Document creation support device, document creation support method, and document creation support program

Abstract

The present invention provides a document creation support technology for creating application documents for patent applications using prior knowledge. [Solution] One aspect of the present disclosure relates to a document creation support device having a reception unit that receives claims, drawings, and creation policy information for an invention that is the subject of an application; a knowledge extraction unit that extracts knowledge data corresponding to the creation policy information; a prompt creation unit that creates a prompt to cause a generative model to create a draft specification for the invention based on the claims, drawings, and knowledge data; and a model interface unit that inputs the prompt into the generative model and retrieves the draft specification from the generative model.

Description

[Technical Field]

[0001] The present disclosure relates to a document creation support device, a document creation support method, and a document creation support program. [Background technology]

[0002] Large-scale language models (LLMs) such as ChatGPT are being increasingly utilized in society and the economy. A typical large-scale language model receives a question as input and generates a response to the question. Such a question is called a "prompt," and the quality of the response generated by a large-scale language model can vary depending on how the question or prompt is written. For this reason, research and development is underway to develop technology (called "prompting technology") that allows large-scale language models to generate more useful responses.

[0003] In addition to prompting technology, research and development is also underway on generative AI (artificial intelligence)-related technologies to improve the accuracy of answer sentences from large-scale language models. One such generative AI-related technology is Retrieval Augmented Generation (RAG), which causes a large-scale language model to generate answer sentences by referencing data stored in a database. Such databases store data that is not used to train the large-scale language model. Data related to a question sentence or a prompt is extracted from the database, and the extracted data is used to generate the question sentence or the prompt. This enables the large-scale language model to generate answer sentences using not only the data used to train itself, but also the untrained data stored in the database.

[0004] The use of generative models such as large-scale language models is also being attempted in intellectual property practice. For example, attempts are being made to use large-scale language models to support the creation of application documents (specifications, claims, drawings, etc.) for patent applications. For example, Japanese Patent Publication No. 7574529 describes how providing various support information tailored to the user's wishes makes it easy to create high-quality documents to be submitted to the patent office. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 7574529 Summary of the Invention [Problem to be solved by the invention]

[0006] In general, it is considered desirable that application documents such as specifications, drawings, and claims be written in accordance with the methodology appropriate to the technical field of the application. In fact, books on how to write patent specifications are published according to the technical field.

[0007] In addition, most patent attorneys have their own technical areas of expertise and are likely to have unique experience and knowledge regarding the preparation of application documents in their areas of expertise. For example, a patent attorney who handles many foreign cases may be able to prepare claims, specifications, and / or drawings that are suitable for examination in foreign countries.

[0008] Application documents such as specifications, drawings, and claims can be prepared using such preparation techniques as prior knowledge. Technology for preparing patent application documents using such prior knowledge is not described in the prior art, including Japanese Patent Publication No. 7574529.

[0009] One object of the present disclosure is to provide a document creation support technology for creating application documents for patent applications using prior knowledge. [Means for solving the problem]

[0010] One aspect of the present disclosure relates to a document creation support device having a reception unit that receives claims, drawings, and creation policy information for an invention that is the subject of an application; a knowledge extraction unit that extracts knowledge data corresponding to the creation policy information; a prompt creation unit that creates a prompt to cause a generative model to create a draft specification for the invention based on the claims, the drawings, and the knowledge data; and a model interface unit that inputs the prompt into the generative model and retrieves the draft specification from the generative model. [Effects of the Invention]

[0011] According to the present disclosure, a document creation support technology can be provided for creating application documents for patent applications using prior knowledge. [Brief explanation of the drawings]

[0012] [Figure 1] FIG. 1 is a schematic diagram illustrating a document creation support process according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a block diagram illustrating a hardware configuration of a document creation support device according to an embodiment of the present disclosure. [Figure 3] FIG. 3 is a block diagram illustrating a functional configuration of a document creation support device according to an embodiment of the present disclosure. [Figure 4] FIG. 4 is a diagram illustrating a hierarchical technology classification tree according to one embodiment of the present disclosure. [Figure 5] FIG. 5 is a diagram showing a knowledge database storing specification templates according to technical fields according to an embodiment of the present disclosure. [Figure 6] FIG. 6 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 7] FIG. 7 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 8] FIG. 8 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 9]FIG. 9 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 10] FIG. 10 is a diagram illustrating a knowledge database storing sample data according to technical fields according to an embodiment of the present disclosure. [Figure 11] FIG. 11 is a diagram illustrating a hierarchical technology classification tree according to one embodiment of the present disclosure. [Figure 12] FIG. 12 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 13] FIG. 13 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 14] FIG. 14 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 15] FIG. 15 illustrates a specification writing prompt according to one embodiment of the present disclosure. [Figure 16] FIG. 16 is a diagram illustrating specification creation using Few Shot prompting according to one embodiment of the present disclosure. [Figure 17] FIG. 17 is a diagram illustrating a knowledge database storing specification templates according to users according to an embodiment of the present disclosure. [Figure 18] FIG. 18 illustrates a specification creation prompt according to one embodiment of the present disclosure. [Figure 19] FIG. 19 illustrates a specification creation prompt according to one embodiment of the present disclosure. [Figure 20] FIG. 20 illustrates a specification creation prompt according to one embodiment of the present disclosure. [Figure 21] FIG. 21 illustrates a specification creation prompt according to one embodiment of the present disclosure. [Figure 22] FIG. 22 is a diagram illustrating a knowledge database storing sample data according to users according to an embodiment of the present disclosure. [Figure 23] FIG. 23 is a diagram illustrating specification creation using Few Shot prompting according to one embodiment of the present disclosure. [Figure 24]FIG. 24 illustrates a specification creation prompt according to one embodiment of the present disclosure. [Figure 25] 25A and 25B are diagrams illustrating example outputs according to one embodiment of the present disclosure. [Figure 26] 26A and 26B are diagrams illustrating example outputs according to one embodiment of the present disclosure. [Figure 27] FIG. 27 is a flowchart illustrating a document creation support process according to an embodiment of the present disclosure. [Figure 28] FIG. 28 is a diagram showing a setting screen of a specification writing tool according to an embodiment of the present disclosure. [Figure 29] FIG. 29 is a diagram illustrating an output screen of a specification writing tool according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0014] In the following embodiment, a document creation support device for patent application documents is disclosed.

[0015] [Summary of the Disclosure] In the embodiment described below, when the document creation support device 100 receives an instruction to create a specification for an invention that is the subject of a patent application from the user terminal 20, it acquires knowledge data related to the creation of the specification from the knowledge database 40, and creates a prompt to be input to the generative model 30 by referring to the knowledge data.The document creation support device 100 then inputs the created prompt to the generative model 30, acquires a draft specification from the generative model 30, and provides it to the user terminal 20.

[0016] 1, the document creation support device 100 according to this embodiment acquires claims and drawings for an invention to be applied for from a user terminal 20, and upon receiving creation policy information indicating a specification creation policy, transmits a search request to the knowledge database 40 to acquire knowledge data related to the received claims, drawings, and / or creation policy information, and obtains search results from the knowledge database 40. The creation policy information here may indicate, for example, the technical field of the invention to be applied for, the user who sent the specification creation instruction, etc.

[0017] The knowledge database 40 stores knowledge data related to specification creation, such as specification templates corresponding to each technical field, claims corresponding to each technical field, sample data of drawings and specifications, specification templates corresponding to each user, claims corresponding to each user, sample data of drawings and specifications, etc. Typically, this knowledge data can be structured, quantified, etc.

[0018] When the document creation support device 100 acquires knowledge data related to the claims, drawings, and / or drafting policy information as search results from the knowledge database 40, the document creation support device 100 uses the acquired knowledge data to create a question or prompt for the generative model 30 to generate a draft specification of the application in accordance with the claims, drawings, and / or drafting policy information, as will be described in more detail below. The document creation support device 100 then transmits the created prompt to the generative model 30 and acquires the draft specification generated by the generative model 30. Typically, the generative model 30 may be any large-scale language model, such as a GPT-based model, that generates a response as text data from a prompt, such as document data (e.g., text data, WORD data, PDF data), or image data.

[0019] When the draft specification is acquired from the generative model 30, the document creation support device 100 provides the acquired draft specification to the user terminal 20. For example, the document creation support device 100 may convert the draft specification, which is text data acquired from the generative model 30, into any document data such as WORD data, and transmit the WORD data of the draft specification to the user terminal 20.

[0020] Here, the document creation support device 100 is realized by a computing device such as a server, a personal computer, a smartphone, or a tablet, and may have a hardware configuration such as that shown in Fig. 2. That is, the document creation support device 100 has a memory 101, a storage 102, a processor 103, a communication device 104, a user interface (UI) device 105, and a peripheral device interface 106, which are interconnected via a bus B.

[0021] Programs or instructions for realizing various functions and processes described below in the document creation support device 100 are stored in a removable storage medium such as a CD-ROM (Compact Disk-Read Only Memory) or flash memory, and are stored in the memory 101 and / or storage 102 via the peripheral device interface 106. However, the programs or instructions do not necessarily need to be installed from the storage medium, and may be downloaded from any external device via a network or the like.

[0022] The memory 101 is realized by a random access memory, a static memory, or the like, and when a program or instruction is activated, the memory 101 reads the program or instruction, data, and the like from the storage 102 and stores them.

[0023] The storage 102 is realized by a hard disk drive or the like, and stores installed programs or instructions as well as files, data, etc. used to execute the programs or instructions. The memory 101, the storage 102, and the removable storage medium may be collectively referred to as a non-transitory storage medium.

[0024] The processor 103 may be realized by one or more CPUs (Central Processing Units), GPUs (Graphics Processing Units), processing circuitry, etc., which may be composed of one or more processor cores, and performs various functions and processes of the document creation support device 100 described below in accordance with programs, instructions, data such as parameters required to execute the programs or instructions, etc. stored in the memory 101.

[0025] The communication device 104 is realized by various communication circuits that execute communication processes with external devices, the Internet, a LAN (Local Area Network), a WLAN (Wireless LAN), WiFi, or other communication networks.

[0026] The user interface (UI) 105 may be composed of input devices such as a keyboard, mouse, camera, microphone, etc., output devices such as a display, speaker, headset, printer, etc., and input / output devices such as a touch panel, and realizes an interface between the user and the document creation support device 100. For example, the user operates the document creation support device 100 by operating a GUI (Graphical User Interface) displayed on a display or touch panel using a keyboard, mouse, etc.

[0027] The peripheral device interface 106 provides various interfaces for connecting to peripheral devices, such as a USB (Universal Serial Bus) port and an HDMI (High-Definition Multimedia Interface) port.

[0028] However, the above-described hardware configuration is merely an example, and the document creation support device 100 according to the present disclosure may be realized by any other appropriate hardware configuration.

[0029] [Document creation support device] Next, a document creation support device 100 according to an embodiment of the present disclosure will be described. FIG. 3 is a block diagram showing a functional configuration of the document creation support device 100 according to an embodiment of the present disclosure. As shown in FIG. 3, the document creation support device 100 includes a reception unit 110, a knowledge extraction unit 120, a prompt creation unit 130, and a model interface unit 140. For example, one or more functional units of the reception unit 110, the knowledge extraction unit 120, the prompt creation unit 130, and the model interface unit 140 may be realized by one or more processors 103 executing one or more programs or instructions stored in memory 101.

[0030] The receiving unit 110 receives claims, drawings, and drafting policy information for an invention to be applied for. For example, when a user creates claims and drawings for filing a patent application for the invention to be applied for on the user terminal 20 and transmits a specification creation instruction to the document creation device 100 to cause the generative model 30 to generate a draft specification corresponding to the claims and drawings in accordance with drafting policy information for creating a specification, the receiving unit 110 receives the specification creation instruction indicating the claims, drawings, and drafting policy information.

[0031] Here, the claims and drawings do not need to be provided separately, but may be provided together in a single document file, for example. The inventions described in the claims are preferably described by claims, similar to those described in patent application documents, and may be described in independent or dependent form. However, the claims are not necessarily limited to this, and may be described in any form that identifies the invention, such as a list of constituent features.

[0032] Furthermore, the drawings may preferably be similar to those described in patent application documents, such as design drawings, image diagrams, block diagrams, flow charts, screen transition diagrams, program code, etc. However, the drawings are not necessarily limited to these and may be described in any format that can visually explain the invention that is the subject of the application.

[0033] The claims and drawings here are assumed to be created independently by the user preparing the patent application documents for the invention that is the subject of the application, but the claims and / or drawings according to the present disclosure are not necessarily limited to those created independently by the user, and may be created by the user performing editing tasks such as adding, deleting, and correcting the draft description generated by the generative model 30.

[0034] The drafting policy information indicates how the invention to be applied for should be described in the specification. For example, it may be in any format indicating the overall structure of the specification, what should be described in each section of the specification, etc. The specification must comply with the description requirements set forth in the Patent Act and Enforcement Regulations, etc., and guidelines such as the Examination Guidelines. It is also considered desirable for the specification to follow a format appropriate to the technical field. For example, in the field of information processing, it is common to describe the invention to be applied for with reference to a flowchart, or to describe the hardware configuration of an information processing device such as a computer that realizes the invention to be applied for. Furthermore, in the technical fields of electrical and mechanical engineering, if the device to be applied for is a component (e.g., a vehicle component), it is desirable to disclose not only the component but also the device (e.g., a vehicle) in which the component is installed.

[0035] For this reason, creation policy information may be prepared corresponding to each technical field, and knowledge data related to specification creation corresponding to each technical field may be registered and stored in the knowledge database 40. In this case, the user may notify the document creation support device 100 of the technical field of the invention to be applied for as creation policy information, and the document creation support device 100 may create a draft specification using the knowledge data corresponding to the notified technical field.

[0036] For example, a user may specify the technical field of the invention to be applied for while searching a hierarchical technology classification tree such as that shown in Fig. 4. In the illustrated example, if the user first selects "information processing" and then selects "AI-related invention," the reception unit 110 may display a "specification template for an invention related to training processing," a "specification template for an invention related to inference processing," a "specification template for an invention related to generation AI," etc. on the user terminal 20, and allow the user to select one of the specification templates.

[0037] Furthermore, the way specifications are written may differ depending on the user, such as a patent attorney. Each patent attorney generally writes patent application documents for the invention being filed in accordance with a specification writing style based on their own background, field of expertise, knowledge, experience, etc. It is likely that even for the same invention, the way patent application documents are written may differ depending on the patent attorney.

[0038] For this reason, creation policy information may be prepared for each user, and knowledge data corresponding to each user may be registered and stored in the knowledge database 40. Typically, a user logs in to the document creation support device 100 using their own account, and the document creation support device 100 may create a draft specification using knowledge data corresponding to the logged-in user.

[0039] The knowledge extraction unit 120 extracts knowledge data corresponding to the creation policy information. For example, the knowledge data is data that is registered and stored in advance in the knowledge database 40 and is to be referenced by the generative model 30 when creating a draft specification. For example, the knowledge database 40 may store, as knowledge data, specification templates and / or sample data of patent application documents such as claims, drawings, and specifications corresponding to each technical field, such as electrical machinery, machinery, information processing, communications, chemistry, and medicine. Additionally or alternatively, the knowledge database 40 may store, as knowledge data, specification templates and / or sample data of patent application documents such as claims, drawings, and specifications corresponding to each user of the document creation support device 100.

[0040] In one embodiment, the way of writing a specification according to such a technical field may be described in a specification template, and the specification may be generated by the generative model 30 by referring to this specification template. The specification template here indicates a policy for creating a specification for each technical field, and may be, for example, any description showing how a specification should be created based on the claims and drawings.

[0041] A specification typically consists of sections such as "title of the invention," "technical field," "background art," "prior art documents," "summary of the invention," "problem to be solved by the invention," "means for solving the problem," "effects of the invention," "brief description of the drawings," "modes for carrying out the invention," and "explanation of symbols." For example, a specification template may indicate how each section of the specification, such as "title of the invention," "technical field," "background art," "prior art documents," "summary of the invention," "problem to be solved by the invention," "means for solving the problem," "effects of the invention," "brief description of the drawings," "modes for carrying out the invention," and "explanation of symbols," should be written based on the claims and drawings.

[0042] In one embodiment, instead of or in addition to such specification templates according to the technical fields, sample data of patent application documents such as claims, drawings, and specifications for each technical field may be prepared, and the generative model 30 may generate a specification by referring to the sample data. For example, such sample data may be electronic data such as application document data, publications, and patent publications for patent applications in each technical field, and may be selected by an administrator, a user, or any other third party of the document creation support device 100.

[0043] In one embodiment, the way of writing a specification for each patent attorney may be described in a specification template, and the specification may be generated by the generative model 30 by referring to this specification template. The specification template here indicates the policy for each patent attorney to prepare a specification, and may be, for example, a description of how a specification should be prepared based on the claims and drawings.

[0044] In one embodiment, instead of or in addition to such specification templates corresponding to patent attorneys, sample data of application documents such as claims, drawings, and specifications prepared by each patent attorney may be prepared, and the generative model 30 may generate a specification by referring to the sample data. For example, such sample data may be electronic data such as application document data, publications, and patent publications of patent applications prepared by each patent attorney, and may be selected by an administrator, a user, or any other third party of the document creation support device 100.

[0045] The knowledge database 40 may store, for example, specification templates and / or sample data of patent application documents for each technical field and / or each user. Specifically, these specification templates and / or sample data may be processed by any known data processing for RAG, such as chunking, vectorization, structuring, etc.

[0046] The prompt creation unit 130 creates a prompt for causing the generative model 30 to create a draft specification of the invention based on the claims, drawings, and knowledge data. Specifically, the prompt creation unit 130 creates a prompt for causing the generative model 30 to create a draft specification of the application based on the claims and drawings received from the user terminal 20, in accordance with knowledge data corresponding to the creation policy information extracted from the knowledge database 40.

[0047] For example, as shown in Figure 5, if specification templates for each technical field such as electrical machinery, machinery, and information processing are registered and stored in the knowledge database 40, the prompt creation unit 130 creates a prompt for creating a specification from the claims and drawings in accordance with the specification template corresponding to the technical field of the invention being applied for, as indicated in the creation policy information received from the user terminal 20.

[0048] Specifically, when the preparation policy information indicates the technical field of the invention to be applied for, the knowledge extraction unit 120 extracts, as knowledge data, a specification template corresponding to the technical field from the knowledge database 40, and the prompt creation unit 130 creates a prompt for creating a draft specification based on the claims and drawings in accordance with the extracted specification template. For example, if the invention to be applied for is an "invention relating to inference processing" of an "AI-related invention," the prompt creation unit 130 writes the specification template for the "invention relating to inference processing" extracted from the knowledge database 40 in the prompt, and instructs the user in the prompt to create a draft specification for the invention to be applied for based on the claims and drawings received from the user terminal 20.

[0049] Specifically, the prompt creation unit 130 may create prompts such as those shown in Figures 6 to 9. That is, as shown in Figure 6, the prompt creation unit 130 may create a prompt for creating a draft specification for the invention to be applied for in accordance with the specification template for an "invention relating to inference processing" extracted from the knowledge database 40, based on the claims file "Claims.docx" and the drawings file "Drawings.pdf" attached to the prompt.

[0050] For example, the specification template for an "invention related to inference processing" describes the description guidelines for each item in the specification, as shown in Figures 7 to 9. For example, for the "title of the invention," the description guideline states, "list the nouns at the end of the inventions related to each independent claim." Furthermore, for the "modes for carrying out the invention," the description guideline states, for example, that the description requirements of Article 36 of the Patent Act must be met, that the invention related to each claim described in the claims must be explained to the extent that a person skilled in the art can carry out the invention with reference to the attached drawings, that a trained machine learning model must be explained for use in the inference processing, and that a trained machine learning model must be explained for training.

[0051] The above-described specification template is merely an example, and the specification may be written in any other appropriate description format that indicates the specification creation policy. Furthermore, in the above-described embodiment, the prompt creation unit 130 created a prompt to cause the generative model 30 to create a draft specification using a specification template corresponding to the technical field of the invention to be applied for, but the prompt creation according to the present disclosure is not limited to this.

[0052] In one embodiment, the prompt creation unit 130 may create a prompt for creating a draft specification based on the claims and drawings by referring to sample data of claims, drawings, and specifications corresponding to the technical field as knowledge data. That is, when the technical field of the invention to be applied for is received as creation policy information from the user terminal 20, the knowledge extraction unit 120 may extract sample data of claims, drawings, and specifications corresponding to the technical field as knowledge data from the knowledge database 40, and provide the extracted sample data to the prompt creation unit 130.

[0053] In this case, sample data of claims, drawings, and specifications corresponding to each technical field such as electrical machinery, machinery, and information processing may be registered and stored in the knowledge database 40 as shown in Fig. 10. Such sample data may be past patent application documents, patent publications, patent bulletins, etc. Such sample data may be specified by the user, or may be selected by an administrator of the document creation support device 100, for example.

[0054] For example, a user may specify the technical field of the invention to be applied for while searching a hierarchical technology classification tree such as that shown in Figure 11. In the illustrated example, the user first selects "information processing" and then selects "AI-related invention." The reception unit 110 may then display "sample data for inventions related to training processing," "sample data for inventions related to inference processing," "sample data for inventions related to generation AI," etc. on the user terminal 20, allowing the user to select one of the sample data. This allows a draft specification to be created using sample data similar to the invention to be applied for.

[0055] A plurality of sample data corresponding to each technical field may be registered and stored in the knowledge database 40, and the prompt creation unit 130 may use one or more of these sample data to create a prompt for causing the generative model 30 to create a draft specification based on the claims and drawings. In addition, the sample data used to create the prompt may be one designated by the user from the plurality of sample data corresponding to each technical field stored in the knowledge database 40.

[0056] Specifically, the prompt creation unit 130 may create prompts such as those shown in Figures 12 to 15. That is, as shown in Figure 12, the prompt creation unit 130 may create a prompt for creating a draft specification for an invention to be applied for in accordance with sample data for an "invention relating to inference processing" extracted from the knowledge database 40, based on a claims file "Claims.docx" and a drawings file "Drawings.pdf" attached to the prompt.

[0057] For example, as shown in Figures 12 to 15, sample data for an "invention related to inference processing" is composed of a specification, claims, and drawings as patent application documents for the "invention related to inference processing." Note that the above-described sample data is merely an example and may be written in any other appropriate description format for patent application documents for inventions in the technical field indicated in the creation policy information. Furthermore, in the above-described embodiment, the prompt creation unit 130 created a prompt to cause the generative model 30 to create a draft specification using sample data corresponding to the technical field of the invention to be applied for, but prompt creation according to the present disclosure is not necessarily limited to this.

[0058] In this embodiment, the prompt creation unit 130 uses sample data of claims, drawings, and specifications corresponding to the technical field of the invention to be applied for, which are registered and stored in the knowledge database 40, to create a prompt for the generative model 30 to create a draft specification from the claims and drawings of the invention to be applied for in the technical field. In other words, the prompt creation unit 130 may use a few-shot prompting method to create a draft specification from the claims and drawings of the invention to be applied for.

[0059] In other words, according to the few-shot prompting technique, as shown in FIG. 16, by utilizing the relationship between sample claims #1 to #N and sample drawings #1 to #N registered and stored in the knowledge database 40 and corresponding sample specifications #1 to #N, a corresponding draft specification can be created from the claims and drawings of the invention that is the subject of the application.

[0060] In the above-described embodiment, the creation policy information indicates the technical field of the invention to be applied for, but in other embodiments, the creation policy information may indicate a user of the document creation support device 100. In this case, the prompt creation unit 130 may create a prompt for causing the generative model 30 to create a corresponding draft specification based on the claims and drawings of the invention to be applied for, in accordance with a specification template corresponding to the user. Specifically, the knowledge extraction unit 120 extracts a specification template corresponding to the user as knowledge data from the knowledge database 40, and the prompt creation unit 130 creates a prompt for causing the generative model 30 to create a draft specification based on the claims and drawings of the invention to be applied for, in accordance with the specification template corresponding to the extracted user.

[0061] For example, as shown in FIG. 17, when a specification template for each user is registered and stored in the knowledge database 40, the prompt creation unit 130 may create a prompt for creating a specification from the claims and drawings in accordance with the specification template corresponding to the creation policy information indicating the user logged in to the document creation support device 100.

[0062] Specifically, when the preparation policy information indicates patent attorney A, the knowledge extraction unit 120 extracts a specification template corresponding to patent attorney A as knowledge data from the knowledge database 40, and the prompt creation unit 130 creates a prompt for creating a draft specification based on the claims and drawings in accordance with the extracted specification template. For example, the prompt creation unit 130 writes the specification template of patent attorney A extracted from the knowledge database 40 in the prompt, and instructs the user in the prompt to create a draft specification for the invention to be applied for based on the claims and drawings received from the user terminal 20.

[0063] Specifically, the prompt creation unit 130 may create prompts such as those shown in Figures 18 to 21. That is, as shown in Figure 18, the prompt creation unit 130 may create a prompt for causing patent attorney A to create a draft specification for the invention to be applied for in accordance with the specification template of patent attorney A extracted from the knowledge database 40, based on the claims file "Claims.docx" and the drawings file "Drawings.pdf" attached to the prompt.

[0064] For example, patent attorney A's specification template, as shown in Figures 19 to 21, describes the description guidelines for each item of the specification. For example, for the "Mode for Carrying Out the Invention," the description guidelines include: satisfying the description requirements of Article 36 of the Patent Act; explaining the invention according to each claim described in the claims with reference to the attached drawings; explaining an invention according to each independent claim with an independent embodiment; explaining the inclusion and connection relationship between the device and the system when an apparatus and a system including the apparatus are described in the claims and / or drawings; explaining the configuration of the apparatus, the operation of each component, and the interrelationship between the components; explaining the operation of the subject of the operation according to a flowchart for an apparatus invention;

[0065] The above-described specification template is merely an example, and the specification may be written in any other appropriate description format that indicates the specification creation policy. Furthermore, in the above-described embodiment, the prompt creation unit 130 creates a prompt for causing the generative model 30 to create a draft specification using a specification template corresponding to each user, but the prompt creation according to the present disclosure is not necessarily limited to this.

[0066] In one embodiment, the prompt creation unit 130 may create a prompt for creating a draft specification based on the claims and drawings by referring to sample data of claims, drawings, and specifications created by a user as knowledge data. That is, when receiving user identification information as creation policy information from the user terminal 20, the knowledge extraction unit 120 may extract sample data of claims, drawings, and specifications corresponding to the user of the identification information as knowledge data from the knowledge database 40, and provide the extracted sample data to the prompt creation unit 130.

[0067] In this case, sample data of claims, drawings, and specifications corresponding to each user may be registered and stored in the knowledge database 40, as shown in FIG. 22. Such sample data may be, for example, past patent application documents, patent publications, and patent bulletins created by the user. Such sample data may be specified by the user, or may be selected by an administrator of the document creation support device 100, for example. Furthermore, the prompt creation unit 130 may display the user's sample data extracted from the knowledge database 40 on the user terminal 20, and allow the user to select any sample data to be referenced by the generative model 30 in order to create a draft specification of the invention to be applied for.

[0068] A plurality of sample data corresponding to each user may be registered and stored in the knowledge database 40, and the prompt creation unit 130 may use one or more of these sample data to create a prompt for causing the generative model 30 to create a draft specification based on the claims and drawings. In addition, the sample data used to create the prompt may be one designated by the user from among the plurality of sample data corresponding to each user stored in the knowledge database 40.

[0069] In this embodiment, the prompt creation unit 130 uses sample data of claims, drawings, and specifications corresponding to each user, which are registered and stored in the knowledge database 40, to create prompts for causing the generative model 30 to create a draft specification from the claims and drawings of the user of the document creation support device 100. In other words, the prompt creation unit 130 uses a few-shot prompting technique to create a draft specification from the claims and drawings of the invention to be applied for.

[0070] That is, according to the few-shot prompting method, as shown in FIG. 23, by utilizing the relationship between the sample claims #1 to #N and sample drawings #1 to #N registered and stored in the knowledge database 40 and the corresponding sample specifications #1 to #N, a corresponding draft specification is created from the claims and drawings of the invention that is the subject of the application.

[0071] Various embodiments of the specification preparation by the document preparation support device 100 have been described above. However, each embodiment is not necessarily implemented alone, and two or more of these embodiments may be combined and implemented. For example, the preparation policy information may indicate both the technical field of the invention to be applied for and the user's identification information. In this case, the prompt creation unit 130 may create a prompt for preparing a specification draft based on the claims and drawings using a specification template and / or sample data corresponding to the technical field and the user extracted from the knowledge database 40. This allows the user to prepare a specification draft according to the same preparation policy as a specification previously prepared in the technical field of the invention to be applied for, and may enable the user to obtain a specification draft that is more suitable for the technical field and the user.

[0072] In one embodiment, the prompt creation unit 130 may include in the prompt an instruction to cause the generative model 30 to explain to which portion of the draft specification the invention according to each claim or each constituent feature of the invention in the claims corresponds. For example, as shown in FIG. 24, the prompt creation unit 130 may add an instruction to the prompt such as, "Please clearly indicate in which wording or paragraph of the draft specification the invention according to each claim or each constituent feature of each invention included in the claims is described."

[0073] For example, if the claims include the invention according to claim 1 as shown in Figure 25A, a note is added to the specification as shown in Figure 25B that paragraph 0036 of the specification generated by the generative model 30 corresponds to the "processing unit" component of the invention according to claim 1. In other words, the specification generated by the generative model 30 can indicate the portions of the specification corresponding to the inventions or components of each claim in the claims. This allows the specification to be generated based on the claims and drawings, and allows the user to confirm which portions of the specification correspond to the inventions or components of each claim in the claims.

[0074] In addition, the portions of the inventions or constituent features of each claim in the claims of the draft specification may be annotated to the corresponding claims or constituent features, as shown in Figures 26A and 26B. Also, by combining Figures 25 and 26, the corresponding portions of the specification and the corresponding claims or constituent features in the claims may be annotated to the claims and the draft specification, respectively.

[0075] According to the document creation support device 100 of the above-described embodiment, it is possible to have the generative model 30 generate a draft specification corresponding to the claims and drawings by utilizing prior knowledge data, and to obtain a draft specification corresponding to a specific technical field and / or user.

[0076] [Document creation support processing] Next, a document creation support process according to an embodiment of the present disclosure will be described with reference to Fig. 27. Fig. 27 is a flowchart showing the document creation support process according to an embodiment of the present disclosure. The document creation support process is executed by the document creation support device 100 described above, and more specifically, may be realized by one or more processors 103 of the document creation support device 100 executing one or more programs or instructions stored in one or more memories 101.

[0077] As shown in FIG. 27, in step S101, the document creation support device 100 receives claims, drawings, and drafting policy information for the invention to be applied for. Here, the claims describe one or more claims that claim a patent right for the invention to be applied for, and may include independent claims and dependent claims. The drawings are used to visually explain the invention to be applied for, and are referred to when explaining the invention to be applied for in the specification. The drafting policy information indicates the policy for drafting the specification, and may indicate, for example, the technical field of the invention to be applied for and / or the user who will be drafting the patent application documents.

[0078] For example, on the setting screen of the specification writing tool provided by the document creation support device 100 displayed on the user terminal 20 as shown in Fig. 28, user A of the user terminal 20 inputs the claims "ProposedClaim#1.docx" and drawings "ProposedFigure#1.pdf" for the invention to be applied for. Note that the claims and drawings do not need to be provided as separate files, and may be provided as a file containing the claims and drawings, for example.

[0079] As shown in the figure, user A can input the technical field of the invention to be applied for by sequentially selecting the major category "information processing," the medium category "AI-related invention," and the minor category "inference processing" on the setting screen. This allows the document creation support device 100 to accept the technical field "inference processing" and the user "user A" as creation policy information from the setting screen.

[0080] In step S102, the document creation support device 100 extracts knowledge data corresponding to the creation policy information. For the setting information input on the setting screen shown in Fig. 28, the document creation support device 100 can extract, from the knowledge database 40, a specification template and / or sample data corresponding to "invention related to inference processing" and "user A" input as the creation policy information.

[0081] In step S103, the document creation support device 100 creates a prompt for causing the generative model 30 to create a draft specification of the invention to be applied for, based on the claims, drawings, and knowledge data. Specifically, for the setting information input on the setting screen shown in Fig. 28, the document creation support device 100 creates a prompt for causing the generative model 30 to create a draft specification of the invention to be applied for, based on the claims "ProposedClaim#1.docx" and drawings "ProposedFigure#1.pdf", using a specification template and / or sample data extracted from the knowledge database 40.

[0082] In step S104, the document creation support device 100 inputs the generated prompt to the generative model 30. For example, the generative model 30 is provided on a server managed by an administrator of the generative model 30, and the document creation support device 100 transmits the prompt to the server. Upon receiving the prompt, the server inputs the prompt to the generative model 30, and causes the generative model 30 to generate a draft specification in accordance with the prompt.

[0083] In step S105, the document creation support device 100 acquires the draft specification from the generative model 30. Specifically, the document creation support device 100 acquires the draft specification as an output result from the generative model 30 and provides the acquired draft specification to the user terminal 20. For example, if the draft specification acquired from the generative model 30 is text data, the document creation support device 100 may display the text data of the draft specification as an output result in a specification creation tool on the user terminal 20, as shown in FIG. 29 . For example, the user may edit the draft specification on the specification creation tool.

[0084] Alternatively, the document creation support device 100 may convert the text data of the draft specification into document data such as WORD specified by the user, and provide the converted data of the draft specification to the user terminal 20.

[0085] According to the document creation support process of this embodiment, it is possible to have the generative model 30 generate a draft specification corresponding to the claims and drawings by utilizing prior knowledge data, and to obtain a draft specification corresponding to a specific technical field and / or user.

[0086] Although the examples of the present disclosure have been described in detail above, the present disclosure is not limited to the specific embodiments described above, and various modifications and variations are possible within the scope of the gist of the present disclosure as set forth in the claims. [Explanation of symbols]

[0087] 20 User terminal 30 Generative Models 40 Knowledge Database 100 Document creation support device 110 Reception 120 Knowledge Extraction Department 130 Prompt Creation Department 140 Model Interface Section

Claims

1. a reception unit for receiving claims, drawings, and technical fields for the invention to be applied for; a knowledge extraction unit that extracts knowledge data corresponding to the received technical field from knowledge data related to specification preparation corresponding to each technical field; a prompt creation unit that creates a prompt for a generative model to create a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; a model interface unit that inputs the prompt into the generative model and acquires the draft specification from the generative model; A document creation support device having the above.

2. The knowledge extraction unit extracts a specification template corresponding to the technical field as the knowledge data, 2. The document creation support device according to claim 1, wherein the prompt creation unit creates a prompt for creating the draft specification based on the claims and the drawings in accordance with the specification template.

3. The knowledge extraction unit extracts sample data of claims, drawings, and specifications corresponding to the technical field as the knowledge data, 2. The document creation support device according to claim 1, wherein the prompt creation unit creates a prompt for creating the draft specification based on the claims and the drawings by referring to the sample data.

4. A reception unit that receives claims, drawings, and user identification information for the invention that is the subject of the application; a knowledge extraction unit that extracts knowledge data corresponding to the received user identification information from knowledge data related to specification creation corresponding to each user; a prompt creation unit that creates a prompt for a generative model to create a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; a model interface unit that inputs the prompt into the generative model and acquires the draft specification from the generative model; A document creation support device having the above.

5. The knowledge extraction unit extracts a specification template corresponding to the user identification information as the knowledge data, 5. The document creation support device according to claim 4, wherein the prompt creation unit creates a prompt for creating the draft specification based on the claims and the drawings in accordance with the specification template.

6. The knowledge extraction unit extracts sample data of claims, drawings, and specifications created by the user of the user identification information as the knowledge data, 5. The document creation support device according to claim 4, wherein the prompt creation unit creates a prompt for creating the draft specification based on the claims and the drawings with reference to the sample data.

7. Accepting the claims, drawings and technical field of the invention to be applied for; extracting knowledge data corresponding to the received technical field from knowledge data related to specification preparation corresponding to each technical field; generating a prompt for a generative model to generate a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; inputting the prompt into the generative model and receiving the proposed specification from the generative model; A document creation support method in which the above is executed by a computer.

8. Receiving claims, drawings, and user identification information for the invention to be applied for; extracting knowledge data corresponding to the received user identification information from knowledge data related to specification creation corresponding to each user; generating a prompt for a generative model to generate a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; inputting the prompt into the generative model and receiving the proposed specification from the generative model; A document creation support method in which the above is executed by a computer.

9. Accepting the claims, drawings and technical field of the invention to be applied for; extracting knowledge data corresponding to the received technical field from knowledge data related to specification preparation corresponding to each technical field; generating a prompt for a generative model to generate a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; inputting the prompt into the generative model and obtaining the draft specification from the generative model; A document creation support program that allows a computer to execute the following.

10. Receiving claims, drawings, and user identification information for the invention to be applied for; extracting knowledge data corresponding to the received user identification information from knowledge data related to specification creation corresponding to each user; generating a prompt for a generative model to generate a draft specification of the invention based on the claims, the drawings, and the extracted knowledge data; inputting the prompt into the generative model and receiving the proposed specification from the generative model; A document creation support program that allows a computer to execute the following.

Images