropriate sections. Specifically, since it has a function to divide the first specification into chapters, it is preferable that the screen 200 has a division button 111. At least one of the divided chapters is referred to as the first chapter. The specification writing support system typically divides the specification into embodiments or examples. Embodiments and examples are examples of chapters that are well-defined technically, and are preferable as divided chapters. Furthermore, it is not necessary to divide the specification into embodiments or examples, but it is preferable to divide the first specification into sections that will result in a technically cohesive document.

[0042] Taking into account the token limit of the language model, the specification writing support system can set an upper limit on the number of characters in Chapter 1. If the number of characters in Chapter 1 exceeds the set value, the specification writing support system can further divide the first chapter into a new chapter by pressing the division button 111 again.

[0043] As for other buttons, it is preferable that the screen 200 has a save button 112, a cancel button 113, an output button 114, etc. Since the explanation of the previously mentioned buttons can be made by referring to FIG. 1(A), an explanation of the save button 112 and the cancel button 113 will be omitted. The output button 114 is a button for outputting the document displayed in the specification display field 202. Furthermore, there are no limitations on the format of the output file.

[0044] <Screen transition 2> Pressing the division button 111 transitions from the screen 200 in FIG. 1(B) to the screen 200 in FIG. 2(A). Specifically, as shown in FIG. 2(A), the specification display field 202 is replaced with a specification division display field 203a, and the first chapter is displayed. At this time, it is preferable that the invention summary input field 201 and the specification division display field 203a are displayed so that they do not overlap. When checking the first chapter in the specification division display field 203a, the user can check whether the contents of the invention summary input field 201 were appropriate.

[0045] The specification division display field 203a preferably has a division button 111, a next button 116, a back button 117, a generate button 118, a check box 119 for automatic generation, and the like. Since the previously mentioned buttons can be seen in Figures 1(A) and 1(B), a description of the division button 111, the generate button 118, and the check box 119 will be omitted. The next button 116 in Figure 2(A) is a button for sequentially advancing to the next chapter when there are multiple chapters displayed in the specification division display field 203a, and the back button 117 is a button for sequentially reverting to the previous chapter. Instead of, or in addition to, the next button 116 and the back button 117, a button for displaying a list of multiple chapters may be provided.

[0046] The specification drafting support system has a function for generating a prompt (referred to as a "second prompt" for distinction) that includes information related to the first chapter retrieved from a database and information for revising the first chapter based on the information to create a new chapter (the revised version may be referred to as the "second chapter" for distinction). Therefore, it is preferable that the screen 200 has a generate button 118. Typically, it is preferable that the specification division display field 203a has the generate button 118. If the specification drafting support system has the function of automatically generating a second prompt without the user having to create a second prompt, it becomes possible to obtain a highly accurate answer regardless of the user's level of proficiency. The function of automatically generating a second prompt includes a function in which the specification drafting support system edits an existing prompt (a past prompt used in the system) according to the invention summary entered by the user. It is preferable that a checkbox 109 be provided adjacent to the generate button 107, allowing the user to select whether to automatically generate a second prompt.

[0047] Furthermore, the specification writing support system can include in the second prompt information related to the first embodiment retrieved from a database. It is preferable that the user can select whether or not to automatically generate the second prompt.

[0048] Such a specification writing support system enables even a user who is unfamiliar with language model processing to obtain a highly accurate answer. However, in the specification writing support system, the user may also create a second prompt, and by unchecking automatic generation and pressing generate button 107, a field for the user to enter the second prompt may be displayed on screen 200.

[0049] Furthermore, since the specification writing support system retrieves the second chapter from the language model as a response to the second prompt, it is preferable that by pressing the generate button 107, the specification writing support system automatically generates the second prompt based on the first chapter, then provides the second prompt to the language model, and executes a series of processes up to retrieving the second chapter as a response from the language model. Such generate button 107 can be said to have the same function as that described for the first prompt.

[0050] Furthermore, the specification drafting support system can obtain a second example by revising the first example based on the information. If there is more than one piece of information, the specification drafting support system can also obtain a third example by adding it to the second example.

[0051] The information related to the first chapter preferably includes information about the company to which the user belongs, and the specification drafting support system can acquire this information by performing a similarity search on the company's database. As a similarity search, a search using keywords included in the first chapter or a phrase search using phrases included in the first chapter is preferably performed. A keyword search is a search that collects characters by breaking them down into spaces before and after the characters. A phrase search is a search that collects characters that include spaces. Such keyword and phrase searches are low-load and can obtain results in a short time. The specification drafting support system can perform a vector search as another form of similarity search. A vector search converts data, including text and images, into numerical vectors and calculates the similarity between the vectors. A vector search can be processed faster than a keyword search and can obtain results that are more relevant. In other words, a vector search is preferable because it can obtain results close to what the user intended, even if common words cannot be identified.

[0052] The specification drafting system can also use language models to extract information from a company's database. For example, by providing prompts based on similarity search information to the language model, the specification drafting system can retrieve that information. The RAG method is preferably used for similarity search.

[0053] The specification writing support system preferably determines whether the information is related to the first chapter. This determination can be made by converting the sentence into a distributed representation and using the similarity value based on the distributed representation. In this case, the specification writing support system preferably has the function of converting the first chapter into a distributed representation of the sentence. It is also preferable to add the information converted into a distributed representation of the sentence to the information stored in the database. Note that the distributed representation of the sentence can be created using, for example, doc2vec, Sentence BERT, Universal Sentence Encoder, BERTScore, etc. Furthermore, the distributed representation can be referred to as an embedding representation. Furthermore, it is also possible to create a distributed representation of the sentence from, for example, the average of the distributed representations of the words in the sentence.

[0054] In distributed representation, methods for calculating similarity include cosine similarity, covariance, unbiased covariance, Pearson's product-moment correlation coefficient, etc. Among these, it is particularly preferable to use cosine similarity.

[0055] The first chapter before revision is often a text that contains a lot of general content that lacks originality, but the second chapter, which has been revised using information from the user's company, can contain a lot of original content. Original information is preferable because it has the potential to ensure novelty.

[0056] Furthermore, the information stored in a company's database can include information about specifications that are more likely to be patentable. As a result, specifications obtained from a specification drafting support system will be specifications that are more likely to be patentable. Furthermore, the information stored in a company's database can include information about specifications that are more likely to be enforced. As a result, specifications obtained from a specification drafting support system will be specifications that are more likely to be enforced. Furthermore, the information stored in a company's database can include information about specifications that have a broad scope of rights. As a result, specifications obtained from a specification drafting support system will be specifications that have a broader scope of rights. Note that the database is not limited to a company's database, and public databases owned by patent offices or courts may also be used. For example, information stored in patent publications, patent bulletins, books, or literature may also be used.

[0057] In this way, a database can contain a wide variety of information. Therefore, it is preferable for users to select a database. By selecting a database and then using the specification preparation support system, appropriate information can be obtained in a short amount of time.

[0058] <Screen transition 3> Pressing the generate button 118 transitions the screen 200 from FIG. 2(A) to FIG. 2(B). Specifically, as shown in FIG. 2(B), the specification division display field 203a is renamed to the pre-revision specification display field 203b, and the revised specification display field 204 is also displayed on the screen 200. It is preferable that the pre-revision specification display field 203b and the revised specification display field 204 are displayed so that they do not overlap. The user can compare the first chapter in the pre-revision specification display field 203b with the second chapter in the revised specification display field 204 and select one or both that best fit the invention summary. For this selection, the screen 200 preferably further includes a selection box 124.

[0059] As shown in Figure 2(B), screen 200 has a combine button 122 and an output button 123. As the previously mentioned buttons can be seen from Figures 1(A) to 2(A), a description of output button 123 will be omitted. The combine button 122 allows the selected chapters to be combined to create a new specification (referred to as a second specification for the sake of distinction), and the specification creation support system can suggest to the user a second specification including the selected chapters.

[0060] <Screen transition 4> Pressing the combine button 122 transitions from the screen 200 in FIG. 2(B) to the screen 200 in FIG. 3(A). Specifically, as shown in FIG. 3(A), the second specification can be displayed in the proposed specification field 205. If there are no problems with the proposed specification, the output button 126 can be pressed, and if there are problems or if the user wants to start over, the back button 125 can be pressed. In particular, if experimental data is received in addition to the invention summary, it is preferable that examples describing the experimental data can be included in the second specification.

[0061] <Screen transition 5> Unlike Figure 3(A), when the combine button 122 is pressed, the screen 200 has a proposed claims field 206 in addition to a proposed specification field 205. If the user has no problems with the second specification and claims, he or she can press the output button 126, or if there are problems or the user wants to start over, he or she can press the back button 125. In particular, if experimental data is received in addition to the invention summary, it is preferable that the second specification include examples describing the experimental data, and that the invention-specifying matters of the examples be described in the claims.

[0062] Such a specification preparation support system allows even inexperienced engineers to prepare high-quality specifications, and also reduces the time required for specification preparation.

[0063] Furthermore, by having experienced intellectual property department personnel use the specification preparation support system, it is possible to obtain specifications that are easier to establish as patents, easier to enforce rights, and broader in scope of rights, which is preferable.Furthermore, it is possible to prevent fatal defects in specification preparation.

[0064] <System configuration example 1 for enjoying the services of the specification drafting support system> The system for receiving the services of the specification drafting support system includes a first information processing device 10, a second information processing device 40, a third information processing device 50, and an information terminal 20. The first information processing device 10, the second information processing device 40, and the information terminal 20 can each be called a component. The first information processing device 10 is a component that enables the specification drafting support system to be implemented, the second information processing device 40 is a component that enables the execution of a language model, and the information terminal 20 is a component used by a user. It is also possible for the information terminal 20 to be a component that enables the specification drafting support system to be implemented. The third information processing device 50 is a database that enables RAG. Furthermore, the above-mentioned RAG may be executed in a database provided in the first information processing device 10 or in a database provided in the information terminal 20.

[0065] 4, the information terminal 20 is connected to a first information processing device 10 via a network 31. The first information processing device 10 is also connected to a third information processing device 50 via the network 31. The first information processing device 10 is also connected to a second information processing device 40 via the network 30.

[0066] <System configuration example 2 for enjoying the services of the specification drafting support system> 5, a system for receiving the services of the specification drafting support system may have a second information processing device 40, a third information processing device 50, and an information terminal 20, without using a first information processing device 10. Although the system does not use the first information processing device 10, the second information processing device 40 is a system for receiving the services that have the functions of the first information processing device 10. In other words, the second information processing device 40 is a component that enables the specification drafting support system to be implemented and enables the execution of a language model, the third information processing device 50 is a component that enables RAG, and the information terminal 20 is a component that is used by the user.

[0067] As shown in FIG. 5, the second information processing device 40 is connected to the information terminal 20 and the third information processing device 50 via a network 32.

[0068] <System configuration example 3 for enjoying the services of the specification drafting support system> Although not shown, as a system for receiving the services of the specification preparation support system, the information terminal 20 may have the functions of the first information processing device 10. In this case, the system for receiving the services of the specification preparation support system includes the second information processing device 40, the third information processing device 50, and the information terminal 20.

[0069] <Examples of information processing devices (hardware) that can implement the specification drafting support system> An example of a first information processing device 10 (sometimes referred to as a first component) that can implement the specification preparation support system will be described with reference to Figure 6. In addition to the first information processing device 10, Figure 6 also shows an information terminal 20, a second information processing device 40, and a third information processing device 50, with arrows indicating the transmission and reception of data. The first information processing device 10 has an input unit 110, a memory unit 120, a processing unit 130, an output unit 140, a transmission path 150, and a display 160. If the contents are displayed only on the information terminal 20, the display on the first information processing device 10 is not necessary.

[0070] [Input section 110] The input unit 110 can have a function of accepting data from outside. For example, the input unit 110 has a function of accepting an invention summary from the information terminal 20. Typically, the input unit 110 has a function of accepting an invention summary entered by a user in an invention summary input field 201 shown in FIG. 1(A).

[0071] Furthermore, the input unit 110 can receive data from the second information processing device 40. Since the second information processing device 40 is a component that enables execution of a language model, an example of the received data is an answer sentence. Furthermore, the input unit 110 can receive data from the third information processing device 50, and since the third information processing device 50 is a database that enables RAG, an example of the received data is information related to a question sentence.

[0072] The input unit 110 can supply the received data to one or more selected from the storage unit 120 , the processing unit 130 and the display 160 via a transmission path 150 .

[0073] [Storage unit 120] The memory unit 120 can store programs and / or data, etc. A representative example of a program is a program that executes the specification drafting support system. The data includes data received by the input unit 110. The data may also include data generated by the processing unit 130 (e.g., calculation results, analysis results, and inference results). It is also possible to store standard prompt phrases as data in the memory unit 120. It is also possible to store standard phrases written by the user. It is also preferable to store past prompts used in the specification drafting support system. Furthermore, the memory unit 120 can be used as a database for RAG. In this case, the third information processing device 50 can be omitted from the system for receiving the services of the specification drafting support system.

[0074] The storage unit 120 has at least one of a volatile memory and a nonvolatile memory. Examples of volatile memories include DRAM (Dynamic Random Access Memory) and SRAM (Static Random Access Memory). Examples of nonvolatile memories include ReRAM (Resistive Random Access Memory, also called resistance change memory), PRAM (Phase change Random Access Memory), FeRAM (Ferroelectric Random Access Memory), MRAM (Magnetoresistive Random Access Memory, also called magnetoresistive memory), and flash memory. The storage unit 120 can be configured with SiLSI (circuits using silicon transistors).

[0075] The storage unit 120 may also include at least one of NOSRAM (registered trademark) and DOSRAM (registered trademark). The storage unit 120 may also include a recording media drive. Examples of recording media drives include a hard disk drive (HDD) and a solid state drive (SSD).

[0076] NOSRAM is an abbreviation for "Nonvolatile Oxide Semiconductor Random Access Memory (RAM)." NOSRAM is a memory technology that uses transistors (also called OS transistors) with metal oxide channel regions, each of which is a two-transistor (2T) or three-transistor (3T) gain cell. OS transistors have extremely low leakage current, the current that flows between the source and drain when they are off. NOSRAM can be used as a nonvolatile memory by retaining a charge corresponding to data within the memory cell using its extremely low leakage current. NOSRAM is particularly suitable for computational processing, which requires repeated large amounts of data read operations, because it can read stored data without destroying it (nondestructive read). NOSRAM memory cells can be stacked. This increases data capacity, enabling it to be used as a large-scale cache memory, main memory, or storage memory, thereby achieving high performance.

[0077] DOSRAM is an abbreviation for "Dynamic Oxide Semiconductor RAM" and refers to RAM with one transistor (1T) one capacitor (1C) type memory cells. DOSRAM is a DRAM formed using OS transistors, and is a memory that temporarily stores information sent from outside. DOSRAM is a memory that takes advantage of the small off-current of OS transistors.

[0078] In this specification and the like, a metal oxide refers to an oxide of a metal in a broad sense. Metal oxides are classified into oxide insulators, oxide conductors (including transparent oxide conductors), oxide semiconductors (also referred to as oxide semiconductors or simply as OSs), and the like. For example, when a metal oxide is used in a semiconductor layer of a transistor, the metal oxide may be referred to as an oxide semiconductor.

[0079] The metal oxide in the channel formation region preferably contains indium (In). An OS transistor using a metal oxide containing indium in the channel formation region has high carrier mobility (electron mobility). The metal oxide in the channel formation region is preferably an oxide semiconductor containing an element M, which will be described later, instead of or in addition to In. The element M is preferably at least one of aluminum (Al), gallium (Ga), and tin (Sn). Other elements that can be used for the element M include boron (B), silicon (Si), titanium (Ti), iron (Fe), nickel (Ni), germanium (Ge), yttrium (Y), zirconium (Zr), molybdenum (Mo), lanthanum (La), cerium (Ce), neodymium (Nd), hafnium (Hf), tantalum (Ta), and tungsten (W). A combination of the elements listed as the element M may be used in the metal oxide. The element M is an element having a high bond energy with oxygen, and the bond energy with oxygen is higher than the bond energy between oxygen and indium. Furthermore, the metal oxide contained in the channel formation region is preferably a metal oxide containing zinc (Zn) instead of or in addition to In. Metal oxides containing zinc may be more likely to crystallize.

[0080] The metal oxide contained in the channel formation region is not limited to the above-mentioned elements, typically, metal oxides containing indium. The metal oxide contained in the channel formation region may be, for example, metal oxides containing zinc but not indium, such as zinc tin oxide and gallium tin oxide, metal oxides containing gallium, or metal oxides containing tin.

[0081] [Processing section 130] The processing unit 130 typically has a function of automatically generating a prompt. Specifically, the processing unit 130 preferably has a function of selecting an arbitrary prompt stored in the storage unit 120 and generating the prompt according to the content of the input unit 110. The processing unit 130 can supply data (e.g., a prompt) generated by itself to one or both of the storage unit 120 and the output unit 140. For example, the data can be supplied to the storage unit 120 via a transmission path 150. Furthermore, the function of automatically generating a prompt in the processing unit 130 preferably has a function of providing the prompt to a language model and acquiring an answer. The processing unit 130 can supply the acquired data (e.g., an answer sentence) to one or both of the storage unit 120 and the output unit 140. For example, the data can be supplied to the output unit 140 via the transmission path 150.

[0082] In addition to the function of automatically generating prompts, the processing unit 130 preferably has the function of executing the processes related to the buttons described in Figures 1(A) to 2(B). By clicking (pressing) each button, the processing unit 130 can execute the process related to that button.

[0083] Furthermore, it is preferable that the processing unit 130 also has a function of determining whether or not the answer has been successfully acquired. It is also possible to have the user confirm the acquired answer, so that the user can determine whether or not the answer has been successfully acquired.

[0084] It is also preferable that the processing unit 130 has a function of constructing display data. It is preferable that the processing unit 130 constructs display data so that the layout is easy for the user to check the display 160. It is also preferable that the processing unit 130 has a function of constructing display data for the user's information terminal 20 instead of or in addition to the display 160. It is preferable that the display data for the display 160 and the user's information terminal 20 are the same data, as this reduces the load on the processing unit 130. Of course, it is also possible for the processing unit 130 to construct data that is different from each other.

[0085] The processing unit 130 has at least an arithmetic circuit. The arithmetic circuit may include, for example, a central processing unit (CPU). The CPU includes an arithmetic unit, a primary cache memory, a secondary cache memory, and the like. The processing unit 130 may also include a graphics processing unit (GPU) in addition to or instead of the CPU. The GPU includes an arithmetic unit, a primary cache memory, a secondary cache memory, and the like. A switch or the like included in the CPU or GPU may include one or both of an OS transistor (a transistor using an oxide semiconductor layer as a channel) and a Si transistor (a transistor using a semiconductor layer containing silicon as a channel).

[0086] The processing unit 130 may have a register and a main memory in addition to a CPU. The register and main memory may also be said to be owned by the CPU. The main memory is capable of sending and receiving data to and from a secondary cache, etc. The main memory has at least one of a volatile memory such as RAM (Random Access Memory) and a non-volatile memory such as ROM (Read Only Memory). The main memory may also have at least one of NOSRAM and DOSRAM. The main memory can have one or both of OS transistors and Si transistors. The configuration of the register and main memory can be understood by replacing CPU in this paragraph with GPU.

[0087] Examples of RAM include DRAM and SRAM. The DRAM or SRAM can be used as a working space for the processing unit 130 by virtually allocating memory space thereto. The operating system, application programs, program modules, program data, lookup tables, and the like stored in the storage unit 120 are loaded into the RAM immediately before execution. The operating system, application programs, program modules, program data, and lookup tables loaded into the RAM can each be accessed by the processing unit 130.

[0088] ROM can store systems that do not require rewriting. Examples of systems that do not require rewriting include BIOS (Basic Input / Output System) and firmware. Examples of ROM include mask ROM, OTPROM (One-Time Programmable Read Only Memory), and EPROM (Erasable Programmable Read Only Memory). Examples of EPROM include UV-EPROM (Ultra-Violet Erasable Programmable Read Only Memory), which allows stored data to be erased by exposure to ultraviolet light, EEPROM (Electrically Erasable Programmable Read Only Memory), and flash memory.

[0089] The processing unit 130 may have a microprocessor such as a DSP (Digital Signal Processor) in addition to a CPU or GPU. Since a DSP is specialized for digital signal processing, it is preferable to install it to control peripheral circuits of the CPU or GPU. The microprocessor may be implemented by a PLD (Programmable Logic Device) that operates on hardware such as an FPGA (Field Programmable Gate Array) or an FPAA (Field Programmable Analog Array). The processing unit 130 may also have a quantum processor. The processing unit 130 can interpret instructions from various programs using a processor such as a quantum processor and can perform various data processing and program control. Programs that can be executed by the processor are stored in at least one of the memory area of ​​the processor and the storage unit 120.

[0090] Because the OS transistors described above have extremely small off-state current, using them as switches for retaining charge (data) flowing into a capacitor element can ensure long-term data retention. By utilizing this characteristic in at least one of the register and cache memory of the processing unit 130, the processing unit can be operated only when necessary, and in other cases, the signal input or power supply to the processing unit 130 can be turned off by saving the information from the previous processing in the corresponding memory element. In other words, the OS transistors enable normally-off computing, thereby enabling the specification preparation support system to consume less power.

[0091] Furthermore, by using a CPU or the like capable of high-speed operation in the processing unit 130, some of the processing performed by the first information processing device 10 can be performed using artificial intelligence (AI). The first information processing device 10 may be provided with an artificial neural network (ANN, hereinafter also simply referred to as a neural network) to enable processing using AI. Since a neural network is realized by a circuit (hardware) or a program (software), it is preferable that the first information processing device 10 has the above circuit or the above program in addition to a CPU capable of high-speed operation.

[0092] In this specification, a neural network refers to a general model that imitates the neural circuit network of a living organism, determines the connection strength between neurons through learning, and has problem-solving capabilities. A neural network has an input layer where information is input, an output layer where information is output, and an intermediate layer (hidden layer) between the input and output layers, and optimizes the weights for the input data to obtain the correct output result.

[0093] In this specification and the like, when discussing neural networks, determining neurons and their weight coefficients from existing information may be referred to as "learning."

[0094] In this specification and the like, constructing a neural network using weighting coefficients obtained by learning and deriving a new conclusion from it may be referred to as "inference."

[0095] [Output section 140] The output unit 140 has a function of outputting data (e.g., the calculation result of the processing unit 130) to the outside. The outside may be one or more selected from the second information processing device 40 and the information terminal 20. For example, the user can check the answer sentence output from the output unit 140 to the information terminal 20.

[0096] [Transmission Line 150] The transmission path 150 has a function of transmitting data. Data can be transmitted and received between the input unit 110, the storage unit 120, the processing unit 130, the output unit 140, and the display 160 via the transmission path 150.

[0097] <<Configuration Example of Second Information Processing Device>> Next, a configuration example of the second information processing device 40 will be described. The second information processing device 40 can perform processing using a language model. For example, it can execute processing using a natural language processing model such as BERT or T5. The second information processing device 40 can execute processing using a general-purpose language processing model that can perform various natural language processing tasks.

[0098] Furthermore, the second information processing device 40 can perform processing using a model that uses an LLM (such as a document generation model or a dialogue model). For example, GPT-3, GPT-3.5, GPT-4, LaMDA, PaLM, Llama2, Llama3, etc. can be used as the LLM. Of course, a model built in-house can be used as the LLM.

[0099] The second information processing device 40 can perform processing such as calculations using the data received from the first information processing device 10. This can reduce the calculation load on the first information processing device 10.

[0100] The provider of the specification preparation support system does not necessarily have to own the second information processing device 40. For example, the service provider can use part of the service provided by another business operator or the like using the second information processing device 40.

[0101] <<Configuration Example of Information Terminal 20>> The information terminal 20 is operated by a user and is a device capable of inputting the information shown in Figures 1(A) to 2(B). For input, the information terminal 20 has an input device such as a keyboard or a speaking device including a microphone. The information terminal 20 can also be called a client computer. It is preferable that the information terminal 20 has an environment in which dedicated application software or a web browser can run. This allows the user to enjoy services using the specification preparation support system.

[0102] 4 and 5 show a desktop computer as an example, but a notebook computer, smartphone, or tablet computer may also be used as the information terminal 20. A tablet computer has a housing that has an input unit (typically a keyboard) that can be folded over and placed on top of the main body. Some tablet computers also have a housing that can be separated from the main body. Because notebook computers, smartphones, and tablet computers are easy to carry, they are suitable for use when a user is away from home and uses the specification preparation support system, which is one aspect of the present invention.

[0103] The information terminal 20 can provide the user with data output by the specification preparation support system or the LLM. In other words, in the specification preparation support system, it is preferable that the user operates the information terminal 20, but does not operate the first information processing device 10 or the second information processing device 40. This configuration can improve security.

[0104] For example, when a provider of a service using the specification preparation support system and a user who receives the service belong to the same organization such as a company, it is preferable that data transmission and reception between the information terminal 20 and the first information processing device 10 be performed using, for example, a network 31 established within the organization. This allows data to be transmitted and received between the information terminal 20 and the second information processing device 40 more securely than when data is transmitted over the Internet. It is also possible to prevent internal company information from leaking to the outside. Alternatively, data transmission and reception between the information terminal 20 and the first information processing device 10 may be performed using a network 30 (for example, the Internet).

[0105] Network 30 The network 30 connects the first information processing device 10 and the second information processing device 40. This allows input data and processed data to be transmitted and received between the two devices. It also distributes the load associated with information processing.

[0106] In this embodiment, the network 30 is mainly described as a computer network that is larger in scale than the network 31. For example, a global network can be used for the network 30. Specifically, the Internet, which is the foundation of the World Wide Web (WWW), can be used.

[0107] Network 31 The network 31 connects the information terminal 20 and the first information processing device 10. This allows data to be sent and received between them. It also distributes the load related to information processing. Furthermore, a service provider can provide a service using the specification preparation support system to a user via the network 31, for example.

[0108] For example, a local network can be used for the network 31. Also, an intranet or an extranet can be used for the network 31. Also, a personal area network (PAN), a local area network (LAN), a campus area network (CAN), a metropolitan area network (MAN), a wide area network (WAN), a global area network (GAN), etc. can be used for the network 31.

[0109] When performing wireless communication, communication standards such as the fourth generation mobile communication system (4G), fifth generation mobile communication system (5G), and sixth generation mobile communication system (6G), or specifications standardized by IEEE such as Wi-Fi (registered trademark) and Bluetooth (registered trademark), can be used as communication protocols or communication technologies.

[0110] Network 32 The second information processing device 40 is connected to the information terminal 20 and the third information processing device 50 via a network 32. This allows input data and processed data to be transmitted and received between the two. It also makes it possible to distribute the load related to information processing. The network 32 can be the same as the network 30 or the network 31 described above.

[0111] According to one aspect of the present invention, it is possible to provide a novel method for operating a specification writing support system. Also, according to another aspect of the present invention, it is possible to provide a method for operating a specification writing support system using a language model with the number of tokens set in the language model or less.

[0112] Next, an example of an operation method of the specification preparation support system according to one embodiment of the present invention will be described with reference to a flowchart.

[0113] <Example 1 of how the specification drafting support system works> A method for preparing specifications, etc., in a specification preparation support system will be described. An example of this method has steps 1 to 10 as shown in Figures 7 and 8. Figure 7 explains steps 1 to 5, and Figure 8 explains steps 6 to 10. Note that (A) after step 5 in Figure 7 means that it continues to (A) shown in Figure 8.

[0114] <Step 1: Receipt of invention summary> In this step, the specification drafting support system accepts an invention summary. Typically, the specification drafting support system accepts an invention summary entered by the user in the invention summary input field 201 shown in FIG. 1(A). The invention summary is preferably written in a natural language, and the user can write the invention summary in a language selected from English, Japanese, and other languages. There are no restrictions on the language of the invention summary, and the above languages ​​may be mixed.

[0115] <Step 2: Generating the first prompt and obtaining the first statement> In this step, the specification drafting support system generates a first prompt including content for creating a first specification corresponding to the invention summary. Typically, the prompt is generated by the processing unit 130 or the like shown in FIG. 6. Furthermore, it is more preferable that the prompt is automatically generated by the processing unit 130 as described above. Of course, it is also possible for the user to create the above prompt without automatic generation. In the case of automatic generation, this step includes, as shown in the figure, from generating the first prompt to obtaining the first specification. In the case of not automatically generating the first specification, instead of this step 2, a step including receiving the first prompt and a step including obtaining the first specification are performed.

[0116] <Step 3: Displaying the first statement> In this step, the specification writing support system displays the first specification. Typically, the first specification is displayed in the specification display field 202 shown in FIG. 1(B). The user can check the first specification. Note that the first specification is within the scope of what the language model, etc. has learned.

[0117] <Step 4> In this step, the specification writing support system determines whether the first specification is appropriate. An appropriate specification means one that satisfies the bibliographic information of the specification. Furthermore, the user can also set the conditions for an appropriate specification, for example, the user can determine whether the specification satisfies the conditions, such as a specification that includes examples. If the specification is determined to be appropriate, the specification writing support system determines that the acquisition was successful (YES) and proceeds to step 5. If the specification is determined to be inappropriate, the specification writing support system determines that the acquisition was unsuccessful (NO), returns to step 2, and executes steps 2 to 4 again. In this case, it is preferable to modify the first prompt in step 2 from the previous one. The user can make the modification.

[0118] <Step 5: Divide the first specification and obtain the first chapter> In this step, the specification writing support system divides the first specification into appropriate sections and acquires any chapters. In this step, at least the first chapter is acquired as the arbitrary chapter. It is preferable that the first chapter be displayed in the specification division display field 203a shown in Figure 2(A) so that the user can confirm the first chapter.

[0119] The first chapter can also be obtained using a language model. For example, by providing a prompt to the language model including the content "Please divide the first specification into embodiments," the specification writing support system can obtain chapters divided into the above embodiments.

[0120] When using a language model, a prompt may be prepared that includes content for obtaining a summary corresponding to each chapter. By providing this prompt to the language model, the specification preparation support system can also obtain the above summary. Typically, the summary can be displayed near the specification division display field 203a shown in Figure 2(A). This is preferable because the user can check the summary and understand the embodiment in a short amount of time. Note that a prompt can be prepared in which the embodiment in the above prompt is replaced with an example.

[0121] <Step 6: First information extraction> In this step, the specification writing support system extracts information about the first chapter from the database. Typically, information extraction is performed by the processing unit 130 shown in FIG. 6 or the like. The database may be owned by the specification writing support system, or it may be owned by an information processing device (component) separate from the specification writing support system. This information may be referred to as "first information" to distinguish it from other information. However, the method for extracting the first information is not limited to the method described here.

[0122] By referring to a large amount of first information, the specification drafting support system can propose a specification that appropriately describes the invention summary. When there is a large amount of first information, it is preferable to reduce the number of characters in the first chapter. This is because the language model processes within the limit of the number of tokens. In particular, when the first chapter corresponds to the core of the invention in the specification, the first information may become enormous. Or, it may be preferable to revise the specification using the large amount of first information. In such cases, it is preferable to reduce the number of characters in the first chapter in the specification drafting support system. This is preferable because it allows more information from the above database to be included in the first chapter.

[0123] On the other hand, if the database contains little first information, it is acceptable to divide the first chapter so that it is longer within the limit of the number of tokens in the language model. It is also possible to treat the specification as a single chapter without dividing it. This is preferable because it allows for revisions that take into account the context of the specification more than when it is divided into chapters.

[0124] <Step 7: Generate the second prompt and get the second chapter> In this step, the specification writing support system creates a prompt for revising the first chapter based on the first information to create a second chapter. The prompt in this step is referred to as the second prompt to distinguish it from other prompts. Typically, this prompt is generated by the processing unit 130 shown in FIG. 6. It is more preferable that the prompt is automatically generated by the processing unit 130 as described above. Of course, the user can create the prompt without automatic generation. When the second prompt is automatically generated, this step includes, as shown in the figure, the generation of the second prompt to the acquisition of the second chapter. When the second prompt is not automatically generated, instead of this step 7, a step including the receipt of the second prompt and a step including the acquisition of the second chapter are performed. Note that when the first chapter is an example, the second chapter is also acquired as an example.

[0125] The language model can be the same as that described in the first prompt generation, and an appropriate number of tokens can be set in the specification writing support system as described in the first prompt generation. The specification writing support system can also generate a prompt specifying the number of output tokens.

[0126] To utilize the above database, first information specific to the company is extracted, and a second chapter revised based on the first information is obtained. In other words, the specification preparation support system can obtain a second chapter revised based on the first information specific to the company. In other words, the second chapter obtained in this step is revised according to the technical information of the company to which the user belongs, and therefore can include unique information. Unique information has the potential to ensure novelty.

[0127] <Step 8: Displaying Chapter 1 and Chapter 2> In this step, the specification writing support system displays Chapter 1 and Chapter 2. Typically, as shown in Figure 2(B), Chapter 1 is displayed in the pre-revision specification display field 203b, and Chapter 2 is displayed in the revised specification display field 204.

[0128] <Step 9: Select either Chapter 1 or Chapter 2> In this step, the user compares the first and second chapters and selects the one that is more suitable for the specification according to the invention summary. In other words, the first chapter may be more suitable for the specification. When the user makes a selection, the first information may be used as a reference. In this case, the first information is also displayed.

[0129] The steps for revising the first chapter into the second chapter (typically steps 6 to 9) and the step for selecting a chapter appropriate for the specification (typically step 10) are repeated as appropriate to obtain a chapter appropriate for the specification.

[0130] <Step 10: Proposal of specification> In this step, the specification writing support system creates a specification by combining all chapters that are appropriate for the specification. This specification is referred to as the second specification to distinguish it from other specifications. The specification writing support system can then suggest the second specification to the user. Typically, the second specification is displayed in the suggested specification field 205, as shown in Figure 3(A). A specification writing support system that has gone through these steps can obtain a high-quality specification.

[0131] <Example 2 of how the specification drafting support system operates> A method for creating specifications, etc. in the specification creation support system will be described below. Although this method is not shown in the figure, the extraction of the first information in step 6 shown in Figure 8 can be obtained using RAG. By using RAG, step 6 shown in Figure 8 can be combined into step 7, making it possible to process in a short time.

[0132] <Steps 1 to 5> Steps 1 to 5 are executed in the same way as in Example 1. Step 6 can also be combined into Step 7 shown below.

[0133] <Step 7: Generate the second prompt using RAG and obtain the second chapter> In the second example of the operation method, in this step the specification writing support system obtains the first information using RAG, and creates a second prompt to revise the first chapter and create a second chapter according to the first information. RAG is a method of searching a database for the question sentence before the language model generates an answer sentence, and can improve the accuracy of the answer sentence of the language model based on the search results.

[0134] The specification writing support system can use the same means for searching the database as in Example 1. The same means for determining whether the data is related to the first chapter as in Example 1 can also be used. The database can also be the same as in Example 1. The database contains a wide variety of information. Therefore, it is preferable for the user to select a database. By selecting a database and then using the specification writing support system, the appropriate first information can be obtained in a short amount of time.

[0135] A database owned by a company can be used as the database. Therefore, first information specific to the company is generated, and a second chapter revised based on the first information can be obtained. In other words, the specification preparation support system can obtain a second chapter revised based on the first information specific to the company. In other words, because the first information is obtained using RAG, it is possible to obtain a second chapter specific to the company, which is considered difficult to obtain using a language model alone.

[0136] In the example of the operation method 2, the specification writing support system preferably uses a summary of the first chapter as the second prompt. For example, if the summary is about a transistor configuration using an oxide semiconductor and silicon oxide, the second prompt can include the content, "Create a second chapter by revising the first chapter, taking into account recent findings about transistors using an oxide semiconductor and silicon oxide."

[0137] If the first information is appropriate in this step, RAG needs to be run once, but if the first information is inappropriate or if the first information needs to be revised, RAG can be run two or more times. In addition, since the content of a specification is diverse, it is preferable to run RAG after dividing the specification and preparing the first chapter, as this makes it possible to obtain appropriate first information.

[0138] <Steps 8 to 10> Steps 8 to 10 are executed in the same manner as in Example 1. A specification preparation support system that has gone through these steps can produce a high-quality specification.

[0139] <Example 3 of operation method of specification drafting support system> A method for preparing a specification and claims in a specification preparation support system will be described. An example of the method has steps 1 to 5 and steps 16 to 23 as shown in Figures 7 and 9. Steps 1 to 5 will be described in Figure 7, and steps 16 to 23 will be described in Figure 9. Note that (A) after step 5 in Figure 7 means that it continues to (A) shown in Figure 9.

[0140] <Steps 1 to 5> Steps 1 to 5 are executed in the same manner as in Example 1. Step 6 can also be combined into step 16 shown below.

[0141] <Step 16: Generate the third prompt using RAG> In Example 3 of the operating method, in this step, the specification drafting support system creates a prompt different from the second prompt. This step is also referred to as a step in which the user creates the prompt, rather than automatically generating it. The prompt in this step is referred to as the third prompt to distinguish it from other prompts. Specifically, this prompt includes content for revising Chapter 1 into Chapter 2 based on one aspect of the first information, and content for creating Chapter 3 based on other aspects of the first information. Because the first information is diverse, information related to Chapter 1 can be used to revisit Chapter 1 into Chapter 2, while information less related to Chapter 1 can be used to create a new Chapter 3. As in Example 1, whether data is related to Chapter 1 can be determined by the similarity of values ​​based on the distributed representation. The first information can be divided according to this similarity. In particular, if experimental data is received in addition to the invention summary, the third prompt can include content for creating a second example corresponding to Chapter 2 and a third example corresponding to Chapter 3.

[0142] <Step 17: Get Chapter 2 and Chapter 3> In this step, the specification writing support system acquires the second and third chapters from the language model as response sentences to the prompt. Note that if the first chapter is an example, the second and third chapters are also acquired as examples.

[0143] The second and third chapters obtained in this step are revised according to the technical information of the company to which the user belongs, and therefore may contain original information. Original information may be novel.

[0144] <Step 18: Displaying Chapters 2 and 3> In this step, the specification writing support system displays Chapters 2 and 3. Typically, Chapters 2 and 3 can be displayed using the revised specification display field 204 in Figure 2(B).

[0145] <Step 19: Select either Chapter 2 or Chapter 3 or both> In this step, the user can compare the second and third chapters and select one or both. When making a selection, the user may refer to the first information. Typically, the selection can be made using the selection button located near the revised specification display field 204 in Figure 2(B).

[0146] The steps for revising the first chapter into the second chapter (typically steps 16 to 18) and the step for selecting a chapter appropriate for the specification (typically step 19) are repeated as appropriate to obtain a chapter appropriate for the specification.

[0147] <Step 20> In this step, the specification drafting support system determines whether both options have been selected in the appropriately repeated step 19. If both options have been selected in one or more cases, the specification drafting support system proceeds to step 21. If only one option has been selected in all cases, the specification drafting support system proceeds to step 22.

[0148] <Step 21: Generate the fourth prompt> In this step, the specification drafting support system generates prompts including content for describing the invention-specifying matters included in Chapter 2 as independent claims and content for describing the invention-specifying matters included in Chapter 3 as dependent claims. Furthermore, for the dependent claims, it is preferable that the prompts include content that will become the reference numbers of the independent claims.

[0149] <Step 22: Generate the fifth prompt> In this step, the specification drafting support system generates a prompt including content for describing the invention-specifying matters included in the second chapter as independent claims.

[0150] <Step 23: Obtaining the third specification including the claims> In this step, the specification drafting support system obtains the specification and claims from the language model as a response to the prompt. This specification is referred to as the third specification to distinguish it from other specifications. Typically, as shown in Figure 3(B), the specification is displayed in the proposed specification field 205, and the claims are displayed in the proposed claims field 206. A specification drafting support system that has gone through these steps can obtain high-quality specifications and claims. In particular, when experimental data is received in addition to the invention summary, it is preferable to propose the invention-specifying features included in the second example, which corresponds to Chapter 2, as independent claims. It is also preferable to propose the invention-specifying features included in the third example, which corresponds to Chapter 3, as dependent claims. In this case, it is preferable to suggest citing the independent claims. [Explanation of symbols]

[0151] 10 First information processing device 20 Information terminal 30 Network 31 Network 32 Network 40 Second information processing device 50 Third information processing device 100 Configuration field 101 Assignment column 102 Effect column 105 Save button 106 Cancel button 107 Generate button 109 Checkboxes 110 Input section 111 Segment button 112 Save button 113 Cancel button 114 Output button 116 Buttons 117 Button 118 Generate button 119 Checkbox 120 Storage section 122 Combine button 123 Output button 124 Selection Box 125 buttons 126 Output button 130 Processing section 140 Output section 150 Transmission Line 160 display 200 screens 201 Invention summary input field 202 Statement display column 203a Specification classification display column 203b Pre-revision specification display column 204 Statement display column 205 Statement 206 Claims

Claims

1. a first component; The first component has a function of accepting an invention summary; a function of generating a first prompt including content for creating a first specification corresponding to the invention summary; retrieving the first specification from a language model as an answer to the first prompt; a function of displaying the first specification; A function of dividing the first specification into chapters; a function of generating a second prompt including a content for obtaining information related to the divided first chapter from a second component and a content for revising the first chapter according to the information to create a second chapter; retrieving the second chapter from the language model as an answer to the second prompt; a function of displaying the first chapter and the second chapter; a function of comparing the first chapter and the second chapter and selecting one that fits the invention summary; and creating a second specification including the selected chapter. Specification creation support system.

2. a first component; The first component has a function of accepting an invention summary; a function of generating a first prompt including content for creating a first specification corresponding to the invention summary; retrieving the first specification from a language model as an answer to the first prompt; a function of displaying the first specification; A function of dividing the first specification into chapters; a function for generating a second prompt including: acquiring a plurality of pieces of information related to the divided first chapter from a second component; revising the first chapter according to the plurality of pieces of information to create a second chapter; and creating a third chapter in addition to the second chapter according to the plurality of pieces of information; a function of obtaining the plurality of pieces of information, the second chapter, and the third chapter from the language model as an answer to the second prompt; a function of displaying the plurality of pieces of information, the second chapter, and the third chapter; a function of comparing the second and third chapters and selecting one or both that are suitable for the invention summary; and creating a second specification including the selected chapter. Specification creation support system.

3. In claim 2, when both chapters suitable for the invention summary are selected, the first component has a function of generating a third prompt including content for describing the invention-specifying matter included in the second chapter as an independent claim in a third specification, and content for describing the invention-specifying matter included in the third chapter as a dependent claim in the third specification, and for describing the dependent claim to cite the independent claim; and obtaining the third specification from the language model as an answer to the third prompt. Specification creation support system.

4. a first component; the first component has a function of receiving an invention summary and first experimental data; a function of generating a first prompt including a content for creating a first specification corresponding to the invention summary and a content for describing the first experimental data in a first example of the first specification; retrieving the first specification from a language model as an answer to the first prompt; a function of displaying the first specification; A function of dividing the first specification into chapters; generating a second prompt including a content for obtaining information related to the first embodiment from a second component and a content for revising the first embodiment according to the first experimental data to create a second embodiment; obtaining the second example from the language model as an answer to the second prompt; A function of displaying the first embodiment and the second embodiment; A function of comparing the first embodiment and the second embodiment and selecting one that is suitable for the outline of the invention; and creating a second specification including the selected embodiment. Specification creation support system.

5. a first component; the first component has a function of receiving an invention summary and first experimental data; a function of generating a first prompt including a content for creating a first specification corresponding to the invention summary and a content for describing the first experimental data in a first example of the first specification; retrieving the first specification from a language model as an answer to the first prompt; a function of displaying the first specification; A function of dividing the first specification into chapters; a function for generating a second prompt including: acquiring a plurality of pieces of information related to the first embodiment from a second component; revising the first embodiment to create a second embodiment according to the plurality of pieces of information; and creating a third embodiment in addition to the second embodiment according to the plurality of pieces of information; a function of obtaining the plurality of pieces of information, the second example, and the third example from the language model as an answer to the second prompt; a function of displaying the plurality of pieces of information, the second embodiment, and the third embodiment; A function of comparing the second embodiment and the third embodiment and selecting one or both that are suitable for the outline of the invention; and creating a second specification including the selected embodiment. Specification creation support system.

6. In claim 5, when both embodiments suitable for the invention outline are selected, the first component has a function of generating a third prompt including content for describing the invention-specifying matters included in the second embodiment as an independent claim in a third specification, and content for describing the invention-specifying matters included in the third embodiment as a dependent claim in the third specification, and for describing the dependent claim to cite the independent claim; and obtaining the third specification from the language model as an answer to the third prompt. Specification creation support system.

7. In any one of claims 1 to 6, the invention summary includes a description of the problem and a description of the effect. Specification creation support system.

Citation Information

Patent Citations

  • Preparation support system for specification of application for patent or specification of application for utility model registration

    JP2013080279A