Patent document creation device, control method of patent document creation device and control program of patent document creation device

Abstract

To provide a patent document creation device, etc. being further user-friendly in supporting creation of a patent document.SOLUTION: A patent document creation device for supporting creation of a patent document using a large-scale language model, includes an acquisition unit that acquires patent information relating to a patent to be described in the patent document, a question generation unit that generates a first question for requesting a detailed description of the patent using the large-scale language model on the basis of the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that when the answer to the first question is not acquired from the user, generates an answer to the first question using the large-scale language model, and a creation unit that creates the patent document using the large-scale language model on the basis of the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a patent document drafting device, a control method for a patent document drafting device, and a control program for a patent document drafting device. [Background technology]

[0002] Conventionally, systems have been developed that automatically prepare patent documents to be submitted to the Patent Office when filing patent applications. For example, applicants have been providing information processing devices that assist users in preparing patent documents (e.g., Patent Documents 1 and 2). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 7086424 [Patent Document 2] Patent No. 7171100 Summary of the Invention [Problem to be solved by the invention]

[0004] In order to support the creation of patent documents, there is a demand for patent document creation devices and the like that are easier for users to use. [Means for solving the problem]

[0005] According to one embodiment of the present invention, a patent document drafting device that supports the drafting of patent documents using a large-scale language model includes an acquisition unit that acquires patent information related to patents described in patent documents, a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates a patent document using a large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0006] In a patent document creation device according to one embodiment of the present invention, the question generation unit generates a second question requesting a detailed explanation of the patent based on the first user answer or the first automatic answer using a large-scale language model, the answer acquisition unit acquires an answer to the second question from the user, and if an answer to the second question is not acquired from the user, the answer generation unit generates an answer to the second question using the large-scale language model, and the creation unit may create a patent document using the large-scale language model based on the patent information, the first user answer or the first automatic answer, and the second user answer to the second question acquired from the user or the second automatic answer to the second question generated by the answer generation unit.

[0007] The patent document drafting device according to one embodiment of the present invention may further include a correction unit that corrects the patent document drafted by the drafting unit using a large-scale language model.

[0008] In the patent document drafting device according to one embodiment of the present invention, the drafting unit may draft the patent document for each of a plurality of items that constitute the patent document.

[0009] In a patent document drafting device according to one embodiment of the present invention, when a large-scale language model is used, a predetermined prompt that has been set in advance is input to the large-scale language model, and the patent document drafting device may further include a setting unit that accepts at least one of a user's change to the predetermined prompt to be input to the large-scale language model and a user's setting of a new prompt.

[0010] A patent document creation device according to one embodiment of the present invention further includes a display processing unit that displays an operation screen on a terminal device used by a user, the operation screen including an input field into which patent information can be input, a display field that displays the patent document created by the creation unit, and a selection field in which multiple prompts can be selected for processing the patent document displayed in the display field using a large-scale language model, and a selection receiving unit that receives from the user a selection of at least one of the multiple prompts displayed in the selection field, and the creation unit may process the patent document displayed in the display field using a large-scale language model according to the selected prompt in accordance with the selection operation of the prompt received by the selection receiving unit.

[0011] In a patent document creation device according to one embodiment of the present invention, the display processing unit may display multiple prompts in a selection field, each of which is an identifiable object, and the selection receiving unit may receive a user's action of superimposing an object on the display field as a selection of a prompt associated with a specified object.

[0012] A control method for a patent document drafting device that supports the drafting of patent documents using a large-scale language model according to one embodiment of the present invention executes the following steps: an acquisition step of acquiring patent information related to a patent to be described in the patent document; a question generation step of generating a first question requesting a detailed explanation of the patent based on the patent information using a large-scale language model; an answer acquisition step of acquiring an answer to the first question from a user; an answer generation step of generating an answer to the first question using a large-scale language model if an answer to the first question is not acquired from the user; and a creation step of creating a patent document using a large-scale language model based on the patent information and the first user answer to the first question acquired from the user or the first automatic answer to the first question generated in the answer generation step.

[0013] A control program for a patent document drafting device that supports the drafting of patent documents using a large-scale language model according to one embodiment of the present invention causes a computer to realize the following: an acquisition function that acquires patent information related to patents to be described in a patent document; a question generation function that generates a first question requesting a detailed explanation of the patent using a large-scale language model based on the patent information; an answer acquisition function that acquires an answer to the first question from a user; an answer generation function that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired from the user; and a creation function that creates a patent document using a large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation function. [Effects of the Invention]

[0014] According to one embodiment of the present invention, it is possible to provide a patent document preparation device and the like that is easier for users to use when supporting the preparation of patent documents. [Brief explanation of the drawings]

[0015] [Figure 1] FIG. 1 is a schematic diagram of the configuration of a patent document preparation support system according to one embodiment of the present invention. [Figure 2]FIG. 2 is an example of a functional block diagram of a patent document drafting device and a user terminal according to an embodiment of the present invention. [Figure 3] Figure 3 is an example of a prompt database. [Figure 4] FIG. 4 is an example of a control flow diagram of a patent document drafting device according to an embodiment of the present invention. [Figure 5] 5(a) and (b) are examples of display screens on a user terminal. [Figure 6] FIG. 6 shows an example of a display screen on a user terminal. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, an embodiment of the present invention will be described using the drawings. Note that the drawings are merely examples, and the present invention is not limited to those shown in the drawings. For example, the illustrated patent document creation device (server), large-scale language model server, and number of user terminals, data sets (tables), flowcharts, and display screen examples are merely examples, and the present invention is not limited to these.

[0017] <System configuration> 1 is a diagram showing an example of the configuration of a patent document drafting support system according to one embodiment of the present invention. The patent document drafting support system 600 includes at least a server (patent document drafting device) 100, a user terminal 200, and a large-scale language model server 300.

[0018] The server 100 functions as a platform for providing patent document preparation support services, and outputs the prepared patent documents to the user terminal 200 connected via the network 500. Here, the "patent document" refers to documents submitted to the Patent Office when filing a patent application, and may refer to the specification, claims, drawings, abstract, etc. As will be described in detail later, the server 100 prepares the patent document while exchanging information with the user terminal 200. At this time, the server 100 prepares the patent document using the large-scale language model function provided by the large-scale language model server 300 via the cloud.

[0019] Large-scale language models (LLMs) are a type of generative artificial intelligence (AI) specialized for natural language processing (NLP). They are language models that realize advanced natural language generation (NLG) and are constructed through deep learning of massive amounts of text data. A language model is a system that predicts the next word after a given word by understanding the patterns of natural language used by humans and analyzing the probability of word occurrence. Examples of large-scale language models include ChatGPT (Chat Generative Pre-trained Transformer) (registered trademark), PaLM2 (registered trademark), and LaMDA2 (registered trademark). Words and sentences used as instructions to be input into these large-scale language models are generally referred to as prompts.

[0020] The server 100 may be any information processing device capable of implementing the functions described in each embodiment. The server 100 may include, for example, a server device, a computer (e.g., a desktop, laptop, tablet, etc.), or a communication platform. Although the server 100 and the large-scale language model server 300 are shown as separate entities in FIG. 1 , this is not limiting. That is, each function described as being provided by the server 100 may be implemented by multiple servers or a single server. Furthermore, the functions provided by the large-scale language model server 300 may be implemented as part of the functions of the server 100. The server 100 may be, for example, a distributed server system that cooperates by communicating via a network, or may be a so-called cloud server. That is, the server 100 is not limited to a physical server, but may also include a virtual server implemented by software. The server 100 may be any information processing device capable of implementing the functions described in each embodiment, and may include, for example, a server device, a computer (e.g., a desktop, laptop, tablet, etc.), a communication platform, etc.

[0021] The user terminal 200 is a communication terminal used by a user of the patent document drafting support system 600, and has an application installed thereon for using the patent document drafting support service, and is capable of transmitting and receiving various types of information to and from the server 100. Note that it is not essential to install an application on the user terminal 200, and the user may access a web page for using the patent document drafting support service provided by the server 100 from the user terminal 200 via a web browser or the like, and transmit and receive various types of information to and from the server 100.

[0022] 1 shows a notebook computer as the user terminal 200, any terminal capable of implementing the functions described below may be used as the user terminal 200. For example, the user terminal 200 may be a computer (e.g., a tablet terminal or a desktop PC), a smartphone, etc.

[0023] The patent document drafting support system 600 may further include a database server (not shown) that stores various information (data) used in the patent document drafting support system 600. The data used in the patent document drafting support system 600 may be, for example, prompts to be input into a large-scale language model or user information.

[0024] Network 500 may include a wireless network or a wired network, such as a wireless LAN (WLAN), a wide area network (WAN), integrated service digital networks (ISDNs), wireless LANs, code division multiple access (CDMA), long term evolution (LTE), LTE-Advanced, fourth generation communication (4G), fifth generation communication (5G), sixth generation communication (6G) or later mobile communication systems. Note that network 500 is not limited to these examples and may also be, for example, a public switched telephone network (PSTN), Bluetooth (registered trademark), an optical fiber line, a satellite communication network, or the like. Network 500 may also be a combination of these.

[0025] Next, the hardware configuration and functional configuration of the server 100 and the user terminal 200 according to one embodiment of the present invention will be described with reference to Figure 2. First, a brief description of the typical operation of each functional unit will be given, and details will be described later using flow charts, display screen examples, etc.

[0026] <User device> (1) Hardware configuration of the user terminal The user terminal 200 includes a control unit 210 , a communication unit 220 , a display unit 230 , an input / output unit 240 , and a storage unit 270 .

[0027] The control unit 210 is typically a processor and includes a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), a microprocessor, a processor core, a multiprocessor, etc. The control unit 210 reads out a program stored in the storage unit 270 and executes code or instructions included in the read out program, thereby performing the functions and methods described in each embodiment.

[0028] The storage unit 270 stores various programs and data required for the operation of the user terminal 200. That is, the storage unit 270 stores programs related to the application of the patent document preparation support service. The storage unit 270 includes, for example, a flash memory and a memory (RAM, ROM, etc.) that provides a working area for the control unit 210.

[0029] The communication unit 220 is implemented as hardware such as a network adapter, communication software, or a combination of these. The communication unit 220 transmits and receives various data to and from the server 100 via the network 500 using any communication protocol.

[0030] The display unit 230 is a monitor that displays data according to the display data written in the frame buffer.

[0031] The input / output unit 240 includes an input device for inputting various operations to the user terminal 200, and an output device for outputting the processing results processed by the user terminal 200. The input device may include, for example, a touch panel, a touch display, a keyboard, a camera, and a microphone. The output device outputs the processing results processed by the control unit 210. The output device may include, for example, a touch panel, a speaker, etc.

[0032] (2) Functional configuration of user terminal The user terminal 200 includes a display processing unit 212 and an input / output processing unit 213 as functions realized by the control unit 210. Of the functional units shown in FIG. 2, functional units that are not essential in the embodiments described below may be omitted. Furthermore, the functions or processes of each functional unit may be realized by machine learning or AI to the extent feasible. Furthermore, some of the various processes described below as being executed by the user terminal 200 may be executed by the server 100.

[0033] The display processing unit 212 controls the display of data on the display unit 230. For example, in response to information transmitted from the server 100, the display processing unit 212 causes the display unit 230 to display an input screen for information required to create a patent document, or to display a patent document created by the server 100.

[0034] The input / output processing unit 213 may control transmission of various types of information to and from external devices via the input / output unit 240. For example, the input / output processing unit 213 may accept input of information relating to the content of an invention related to a patent application from the user 20, or input of an answer to a question sent from the server 100.

[0035] <server> (1) Server hardware configuration The server 100 includes a control unit 110, a communication unit 120, and a storage unit 170.

[0036] The control unit 110 is typically a central processing unit (CPU) that reads a program stored in the storage unit 170 and executes the code or instructions included in the read program to perform the functions and methods described in each embodiment. The control unit 110 may also include an MPU, a GPU, a microprocessor, a processor core, a multiprocessor, etc.

[0037] The storage unit 170 stores various programs and various data required for the operation of the server 100. The storage unit 170 may include, for example, a hard disk drive (HDD), a solid state drive (SSD), a flash memory, etc. The storage unit 170 may also include a memory that provides a working area for the control unit 110.

[0038] The communication unit 120 is implemented as hardware such as a network adapter, communication software, or a combination of these, and transmits and receives various data to and from the user terminal 200 and the large-scale language model server 300 via the network 500.

[0039] (2) Server functional configuration The server 100 includes, as functions realized by the control unit 110, an acquisition unit 111, a question generation unit 112, an answer acquisition unit 113, an answer generation unit 114, a creation unit 115, a correction unit 116, a setting unit 117, a display processing unit 118, and a selection receiving unit 119. Of the functional units shown in FIG. 2 , functional units that are not essential for the embodiments described below may be omitted. Furthermore, the functions or processes of the functional units may be realized by machine learning or AI to the extent feasible.

[0040] <Server processing flow> The processing flow of the server 100 according to one embodiment of the present invention will be described below with reference to Figures 3 to 6. Figure 3 shows the processing flow of the server 100 according to one embodiment of the present invention, Figure 4 shows an example of a prompt input to a large-scale language model in one embodiment of the present invention, and Figures 5 and 6 show examples of the display screen of the user terminal 200.

[0041] 3, the acquisition unit 111 acquires patent information related to the patent described in the patent document from the user terminal 200 (step S11). Here, the "patent information" refers to information indicating the content of the patent, and may be, for example, text data such as sentences or notes outlining the invention for which a patent is sought, papers, information on related or similar patents (publication publications, etc.), drawings, etc.

[0042] When acquiring patent information, the display processing unit 118 displays an operation screen including an input field into which the patent information can be input on the user terminal 200. Fig. 5(a) is an example of the operation screen displayed on the display unit 230 of the user terminal 200. Note that the figure is an example, and the present invention is not limited to this.

[0043] The operation screen 10 includes a patent information input field 11 that accepts patent information input from the user. The user may input patent information into the patent information input field 11 by directly inputting an outline of the invention for which the user wishes to obtain a patent, or by uploading data containing an outline of the invention (e.g., a Word file, a text file, a PDF file, an image file, etc.). When the "Create a patent document with this content" button is selected on the operation screen 10, the patent information input into the patent information input field 11 is sent to the server 100.

[0044] The question generation unit 112 generates a first question requesting a detailed explanation of the patent using a large-scale language model based on the patent information (step S12). The question generation unit 112 inputs a prompt obtained by combining a preset prompt with the patent information into the large-scale language model to generate the first question. FIG. 4 shows an example of a prompt that is preset and stored in the storage unit 170 or a database server (not shown). Note that the diagram is an example and is not precise, and the present invention is not limited to these.

[0045] The prompt table TB10 predefines prompts necessary for creating a patent document. FIG. 4 shows examples of prompts for creating items constituting a patent document, such as "Title of the Invention," "Technical Field," "Background Art," "Claims (Claim 1)," "Claim 2," "Prior Art Documents," "Problem to be Solved by the Invention," "Means for Solving the Problem," and "Mode for Carrying Out the Invention." The "First Question" and "Second Question" in the prompt table TB10 are prompts for outputting information to the user in the form of a question to facilitate the creation of the patent document. Here, the "First Question" and "Second Question" may, for example, request a more detailed explanation from the user about the meaning of a word contained in the patent document or the purpose achieved by a sentence contained in the patent document. Note that in the prompts, curly brackets {} indicate that the sentence corresponding to the item enclosed in the curly brackets should be entered.

[0046] The creation unit 115 may create items in a patent document that can be created using prompts stored in the prompt table TB10, based on patent information acquired from the user terminal 200. For example, in the example of the prompt table TB10 in FIG. 4, items such as "Title of the Invention," "Technical Field," and "Background Art" can be created using patent information. In response to acquiring patent information from the user terminal 200, the creation unit 115 may input prompts including {patent information} into a large-scale language model and create items obtained from each prompt. In other words, the creation unit 115 may create a patent document for each of the multiple items that make up the patent document. This makes it possible to prevent omission of items necessary for the patent document.

[0047] The communication unit 120 transmits the question generated by the question generation unit 112 to the user terminal 200 (step S13). At this time, the display processing unit 118 causes the user terminal 200 to display an operation screen including an answer input field that prompts the user to answer the first question. Fig. 5(b) is an example of the operation screen displayed on the display unit 230 of the user terminal 200. Note that the diagram is an example, and the present invention is not limited to this.

[0048] The operation screen 20 may include at least a patent information display field 21 that displays patent information received from the user, a question display field 23 that displays a first question sent from the server 100, and an answer input field 25 in which an answer to the first question can be input. In addition, the operation screen 20 may further include a patent document display field 22 that displays patent documents that the creation unit 115 has created so far.

[0049] As described above, the first question is generated based on the patent information entered in the patent information input field 11 (i.e., the patent information displayed in the patent information display field 21). In the example of FIG. 5(b), based on the patent information "This invention is...by XX.... XX functions to...," displayed in the patent information display field 21, questions such as "What is XX?", "How is XX...," and "What does □□ mean?" may be generated and displayed in the question display field 23. The user can input an answer to the question displayed in the question display field 23 in the answer input field 25. The answer entered in the answer input field 25 may be sent to the server 100 in response to the user selecting the "Send Answer" button on the operation screen 20.

[0050] The answer acquisition unit 113 determines whether an answer to the first question has been received from the user terminal 200 (step S14). If an answer to the first question has been received from the user terminal 200 (YES in step S14), i.e., if the server 100 has received an answer entered in the answer input field 25 of the operation screen 20, the creation unit 115 creates a patent document using a large-scale language model based on the patent information and the first user answer to the first question acquired from the user (step S17). The creation unit 115 combines the acquired {patent information} and {answer to the first question} (first user answer) with a prompt that uses {patent information} and {answer to the first question} in the prompts pre-stored in the prompt table TB10, and inputs the combined information into the large-scale language model. The created patent document may be transmitted to the user terminal 200, and the patent document display field 22 on the operation screen 20 may be updated.

[0051] According to one embodiment of the present invention, the server 100 may be able to generate an answer to a question sent from the server 100. For example, the user can request the server 100 to answer the question displayed in the question display field 23 by selecting the "Request an Answer" button 24 displayed near the question display field 23 on the operation screen 20.

[0052] If the "Request an answer" button 24 is selected, then in step S14, the answer acquisition unit 113 determines that an answer to the first question has not been received from the user terminal 200 (NO in step S14). If an answer to the first question is not acquired from the user, the answer generation unit 114 generates an answer to the first question using a large-scale language model (step S15). That is, the answer generation unit 114 may input the prompt "Please generate an answer to {question content}" in the prompt table TB10 into the large-scale language model to generate an answer.

[0053] The creation unit 115 creates a patent document using a large-scale language model based on the patent information and the generated first automatic answer to the first question (step S16). The creation unit 115 combines the acquired {patent information} and {answer to the first question} (first automatic answer) with a prompt using {patent information} and {answer to the first question} in the prompts pre-stored in the prompt table TB10, and inputs the combined results into the large-scale language model. Note that the prompts used in steps S16 and S17 may be identical, except that the {answer to the first question} is entered as "first automatic answer" in the former and "first user answer" in the latter. The created patent document may be transmitted to the user terminal 200, and the patent document display field 22 on the operation screen 20 may be updated.

[0054] In this way, according to one embodiment of the present invention, the creation of a patent document can be advanced while the content of the patent document is enriched through dialogue with the user. Furthermore, a large-scale language model can be used to enrich the content, reducing the user's workload, and the amount of information contained in the large-scale language model allows information beyond the user's knowledge to be reflected in the patent document.

[0055] <Other embodiments> In preparing the patent document, further questions and answers may be exchanged between the user terminal 200 and the server 100. For example, the question generator 112 may use a large-scale language model to generate a second question requesting a detailed explanation of the patent based on the first user answer or the first automatic answer. That is, a question for further probing the contents of the patent document may be generated using the large-scale language model based on the content of the answer to the first question. Similar to the first question, the generated second question may be transmitted to the user terminal 200 and displayed in the question display field 23 on the operation screen of the user terminal 200.

[0056] Furthermore, in a manner similar to that for the first question, the answer acquisition unit 113 may acquire an answer to the second question from the user. When an answer to the second question is not acquired from the user (i.e., when the user selects to request an answer from the server 100), the answer generation unit 114 may generate an answer to the second question using a large-scale language model. The creation unit 115 may create a patent document using a large-scale language model based on the patent information, the first user answer or the first automatic answer, and the second user answer to the second question acquired from the user or the second automatic answer to the second question generated by the answer generation unit 114. The created patent document may be transmitted to the user terminal 200, and the patent document displayed in the patent document display field 22 may be updated.

[0057] In this way, according to one embodiment of the present invention, patent documents are enriched through repeated exchange of questions and answers with the user. Furthermore, since the user can select to generate answers using a large-scale language model in response to questions generated by the server 100 using a large-scale language model, a patent document generation system with high usability can be realized.

[0058] The correction unit 116 may correct the patent document created by the creation unit 115 using a large-scale language model. For example, a correction instruction input field for inputting correction instructions for the patent document may be displayed on the operation screen of the user terminal 200, allowing the user to input correction instructions for the patent document. The correction unit 116 may cause the large-scale language model to correct the patent document created so far, using the correction instructions transmitted from the user terminal 200 as a prompt. Note that an upper limit may be set on the number of characters that can be input into the large-scale language model, and a guide may be provided to the user to input correction instructions for the patent document for each item constituting the patent document (e.g., "title of the invention," "claims," ​​"problems to be solved by the invention," etc.). For example, in response to receiving a request for correction of the patent document from the user, a screen may be displayed on the user terminal that allows the user to select the items to be corrected and the correction content.

[0059] Thus, according to one embodiment of the present invention, a large-scale language model can be used to further refine patent documents, and a patent document drafting system with high usability can be provided.

[0060] The above description deals with an embodiment in which prompts used in creating patent documents are preset. However, the setting unit 117 may accept at least one of a user's change of a prompt to be input to a large-scale language model and a user's setting of a new prompt. For example, a prompt update field for inputting an instruction to update the prompt may be displayed on the operation screen of the user terminal 200, allowing the user to input an instruction to update the prompt. Alternatively, the setting unit 117 may learn the user's tendency to make corrections based on the content and frequency of corrections made by the correction unit 116 in response to user instructions, and set a new prompt.

[0061] <Other operation screens> 6 shows another example of an operation screen displayed on the user terminal 200. The display processing unit 118 may cause the user terminal 200 to display an operation screen 50 including an input field (patent information input field) 51 into which patent information can be input, a display field (patent document display field) 52 that displays the patent document created by the creation unit 115, and a prompt selection field 54 in which a plurality of prompts 54a to 54d for performing processing using a large-scale language model on the patent document displayed in the patent document display field 52 can be selected.

[0062] The patent information input field 51 is an area where patent information can be input or uploaded, and the creation unit 115 may combine the patent information with a preset prompt and input the combined information into the large-scale language model in response to the patent information being input into the patent information input field 51. In addition, in the patent information input field 51, it may be possible to input questions sent from the server 100 (such as the first question and second question described above), answers to the questions, or to issue instructions requesting answers to the server 100.

[0063] The operation screen 50 may include an item selection field 53 that allows selection of each item that constitutes the patent document. In the operation screen 50, the patent document display field 52 may display a portion of the patent document created by the creation unit 115 that corresponds to the item selected in the item selection field 53. For example, when "claims" is selected in the item selection field 53, the patent document display field 52 may display the "scope of claims" of the patent document.

[0064] The prompt selection field 54 is a field in which a user can select a prompt to cause the large-scale language model to process the patent document displayed in the patent document display field 52. While the example in FIG. 6 shows four selectable prompts, the present invention is not limited to this number and may include more or fewer. In the prompt selection field 54, for example, prompt 54a (raw) may be a prompt instructing the user to generate more text. Prompt 54b (paragraph) may be a prompt instructing the user to add or update a paragraph. Prompt 54c (number) may be a prompt instructing the user to add a symbol. Prompt 54d (times) may be a prompt instructing the user to generate multiple answer plans for a question using the large-scale language model. The user may select an answer plan to use in creating the patent document from the multiple answer plans generated. The selection receiving unit 119 receives a user's selection of at least one of the multiple prompts displayed in the prompt selection field 54. The creation unit 115 may process the patent document displayed in the patent document display field 52 using a large-scale language model according to the selected prompt, in response to the selection operation of the prompt received by the selection receiving unit 119. Then, the patent document created or updated in response to the prompt may be displayed in the patent document display field 52.

[0065] The prompt selection operation is not limited to the above. For example, the display processing unit 118 may display multiple prompts as identifiable objects in the prompt selection field 54. The identifiable object may be, for example, a circular object with text indicating the content of the prompt superimposed on it, as shown in FIG. 6, or the object may have a different shape, such as a circle, triangle, or square. The prompt-identifiable object may be displayed so that it can be dragged (moved) according to the user's selection, and the selection receiving unit 119 may receive the user's action of superimposing (dragging) the object on the patent document display field 52 as a selection of a prompt associated with a specific object.

[0066] As described above, according to one embodiment of the present invention, patent documents can be created with simple operations on an operation screen.

[0067] While the present invention has been described with reference to the drawings and examples, it should be noted that those skilled in the art will readily be able to make various modifications and alterations based on this disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of the present invention. For example, the functions included in each component, step, etc. may be rearranged so as not to cause logical inconsistencies, and multiple components, steps, etc. may be combined or separated into one. Furthermore, the configurations described in the above embodiments may be combined as appropriate. For example, the components described as being included in the patent document creation device 100 may be physically distributed across multiple computers or implemented as a single computer.

[0068] For example, in the above description, the server 100 and the user terminal 200 exchange questions and answers twice. However, the exchange of questions and answers may occur more than twice. In this case, the server 100 generates questions based on the answers obtained from the user, thereby enhancing the content of the patent document.

[0069] The programs of the embodiments of the present disclosure may be provided in a state stored in a storage medium readable by an information processing device. The storage medium may store the programs in a "non-transitory tangible medium." The programs include, for example, software programs and control programs. When the functional units of the server 100 are realized by software, the server 100 functions as an acquisition unit 111, a question generation unit 112, an answer acquisition unit 113, an answer generation unit 114, a creation unit 115, a correction unit 116, a setting unit 117, a display processing unit 118, and a selection reception unit 119 by the processor executing the programs loaded on the memory.

[0070] The storage medium may, where appropriate, include one or more semiconductor-based or other integrated circuits (ICs) (e.g., field programmable gate arrays (FPGAs), application specific ICs (ASICs), etc.), hard disk drives (HDDs), hybrid hard drives (HHDs), optical disks, optical disk drives (ODDs), magneto-optical disks, magneto-optical drives, floppy diskettes, floppy disk drives (FDDs), magnetic tapes, solid state drives (SSDs), RAM drives, secure digital cards or drives, any other suitable storage media, or any suitable combination of two or more of these. The storage medium may, where appropriate, be volatile, non-volatile, or a combination of volatile and non-volatile.

[0071] Furthermore, the program of the present disclosure may be provided to the server 100 via any transmission medium (such as a communication network or broadcast waves) capable of transmitting the program.

[0072] The programs disclosed herein can be implemented using, for example, scripting languages ​​such as Actionscript and JavaScript (registered trademark), object-oriented programming languages ​​such as Python, Objective-C, C language, Go language, Swift (registered trademark), Koltin, and Java (registered trademark), and markup languages ​​such as HTML5. [Explanation of symbols]

[0073] 100 Server (Patent Document Creation Device) 110 control section 111 Acquisition Department 112 Question generation part 113 Answer acquisition part 114 Answer generation part 115 Creation Department 116 Correction section 117 Setting section 118 Display processing section 119 Selection Reception Department 120 Communications Department 170 Storage section 200 User terminal (terminal device) 210 Control Unit 212 Display processing unit 213 Input / Output Processing Unit 220 Communications Department 230 Display section 240 Input / output section 270 Storage section 300 Large-scale language model server 500 Network 600 Patent Document Creation Support System

[0074] The patent document drafting device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. Furthermore, if an answer cannot be obtained from the user, the device generates an answer using the large-scale language model. Then, the device creates a patent document using the large-scale language model based on the patent information and the answer from the user. In this way, the patent document drafting device can support the creation of patent documents.

[0075] The present invention aims to provide a patent document creation support device and method using LLM technology. Specifically, LLM technology built from large amounts of text data can be used to generate questions, answers, and other documents, and this technology is applied to provide a patent document creation support device and method.

[0076] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained, the device also generates an answer using the large-scale language model. The device then creates a patent document based on the acquired information and the answer. This can assist in the creation of patent documents.

[0077] The present invention provides a patent document drafting device using a large-scale language model. This device acquires information about patents to be described in patent documents, generates questions and answers using a large-scale language model based on that information, and creates patent documents based on the acquired information and answers. This eliminates the need for manual work as in the past, enabling efficient patent management. Furthermore, the use of a large-scale language model enables questions and answers to be generated with high accuracy and speed, enabling more accurate and rapid patent management.

[0078] The present invention relates to an information processing device. The information processing device of the present invention includes a collection unit that collects similar document information, a presentation unit that presents a collection of the collected document information to a user, and a calculation unit that calculates the similarity of the document information based on feedback collected from the user, including indications of documents that have been mistakenly included in the collection and indications of document information that should be added but that is not included in the collection. The calculation unit optimizes the calculation of the similarity of the document information based on the feedback.

[0079] Although the word "prompt" is used in the embodiment of the present invention, the same function can be achieved without using the word "prompt."

[0080] The system determines a set of search terms and a type of question from the question, searches for a set of related documents from a set of documents stored in a document set storage device according to the determined set of search terms and type of question, extracts answers to the question from the related documents, and provides the extracted answers and the document information from which the answers were extracted as response results to the question.

[0081] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question. The creation unit creates a patent document using a large-scale language model based on the patent information and the first user answer or the first automatic answer. The answer information acquisition unit calculates the value of a first formula using two terms selected from additional terms associated with the classification information by correspondence information and terms extracted by the term extraction unit, and acquires answer information in which the value of the first formula is larger than the others.

[0082] The patent document creation device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. In addition, during machine learning, the answer information acquisition unit can acquire answer information using an equation (first equation) corresponding to additional terms and classification information by using the value of a first equation as training data.

[0083] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The answer information acquisition unit may further include a reranking unit (not shown) that reranks documents acquired by the document acquisition unit. The calculation unit may calculate the value of a first formula for answer candidate information included in documents reranked by the reranking unit instead of documents acquired by the document acquisition unit.

[0084] The patent document drafting device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a question based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer from a user, an answer generation unit that generates an answer using a large-scale language model if an answer cannot be acquired, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The acquisition unit may also acquire, as answer history information, at least one of information on the category of a question answered by a user, a questioner, a user attribute of the questioner, a question statement, a user's answer to the question, and an evaluation of the answer statement.

[0085] The present invention relates to a method for acquiring a work log in an information processing device. According to the present invention, when a user performs a work using an application installed in the information processing device, information about the work can be acquired and stored in a storage device as work log information. This allows the history of the work performed by the user to be acquired, and by analyzing the user's work history, it is possible to improve the user's work efficiency and improve the application. The information processing device includes a work log acquisition unit that, when a user performs a work using an application installed in the information processing device based on setting information stored in the storage device, acquires information about the work and stores it in the storage device as work log information.

[0086] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer using a large-scale language model if an answer to the first question is not acquired, and a creation unit that creates a patent document using a large-scale language model based on the patent information and a first user answer or a first automatic answer. The device also includes a reference information acquisition unit that can acquire an answer created by referencing the same document from a past question and answer record database and a corresponding question sentence, and pass them to an output unit. The present invention is an application of large-scale language model technology, eliminating the need for cumbersome procedures and time-consuming processes that are conventionally performed manually. Furthermore, the device facilitates access to past similar cases and past similar cases, enabling more efficient business processing.

[0087] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, patent documents can be drafted using a large-scale language model by acquiring information about patents to be described in patent documents, generating questions based on that information, and acquiring answers. Furthermore, as information on answering ability, at least one of user information indicating the user's attributes and answer history information, which is information on answers given by the user in the past, can be acquired. This is expected to enable the generation of more appropriate questions and answers.

[0088] The patent document drafting device according to this embodiment is configured as shown in FIG. 1. The patent document drafting device 1 includes an acquisition unit 2, a question generation unit 3, an answer acquisition unit 4, an answer generation unit 5, and a creation unit 6. The acquisition unit 2 is a component for collecting information related to patents, and can acquire patent information from, for example, a database of the Japan Patent Office. The question generation unit 3 is a component for generating questions using a large-scale language model based on the patent information collected from the acquisition unit 2. The answer acquisition unit 4 is a component for receiving answers from a user, and can accept, for example, text data entered by a user. The answer generation unit 5 is a component used when an answer is not received from the user, and can generate an answer using a large-scale language model. The creation unit 6 is a component for creating a patent document based on question information and answer information.

[0089] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A feature of the present invention is that it is an automatic response system using a large-scale language model. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and an automatic response system that is constructed using the large-scale language model based on the patent information and the various information and the like acquired from the user. The present invention also includes the above-mentioned configuration.

[0090] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired, and a creation unit that creates a patent document using the large-scale language model. Furthermore, the statistical data generated by the answer generation means described above with reference to the drawings may be used as a primary answer to be presented to a user, and the user may be allowed to select a predicate as an answer from this list of primary answers. When a specific predicate is selected by the user, a passage search may be performed again together with the question keyword, and elements (subject and modifiers) related to the selected predicate may be extracted by a syntactic and semantic analysis unit and presented to the user.

[0091] The patent document drafting device according to the present invention includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0092] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A feature of the present invention is "automatic response" using the "large-scale language model." This eliminates the need for "manual response" and enables "quick and accurate response." Furthermore, the "large-scale language model" has a "learning function" that learns from "past response history" and "similar response examples" to enable "more appropriate responses." This invention provides an "automated response" using a "large-scale language model," eliminating the need for "manual response" and enabling "quick and accurate response." Furthermore, the "large-scale language model" has a "learning function," learning from "past response history" and "similar response examples," enabling "more appropriate responses." These effects enable the system to meet the requirements of "high speed, high accuracy, diversity, flexibility, scalability, and reliability," which were difficult to achieve with conventional technology. Furthermore, this invention has been realized as a "patent document preparation device," and is expected to contribute to improving the efficiency and quality of patent document preparation work.

[0093] The patent document creation device is composed of an acquisition unit that acquires patent information related to the patent to be described in the patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0094] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question. The creation unit creates the patent document using the large-scale language model. (Variant Example) In the above explanation, the question generation unit generates only the first question, but it is also possible to generate multiple questions, generate multiple corresponding automatic answers, and extract the necessary information from them to create a patent document.

[0095] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question cannot be acquired, and a creation unit that creates patent documents using the large-scale language model. The patent document drafting device of the present invention uses a large-scale language model to enable the drafting of patent documents with high accuracy and speed.

[0096] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A feature of the present invention is that it is an automatic response system using a large-scale language model.

[0097] The patent document creation device of this embodiment includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using a large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit.

[0098] The document search unit requests the basis document adopted in the previous question instead of searching the document database for the current question. The document search unit then forms a group of documents containing the basis document and searches only within that group, thereby reducing processing time.

[0099] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if the answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. In some embodiments of the present invention, the device can also be configured to provide answers to other designated terminals.

[0100] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires a first answer from a user, an answer generation unit that generates a first automatic answer when a first answer cannot be acquired, and a creation unit that creates a patent document using the large-scale language model. The expression acquisition means is composed of four expression acquisition means 141 to 144, namely, first expression acquisition means to fourth expression acquisition means. The expression acquisition means may further include a fifth expression acquisition means, etc.

[0101] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if no answer is acquired from the user, and a drafting unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. In an embodiment of the present invention, four candidate generation units are provided, each of which generates candidate sentences using a different method.

[0102] The present invention relates to an information processing device. The information processing device of the present invention includes a control unit that generates answer candidates and a display unit that presents the answer candidates. When there are multiple answer candidates, the control unit generates information for presenting multiple answer candidates that are likely to be answers. The display unit presents the answer candidates based on the information generated by the control unit. According to the present invention, multiple answer candidates that are likely to be answers can be presented, making it easier for the user to select a more appropriate answer. Furthermore, according to the present invention, more accurate information can be provided based on the answer selected by the user.

[0103] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0104] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires a first answer from a user, an answer generation unit that generates a first automatic answer when a first answer is not acquired, and a creation unit that creates patent documents using the large-scale language model. The present invention improves the efficiency of patent document drafting.

[0105] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. 1. A speech recognition device that outputs a speech recognition result from a speech signal, comprising: a speech recognition processing unit that outputs the speech recognition result from the speech signal; an environment estimation processing unit that estimates a type and state of an environment in which the speech signal is uttered based on a time change and a frequency change of a frequency spectrum distribution of the speech signal; and a control processing unit that controls parameters of the speech recognition processing unit in accordance with the type and state of the environment estimated by the environment estimation processing unit. The control processing unit is characterized in that it can control parameters such as the number of filter banks and frame length used by the speech recognition processing unit in accordance with the type and state of the environment estimated by the environment estimation processing unit, so as to output more accurate speech recognition results from the speech signal. The present invention relates to a speech recognition device that outputs a speech recognition result from a speech signal. Specifically, the speech recognition device of the present invention includes a speech recognition processing unit for outputting the speech recognition result from the speech signal, an environment estimation processing unit for estimating the type and state of an environment from the amount of change over time and the amount of change in frequency of a frequency spectrum distribution, and a control processing unit.

[0106] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, patent information is acquired, a question is generated based on the acquired information using a large-scale language model, and an answer is obtained from the user. If an answer cannot be obtained, an answer is generated using the large-scale language model. Then, a patent document is drafted based on the acquired information and answer. According to the present invention, the drafting of patent documents can be performed easily and efficiently. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit; and the large-scale language model must be trained. The present invention is characterized by the requirement of a large-scale, trained language model. As a result, the present invention enables high-precision, high-speed processing, and efficient creation of patent documents. Furthermore, because the present invention uses a large-scale, trained language model, it does not require specialized knowledge or technical skills, allowing for easy and quick operation, improving convenience.

[0107] The patent document creation device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. If an answer to the first question is not acquired from the user, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer to the first question acquired from the user or the first automatic answer to the first question generated by the answer generation unit. Moreover, since the basis document information used to answer the related question is likely to include answer elements useful for the question information being processed, appropriate answer information can be presented to the user.

[0108] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it is composed of an acquisition unit that acquires patent information, a question generation unit that generates a first question, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer when an answer cannot be acquired, and a drafting unit that creates a patent document. Furthermore, if the revised question generation model is a neural network, an error function related to correct answer data can be defined for each of machine comprehension and revised question generation, and the sum of these error function values ​​can be treated as the error of the entire neural network, and parameters can be updated to reduce this error.

[0109] The present invention is a patent document drafting device that supports the drafting of patent documents using a large-scale language model, and includes: an acquisition unit that acquires patent information related to patents described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The question answering execution unit of the revised question generation unit discovers and extracts information related to the input question from the related documents X to perform machine reading comprehension using the document vector sequence and the question vector sequence, generates a revised question Q' and its answer A', generates and outputs a new document vector sequence and a new question vector sequence when the revised question Q' and its answer A' are input, and outputs them as the processing results of the new document-question matching layer when the revised question Q' and its answer A' are input.

[0110] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device comprises an acquisition unit that acquires patent information, a question generation unit that generates a first question, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer when an answer cannot be acquired, and a creation unit that creates a patent document. The question generation unit can use the large-scale language model to generate a second question that requests a detailed explanation of the patent based on the first user answer or the first automatic answer.

[0111] The present invention relates to a patent document drafting device and a patent document drafting method. According to the present invention, a patent document drafting device and a patent document drafting method can be provided that support the creation of patent documents using a large-scale language model. Specifically, the creation of patent documents can be supported by generating questions based on patent information using a large-scale language model and obtaining answers. Furthermore, if an answer cannot be obtained, an answer can also be generated using the large-scale language model. This enables the creation of patent documents more efficiently and accurately than conventional manual patent document drafting. Step S103: The question restoration unit of the revised question generation unit uses the matching information (hidden state vector and matching matrix M) generated by the matching unit, the input question, and the related documents to generate a natural sentence that will serve as a revised question through the following steps S103-1 to S103-7.

[0112] The present invention relates to a patent document drafting device and a patent document drafting method. The patent document drafting device supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer when no answer is acquired, and a creation unit that creates a patent document. As processing in the document encoding layer and question encoding layer of the revised question generation model, related documents and an input question are each converted (encoded) into a d-dimensional word vector sequence. Step S303-2: Next, the question answering execution unit of the revised question generation unit outputs a revised question from the encoding results of the related documents and the input question as processing of the attention layer and decoder layer of the revised question generation model shown in the figure.

[0113] 2. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The revised question generation unit uses an RNN to create a vector for outputting words that constitute the revised question as processing in the decoding layer of the revised question generation model shown in the figure. Step S305-2: Revised Question Generation Unit The revised question generation unit outputs a word expression vector sequence from the input word sequence using an RNN as processing in the encoding layer of the revised question generation model shown in the figure, which is then processed in the decoding layer. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model when an answer to the first question is not acquired from the user; and a query processing unit that generates an answer to the first question using the patent information, the first user answer to the first question acquired from the user, or a previous answer generated by the answer generation unit. Step S305-2: The revised new word string is the word string following "new" in the old word string that includes the "new word string."

[0114] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. The device is composed of an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a drafting unit. The question generation unit generates a first question based on the patent information using a large-scale language model, and the answer acquisition unit acquires a first user answer from a user. If a first user answer is not available, the answer generation unit generates a first automatic answer using the large-scale language model. The drafting unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The related question generation unit extracts information from the input question, searches a question record database as shown in the figure using the information extraction results to find similar questions, and generates related questions that follow the similar questions.

[0115] The present invention is a patent document drafting device that supports the drafting of patent documents using a large-scale language model, and includes: an acquisition unit that acquires patent information related to patents described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A feature of the present invention is that it enables the rapid and accurate creation of patent documents through automation technology using large-scale language models. Furthermore, the present invention applies natural language processing technology, a type of artificial intelligence technology, making it possible to generate natural-sounding sentences that sound like they are written by a human. Another advantage of the present invention is that it is expected to improve productivity by automating tasks that were previously performed manually. Step S305-2: The question generation unit of the revised question generation model processes the input layer and intermediate layer of the revised question generation model shown in the figure by outputting revised answer candidate B (i.e., the starting point and ending point tend) from the input answer candidate A (i.e., the starting point and ending point tend), and creating and outputting an answer vector from the revised answer candidate B (i.e., the starting point and ending point tend).

[0116] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents is acquired, a question is generated based on the information using a large-scale language model, and an answer is obtained from a user. If an answer cannot be acquired, an answer is generated using the large-scale language model. Then, a patent document is drafted using the large-scale language model based on the acquired information and answer. This improves the efficiency of patent document drafting. Step S102: The matching unit of the revised question generation unit generates, as matching information, a hidden state vector to be the initial state of the Decoder and a matching matrix that is a matching model used in the machine reading comprehension task, through the following steps S102-1 to S102-4. The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, a question is generated from information about a patent described in a patent document, and an answer is obtained from a user. If an answer cannot be obtained, an answer is generated using a large-scale language model. Then, a patent document is drafted using the large-scale language model based on the obtained information and answer. This improves the efficiency of patent document drafting.

[0117] The present invention relates to a patent document drafting device. The patent document drafting device supports the drafting of patent documents using a large-scale language model. The device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user; and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The question generation unit generates the first question by inputting a prompt obtained by combining the patent information with a preset prompt into the large-scale language model.

[0118] The patent document creation device includes an acquisition unit that acquires patent information related to a patent to be described in a patent document, a question generation unit that generates a first question based on the patent information using a large-scale language model, an answer acquisition unit that acquires a first user answer from a user, an answer generation unit that generates a first automatic answer using a large-scale language model if a first user answer is not acquired, and a creation unit that creates a patent document using a large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0119] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. This system is also applicable to cases where document information is classified by multiple users. 2. The patent document drafting device according to claim 1, wherein the large-scale language model is based on a Transformer architecture. A patent document creation device is realized using a large-scale language model based on the Transformer architecture.

[0120] The present invention is a patent document creation device in which, in response to a question from a user, a question and answer list management means retrieves one question and answer record from a question and answer list in a database, calculates the degree of match between the words contained in the question received from the user and the words contained in the background information, calculates the degree of match between the words contained in the question received from the user and the words contained in the main part of the question, calculates the degree of match between the words contained in the question received from the user and the words contained in the content of the answer, and calculates a final degree of match based on these calculated degrees of match, degree of match, and degree of match, thereby providing the optimal answer.

[0121] The sentence judgment function management unit manages judgment functions, which receive as input the sentence to be judged, the type of the sentence to be judged, and the judgment result for the immediately preceding sentence, and by comparing this received information with heuristic knowledge, determine whether the sentence to be judged is a background information candidate, a question main part candidate, an answer content candidate, or any of the above candidates at the time the sentence to be judged is read in. If the sentence to be judged is a background information candidate, a "background candidate" is returned; if it is a question main part candidate, a "question candidate" is returned; if it is an answer content candidate, an "answer candidate" is returned; and if it is none of these candidates, a "non-candidate" is returned. Then, when candidates for the question main part are obtained, it is possible to select an appropriate one as the question main part.

[0122] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question. The creation unit creates a patent document using a large-scale language model based on the patent information and the first user answer or the first automatic answer. The answer information acquisition unit calculates the value of a first formula using additional terms associated with the classification information by correspondence information and two terms selected from the terms extracted by the term extraction unit, and acquires answer information whose value is larger than the others.

[0123] The question answering system of the present invention includes a question input unit that inputs a question sentence, a document search unit, an answer extraction unit, and an answer output unit. The question sentence input by the question input unit is subjected to morphological analysis by the question analysis unit to generate search terms. The document search unit acquires document identifiers associated with words in the document index information based on the search terms. The answer extraction unit extracts answer words associated with the acquired document identifiers based on the question type and the search terms, thereby acquiring identifiers indicating documents that are candidate answers to the question sentence even if the key words of the question sentence are not directly contained in the text of the searched document. The answer output unit then generates appropriate answers from the acquired answer words and presents them to the user. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention utilizes the latest technologies such as large-scale language model technology and natural language processing technology. The patent document drafting device of the present invention utilizes the latest technologies such as large-scale language model technology. The patent document drafting device of the present invention utilizes the latest technologies such as large-scale language model technology. The patent document drafting device of the present invention utilizes the latest technologies such as large-scale language model technology. The patent document drafting device of the present invention utilizes the latest technologies such as large-scale language model technology.

[0124] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit that generates a first question based on the patent information, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired, and a creation unit that creates a patent document using a large-scale language model based on the patent information and a first user answer or a first automatic answer. In addition, the respondent chat screen, which is a screen for text chatting with users, has a user display area, a message display area, an input area, a transmission object, and an answer candidate generation object.

[0125] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0126] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit.

[0127] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. If the first user answer is not acquired, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0128] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. It receives an answer from the user, but if no answer is received, it generates an answer using the large-scale language model. It creates a patent document using the large-scale language model based on the patent information and the answer from the user or the answer automatically generated.

[0129] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0130] The present invention relates to a dialogue processing device and a dialogue processing method that can accurately grasp the user's intention and provide an appropriate answer by generating a question summary for a question entered by a user using a large-scale language model and generating an answer based on the question summary. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit that generates a question summary using a large-scale language model, an answer acquisition unit, an answer generation unit, and a creation unit. In addition, if an appropriate answer is not obtained, log data is output by a log output processing unit. A sentence suitable for an example patent specification, related to the content of claim 1: In one embodiment of the present invention, a patent document drafting device is provided. The patent document drafting device includes an acquisition unit for acquiring patent information, a question generation unit for generating a question summary using a large-scale language model, an answer acquisition unit, an answer generation unit, and a creation unit. The patent document drafting device is capable of providing an appropriate answer to a question entered by a user. Furthermore, if an appropriate answer is not obtained, log data is output by a log output processing unit.

[0131] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. If the first user answer is not acquired, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. A sentence suitable for an example patent specification, related to the content of claim 1: In an embodiment of the present invention, a patent document drafting device is provided. The device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. If the first user answer is not acquired, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0132] 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0133] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires a first user answer from a user, an answer generation unit that generates a first automatic answer when a first user answer is not acquired, and a creation unit that creates a patent document using the large-scale language model. It is also possible to configure the device so that a user can select a predicate as an answer from a list of primary answers, and elements related to the selected predicate (such as a subject or modifier) ​​are extracted and presented to the user.

[0134] The question-answering system of the present invention can generate answers to questions from users using a large-scale language model. Specifically, the system is realized by including an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit in a patent document drafting device. Furthermore, the answer filter unit can notify relevant answers. This configuration allows users to easily and quickly obtain the information necessary for drafting patent documents.

[0135] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. The device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed description of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. Furthermore, if a product is primarily used by users of a specific gender, a specific age, or users living in a specific region, the device may present a screen for inputting gender, age, or address to the user asking the question, and determine the user's knowledge level based on the input information.

[0136] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The patent document drafting device of the present invention enables highly accurate question answering by using the large-scale language model, enabling efficient and accurate patent document drafting. A voice search system that utilizes artificial intelligence technologies such as voice recognition and natural language processing to extract necessary information from voice input, extract necessary information from relevant information and related information, etc., and present it as search results, and extract and display necessary information from information selected from the search results and related information, etc. In this way, the function of "extracting necessary information from voice input" makes it possible to present not only the "voice input" but also "necessary information" related to the "voice input" as search results. For example, if there is a voice input of "restaurants around Tokyo Station," not only the keyword "Tokyo Station area" but also the keyword "restaurant" can be extracted and presented as search results. The present invention relates to a voice search system, which utilizes artificial intelligence technologies such as voice recognition and natural language processing to extract necessary information from voice input, extract necessary information from relevant information and related information, present the extracted information as search results, and extract and display necessary information from information selected from the search results and related information.

[0137] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, an answer is generated using the large-scale language model. The patent document is created using the large-scale language model based on the patent information and the answer from the user or an automatically generated answer. Instructions for correcting the patent document may be provided to the user by providing a guide so that the user can input each item that makes up the patent document. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires information related to patents to be described in the patent document; a question generation unit that generates a question using the large-scale language model from the information acquired by the acquisition unit; an answer generation unit that generates an answer using the large-scale language model when no answer is obtained from a user to the question generated by the question generation unit; a storage unit that saves the answer acquired by the answer acquisition unit when an answer is obtained from a user to the question generated by the question generation unit; and a creation unit that creates a patent document consisting of the answer saved in the storage unit or the answer generated by the answer generation unit, the question generated by the question generation unit, and the information acquired by the acquisition unit. The device of the present invention uses a large-scale language model to assist in the creation of patent documents. The device begins with an information gathering section, in which the device collects necessary information (e.g., the inventor's name and the type of invention). Next, the device automatically generates a question set (e.g., "What is the inventor's name?") from this information. If the user does not respond to this set (e.g., "I don't know those details"), the device automatically generates a reply set (e.g., "The inventor's name is unknown"). If the user responds to the reply set (e.g., "We don't know those details"), the device saves the reply. Finally, the device creates a patent document consisting of the saved and automatically generated replies.

[0138] The present invention relates to a patent document drafting device. According to the present invention, there is provided a patent document drafting device that supports the drafting of patent documents using a large-scale language model. The patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model. The answer acquisition unit receives an answer to the first question from a user. The answer generation unit generates an answer to the first question when an answer to the first question is not received. The creation unit creates a patent document using the large-scale language model. a display control unit that creates and displays the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user or a first automatic answer to the first question generated by the answer generation unit; a control unit that controls the display content to change and be displayed according to an item selected on a screen displayed by the display control unit; a display content change control unit that controls the display content to change and be displayed according to the content of the item selected on a screen controlled by the control unit; and an input field change control unit that controls the number and type of input fields to change and be displayed according to the content of the item selected on a screen controlled by the control unit. The present invention is provided as a means for solving technical problems in the field of "large-scale language models," and known technical means and general knowledge in the field of "large-scale language models" are outside the scope of the present invention.

[0139] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires an answer to the first question from a user. If an answer to the first question is not acquired from the user, the answer generation unit generates an answer to the first question using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0140] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it is composed of an acquisition unit that acquires patent information, a question generation unit that generates a first question, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer when an answer is not available, and a creation unit that creates a patent document. The answer acquisition unit determines whether an answer to the first question has been received from a user terminal. According to the present invention, the use of a large-scale language model makes it possible to draft patent documents with high accuracy. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention is an "apparatus" that can "create" highly accurate "patent documents" by utilizing "large-scale language model" technology. Specifically, it is composed of an "acquisition unit," "question generation unit," "answer acquisition unit," "answer generation unit," and "creation unit." The "answer acquisition unit" is "determined" in "step S14."

[0141] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, patent information is acquired and a first question is generated using the large-scale language model. If an answer to the first question cannot be acquired, an answer is generated using the large-scale language model. Then, if an answer to the first question is acquired, the answer and the large-scale language model are used to create a patent document. The patent document drafting device of the present invention can also select documents for which tags are adjusted based on the user's behavior history.

[0142] The present invention is a patent document drafting device that supports the drafting of patent documents using a large-scale language model, and includes: an acquisition unit that acquires patent information related to patents described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit.

[0143] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the creation of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, an answer acquisition unit that acquires an answer to a first question generated using the large-scale language model, an answer generation unit that generates an answer to the first question generated using the large-scale language model if an answer to the first question cannot be acquired, and a creation unit that creates a patent document using the large-scale language model based on the patent information and a first user answer or a first automatic answer. The patent document drafting device of the present invention supports the creation of patent documents and enables efficient creation of patent documents.

[0144] The patent document drafting device according to the present invention includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired from the user, and a drafting unit that creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. For example, the answer generation unit matches an input question sentence with text data included in each item of recorded document information to extract multiple items related to the question sentence, generates an answer (first answer) including one or more levels leading to each of the extracted items and their headings, and returns the answer to the question sentence to the user.

[0145] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a question using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer using the large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The answer generation unit shown in the figure matches the input question with text data included in each item of recorded document information to extract multiple items related to the question, generates answer output information including one or more levels leading to each extracted item and their headings, and returns the answer to the user as an answer to the question.

[0146] The patent document creation device according to the present invention includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if a first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The answer generation unit also expands the patent information into multiple answer category information based on predetermined expansion rules, and searches an answer candidate database using the multiple answer category information obtained by this expansion and documents included in the search results sent from the document search unit as keys to obtain answer candidate information.

[0147] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. If an answer to the first question is not acquired from the user, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0148] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires a first answer from a user, an answer generation unit that generates a first automatic answer if the first answer cannot be acquired, a creation unit that creates a patent document using the large-scale language model, and a service control unit. If the service control unit acquires an answer result from a server to which an answer request has been made and no answer is obtained, the service control unit moves to another answer service candidate if one is available, or sends information to the user terminal indicating that an answer cannot be obtained if no other answer service candidate is available. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit; and a service control unit. If there are other answer service candidates, the service control unit moves to that service (Yes); if there are no other answer service candidates, the service control unit sends information to the user terminal indicating that an answer is not available (No).

[0149] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. The device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer when an answer to the first question is not acquired from the user. The creation unit creates a patent document using a large-scale language model. According to the present invention, the use of a large-scale language model enables highly accurate question answering, improving the efficiency of patent document drafting. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention enables highly accurate question and answering by utilizing large-scale language model technology, and eliminates the need for the cumbersome procedures traditionally performed manually, enabling efficient and rapid creation of patent documents. Furthermore, in embodiments of the present invention, an input question is matched with text data contained in each item of recorded document information, and an answer is generated that includes one or more levels leading up to each extracted item and their headings, which is expected to provide high-quality and highly accurate answers.

[0150] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents is acquired, a question is generated based on the information using a large-scale language model, and an answer is obtained from a user. If an answer cannot be obtained, an answer is generated using the large-scale language model. Then, a patent document is drafted based on the acquired information and answer. The present invention makes it easier to draft patent documents and enables highly accurate question and answering. A unified account management system for unifiedly managing account information of multiple financial institutions, wherein the account information includes information such as account number, account holder name, balance, etc., and the multiple financial institutions are financial institutions such as banks. The present invention relates to a unified account management system for unified management of account information from multiple financial institutions. According to the present invention, account information held by financial institutions such as banks can be unified. Specifically, account information collected from each financial institution can be integrated and managed centrally. This allows customers to centrally check all of the account information they hold, improving convenience. Furthermore, it also makes it easier for financial institutions to share information collected from customers, which is expected to improve customer service and strengthen risk management.

[0151] The present invention relates to a patent document drafting device. According to the present invention, it is possible to support the drafting of patent documents using a large-scale language model. Specifically, patent information is acquired, a first question is generated, and an answer is acquired from a user. If an answer cannot be acquired, an answer is generated using a large-scale language model. Then, a patent document is drafted based on the patent information and the answer. In this way, patent documents can be drafted efficiently and accurately. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention provides a device and method for assisting in the creation of "patent documents" by utilizing a "large-scale language model." Specifically, by utilizing a "large-scale language model," "questions" and "answers" such as "detailed explanations" are automatically, accurately, quickly, appropriately, efficiently, simply, inexpensively, and reliably generated from "patent information," thereby providing a device and method for assisting in the creation of "patent documents."

[0152] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information and the first user answer or an automatically generated first automatic answer.

[0153] The patent document creation device comprises an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The creation unit may create a patent document using a large-scale language model based on the patent information, a first user answer or a first automatic answer, and a second user answer to a second question acquired from a user or a second automatic answer to the second question generated by the answer generation unit. This can assist in the creation of patent documents.

[0154] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. The patent document drafting device of the present invention includes an acquisition unit, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit acquires information about patents. The question generation unit generates a first question using the large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. In the examples of this invention, the word "prompt" appears, but the word "prompt" can mean "waiting for input" or "waiting for input", etc. Therefore, care must be taken in the meaning of the word "prompt".

[0155] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0156] The patent document creation device includes an acquisition unit that acquires patent information related to a patent to be described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired from the user, and a creation unit that creates a patent document using a large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The creation unit combines the acquired {patent information} and {answer to the first question} (first user answer) with a prompt that uses {patent information} and {answer to the first question} in a prompt that is stored in advance in a prompt table, and inputs the combined information into the large-scale language model.

[0157] The patent document drafting device includes an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a drafting unit. In response to acquiring patent information from a user terminal, the drafting unit may input prompts containing {patent information} into a large-scale language model and create items obtained from each prompt. For example, these items may include "title of invention," "outline of invention," "embodiment," and "explanation of drawings." This allows users to easily draft patent documents. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The creation unit may provide a template document consisting of items such as {Name of the invention}, "Summary of the invention," "Embodiment," and "Explanation of the drawings," and may automatically generate sentences using a large-scale language model according to this template document.

[0158] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents can be acquired, and questions can be generated and answers can be obtained using a large-scale language model. Furthermore, if an answer cannot be acquired, an answer can be generated using the large-scale language model. Furthermore, patent documents can be drafted based on the acquired information and answers. This allows for efficient and accurate drafting of patent documents. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0159] 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0160] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired, and a creation unit that creates a patent document using the large-scale language model based on the patent information and a first user answer or a first automatic answer. This configuration improves the efficiency of patent document drafting.

[0161] The record document creation unit extracts necessary blocks from the fixed phrase information based on the setting information and question information, reads information to be filled in the blanks of the extracted blocks from the storage device, and fills each blank, thereby creating a record document (diary) (step S2).The record document creation unit also extracts necessary blocks from the fixed phrase information based on the setting information and question information, reads information to be filled in the blanks of the extracted blocks from the storage device, and fills each blank, thereby creating the record document (diary) (step S2).

[0162] The patent document creation device uses a large-scale language model to support the creation of patent documents. Information about the patent to be described in the patent document is acquired, and a question generation unit generates a first question requesting a detailed explanation. An answer acquisition unit acquires an answer from the user, and if no answer is available, the answer generation unit generates an answer. The patent document is then created using the large-scale language model. A correction unit may cause the large-scale language model to correct the patent document using a correction instruction sent from the user terminal as a prompt.

[0163] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it acquires patent information, generates questions using the large-scale language model, and acquires answers from the user. If no answers are available, it generates answers using the large-scale language model. Then, it creates a patent document based on the patent information and answers. In addition, the patent document drafting device of the present invention can learn the user's revision tendencies and set new prompts based on the content and frequency of revisions made by the revision unit.

[0164] The patent document creation device includes an acquisition unit that acquires information about the patent to be described in the patent document, a question generation unit that presents a first question generated using a large-scale language model to a user, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer using the large-scale language model if there is no answer, and a creation unit that creates the patent document using the large-scale language model. The created patent document can also be modified using the large-scale language model.

[0165] When displaying a document containing specific answer information, the answer information and the document portions other than the answer information can be displayed in a distinguishable manner. For example, the answer information and the document portions other than the answer information can be displayed in different colors. Alternatively, the font size or font type can be changed. Alternatively, only the answer information portions can be underlined. Alternatively, only the answer information portions can be framed or shaded. Alternatively, the document portions other than the answer information can be displayed in a lighter color than the answer information.

[0166] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer is not acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A feature of the present invention is that it is an automatic response system using a large-scale language model.

[0167] a first question requesting a detailed explanation of the patent based on the patent information, using the large-scale language model; an answer acquisition unit for acquiring an answer to the first question from a user; an answer selection unit for selecting an optimal answer candidate from answer candidates generated using the large-scale language model if an answer to the first question is not acquired from the user; an answer generation unit for generating an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit for creating the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or the optimal answer candidate selected by the answer selection unit.

[0168] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired, and a creation unit that creates a patent document using the large-scale language model based on the patent information and a first user answer or a first automatic answer. The patent document drafting device of the present invention uses a large-scale language model to enable the rapid drafting of patent documents with high accuracy. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a patent document drafting device that uses the large-scale language model to draft the patent document based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. An apparatus for supporting the creation of patent documents using a large-scale language model, the apparatus having the following functions: An extraction unit that extracts necessary information, knowledge, etc. from a knowledge graph constructed from a dictionary or knowledge data consisting of a large-scale language model learned from the knowledge and experience of experts in a specific theme or field; An extraction result display unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc.; A selection unit that selects necessary information, knowledge, etc. from the information, knowledge, etc. displayed in the extraction result display unit; An editing unit that edits necessary information, knowledge, etc. from the selected information, knowledge, etc.

[0169] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question using a large-scale language model if an answer to the first question is not acquired, a creation unit that creates a patent document using the large-scale language model based on the patent information and a first user answer or a first automatic answer, a final answer storage unit, a final answer correction unit, and a final answer change unit. The patent document drafting device of the present invention supports the drafting of patent documents and enables efficient drafting of patent documents. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a patent document drafting device that uses the large-scale language model to draft the patent document based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. An automatic summarization system using a large-scale language model, comprising an extractor as a processing means for extracting from an input text parts that are of high importance and should be summarized and their contents, an excluder as a processing means for extracting from an input text parts that are of low importance and should be summarized and their contents, a summarizer as a processing means for generating a summary sentence from the contents extracted by the extractor and excluder, a display device as a processing means for displaying the input text and the summary sentence, and a storage device as a processing means for saving the input text and the summary sentence.

[0170] The patent document drafting device according to the present invention comprises an acquisition unit that acquires patent information, a question generation unit that generates questions using a large-scale language model, an answer acquisition unit that acquires answers, an answer generation unit that generates answers when answers cannot be acquired, and a drafting unit that drafts patent documents using a large-scale language model. The present invention makes it possible to draft patent documents easily and efficiently. A sentence suitable for an example patent specification, related to the content of claim 1: In an embodiment of the present invention, the necessary information is collected from the patent office database to obtain patent information. The question generation unit automatically generates a first question using a large-scale language model. The answer acquisition unit collects answers from the user, and if no answer is obtained, the answer generation unit automatically generates an answer. The creation unit then automatically creates a patent document using a large-scale language model. All of these processes are automated, eliminating the need for manual work. As such, the patent document creation device of the present invention is highly efficient and can significantly reduce the time required to create patent documents.

[0171] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired, and a creation unit that creates a patent document based on the patent information and a first user answer or a first automatic answer. The patent document drafting device of the present invention uses a large-scale language model, enabling the drafting of patent documents with high accuracy and speed. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and creating the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0172] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired, and a creation unit that creates a patent document based on the patent information and a first user answer or a first automatic answer. The patent document drafting device of the present invention uses a large-scale language model, enabling the drafting of patent documents with high accuracy and speed. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a patent document drafting device that uses the large-scale language model to draft the patent document based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. An apparatus for supporting the creation of patent documents using a large-scale language model, the apparatus having the following functions: An extraction unit that extracts necessary information, knowledge, etc. from a dictionary, knowledge database, etc. consisting of a large-scale language model learned from the knowledge and experience of experts in a specific theme or field. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc. An extraction unit that extracts necessary information, knowledge, etc. from the extracted information, knowledge, etc.

[0173] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0174] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer when an answer to the first question is not acquired from the user. The creation unit creates a patent document using a large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0175] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in the patent document, a question generation unit that generates a first question requesting a detailed description of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if no answer is acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. Furthermore, if two or more additional terms are acquired, the two or more additional terms may be added to all of T, or the device may repeatedly add one of the two or more acquired additional terms to T, calculate the value of the above-mentioned formula, and then delete the added additional term, add a new additional term to T, and calculate the value of the above-mentioned formula.

[0176] The question-answering system of the present invention can generate answers to questions from users using a large-scale language model. Specifically, the system is realized by including an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit in a patent document drafting device. Furthermore, the answer filter unit can notify relevant answers. This configuration allows users to easily and quickly obtain the information necessary for drafting patent documents.

[0177] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. Points may be added or subtracted based on a suitability evaluation of a new FAQ registered in the FAQ-DB, and the FAQ may be deleted if the total point falls below zero.

[0178] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. The patent document drafting device of the present invention includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using the large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. A speech recognition system comprising: a speech recognition engine that outputs text data from speech input speech data; a natural language processing engine that performs natural language processing on the text data; a control unit that causes the speech recognition engine to perform analysis processing of the speech input speech data again based on the analysis results output by the natural language processing engine; a control information storage unit that stores control information such as parameter values ​​used by the control unit to perform analysis processing of the speech input speech data again based on the analysis results; and control means that controls the control unit to perform the analysis processing based on the control information stored in the control information storage unit. In the embodiments of the present invention, examples of "control information such as parameter values" include parameter values ​​such as "volume," "speed," and "pitch," as well as dictionary information such as "word dictionary" and "grammar dictionary," but these are merely examples and are not limiting. The present invention relates to a speech recognition system. The speech recognition system of the present invention includes a speech recognition engine that outputs text data from speech input data, a natural language processing engine, a control unit, a control information storage unit, and control means. The natural language processing engine performs natural language processing on the text data. The control unit performs analysis processing on speech input data again based on the analysis results.

[0179] In a document information classification device comprising a classification means for classifying document information based on a document information classification request, an attribute assignment means for assigning attribute information associated with 111cS1 to a set of document information including document information by the classification means, a search means for searching the set of document information using the attribute information assigned by the attribute assignment means, and a display means for displaying the classification result for the set of document information searched by the search means using the attribute information, the notification means can notify the user to suggest assigning the new attribute information only when the user determines that new attribute information should be assigned to the classification result. A recommendation system that extracts user preferences related to specific products or services from the user's behavior history data related to the products or services, and recommends products or services based on the preferences, comprising: a keyword extraction unit that extracts keywords related to the specific products or services from the user's behavior history data; a preference extraction unit that extracts the user's preferences from the keywords extracted by the keyword extraction unit; a recommendation unit that recommends products or services based on the preferences extracted by the preference extraction unit; an update unit that, when a product or service recommended by the recommendation unit is purchased, obtains purchase history data and updates the user's preferences from the purchase history data; and a control unit that is configured to recommend products and services based on the preferences updated by the update unit. The present invention is a system that can extract user preferences related to specific products or services from the user's behavioral history data related to those products or services, and can recommend products or services based on those preferences.The present invention also provides a system that can update user preferences from purchase history data and perform recommendation processing again.

[0180] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer cannot be acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. In addition, a screen such as "Answer generating" may be displayed until the answer is ready, and the answer may be displayed once it has been generated.

[0181] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed description of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. Furthermore, if there are no other basis documents in the answer history, instead of terminating the search process at that point, new document information unrelated to the answer history may be acquired from a document database and an answer element may be searched for from the document information.

[0182] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The device can also detect relevant answers from searched answer candidates.

[0183] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it acquires patent information, generates questions using the large-scale language model, and acquires answers from the user. If no answers are available, it generates answers using the large-scale language model. Then, it creates a patent document based on the patent information and answers. In addition, the patent document drafting device of the present invention can learn the user's revision tendencies and set new prompts based on the content and frequency of revisions made by the revision unit.

[0184] The patent document creation device acquires information about the patent to be described in the patent document, generates a question using a large-scale language model, and acquires an answer from the user. If an answer cannot be acquired, the device generates an answer using the large-scale language model. Then, the patent document is created using the large-scale language model based on the patent information and the answer from the user or the automatically generated answer. The server can further train the trained model using newly added combinations of questions and answers.

[0185] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. It receives an answer from the user, but if no answer is received, it generates an answer using the large-scale language model. It creates a patent document using the large-scale language model based on the patent information and the answer from the user or the answer automatically generated.

[0186] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0187] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired, and a creation unit that creates a patent document using the large-scale language model. Furthermore, the statistical data generated by the answer generation means described above with reference to the drawings may be used as a primary answer to be presented to a user, and the user may be allowed to select a predicate as an answer from this list of primary answers. When a specific predicate is selected by the user, a passage search may be performed again together with the question keyword, and elements (subject and modifiers) related to the selected predicate may be extracted by a syntactic and semantic analysis unit and presented to the user.

[0188] The patent document drafting device according to the present invention comprises an acquisition unit that acquires patent information, a question generation unit that generates a first question based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question is not acquired, and a creation unit that creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. Also, the patent document drafting method according to the present invention is a method using the above device. a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit; a judgment display unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit, and determines whether the content of the patent document is correct and displays the result; and a correction request unit that requests further correction if the content displayed by the judgment display unit is inaccurate. The present invention provides a system and method for supporting the creation of "large-volume, high-quality" patent documents by utilizing technical means such as a "large-scale language model" and an "automatic response system." Furthermore, the system and method also improve "accuracy," "efficiency," "reliability," "flexibility," "scalability," "adaptability," "ease of use," "response speed," and "cost performance."

[0189] An information processing device according to one embodiment of the present invention includes an input unit, an output unit, and a processing unit. The input unit accepts input from a user. The output unit outputs processing results. The processing unit includes a search unit that searches for information that meets specific conditions based on the input information, an extraction unit that extracts information that meets the specific conditions based on the searched information, an analysis unit that analyzes the information that meets the specific conditions based on the extracted information, and a display unit that displays the information that meets the specific conditions based on the analysis results. A speech recognition device comprising: a speech recognition engine that receives a speech signal and generates text data from the speech signal; and a dialogue engine that generates a dialogue context from the text data and generates a response context based on the dialogue context. The present invention provides a speech recognition device that can generate text data from a speech signal uttered by a user, generate a response context from a dialogue context, and determine a response content. With this configuration, it is expected that an appropriate response content will be returned in response to the content uttered by the user. An image processing device, characterized in that the image processing device extracts feature data from image data, generates an object detection result from the feature data, obtains object position information and object type information from the object detection result, and processes the image data based on the object position information and object type information. The present invention provides an image processing device in the field of image processing technology that can extract feature data such as object position information and object type information from image data and perform various processes such as image data processing based on object detection results and various information such as object position information and object type information.

[0190] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, an answer is generated using the large-scale language model. The patent document is created using the large-scale language model based on the patent information and the answer from the user or an automatically generated answer. Instructions for correcting the patent document may be provided to the user by providing a guide so that the user can input each item that makes up the patent document. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires information related to patents to be described in the patent document; a question generation unit that generates a question using the large-scale language model from the information acquired by the acquisition unit; an answer generation unit that generates an answer using the large-scale language model when no answer is obtained from a user to the question generated by the question generation unit; a storage unit that saves the answer acquired by the answer acquisition unit when an answer is obtained from a user to the question generated by the question generation unit; and a creation unit that creates a patent document consisting of the answer saved in the storage unit or the answer generated by the answer generation unit, the question generated by the question generation unit, and the information acquired by the acquisition unit. The device of the present invention uses a large-scale language model to assist in the creation of patent documents. The device begins with an information gathering phase, in which the device collects necessary information (e.g., inventor's name and type of invention). Next, the device automatically generates a question (e.g., "What is this invention?") from this information. If the user does not respond to this question, an automatic response such as "This invention is ____" is generated. If the user responds, the response is saved. Finally, the device automatically creates a complete patent document (e.g., "title of invention," "claims," ​​"details of invention," etc.) from this information.

[0191] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if the answer is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. Furthermore, an answer deadline setting and management subprogram manages the deadline based on the answer deadline set in the figure, and has a function of sending a reminder email to the answerer if the answer (second answer information) is not transmitted to the question and answer server within the deadline.

[0192] A question-answering system generates answers to questions from users by referencing past question histories. In this process, weights are assigned to the basis documents in the answer history, and by referencing the basis documents in descending order of weight, a more appropriate answer can be generated. Furthermore, because the weights of the basis documents in the answer history are automatically set based on the past question histories, the user does not need to set them manually, enabling more efficient question-answering. The speech recognition device includes a speech recognition unit that receives speech data input from a speech input unit, analyzes the speech data, and converts it into text data; an intention identification unit that analyzes the text data and identifies an intention; a response generation unit that generates a response based on the intention identified by the intention identification unit; and a speech synthesis unit that converts the response generated by the response generation unit into speech data and outputs it. The present invention relates to a speech recognition device that combines speech recognition technology and natural language processing technology. This device receives speech input from a user, analyzes the content to identify the intent, and generates an appropriate response. The generated response is then converted into voice data using speech synthesis technology and output. This configuration allows users to exchange information using natural speech, enabling smoother and more efficient communication. The image processing device includes an object detection unit that detects and extracts an object region from image data, a feature extraction unit that extracts features within the object region, an estimation unit that estimates object type information by comparing the features extracted by the feature extraction unit with a trained classifier, a correction unit that corrects attribute information such as color information within the object region in the image data based on the object type information estimated by the estimation unit, and an output unit that outputs the corrected image data. The present invention relates to the field of image processing technology, and relates to an image processing device and method capable of extracting attribute information (for example, color information) within an object area from image data with high accuracy and high speed.

[0193] In the present invention, if a next question is set in the option data (302 in the figure), the user is guided to the question specified by the next question (301 in this case), and if an FAQ ID is set (303 in the figure), the user is guided to the FAQ specified by the FAQ ID (201 in this case). This makes it possible to quickly and accurately provide the information the user desires. Furthermore, since the option data allows the user to accurately grasp the information they desire, the burden on the user can be reduced. Furthermore, since the present invention can reduce the burden on the system developer, it can also contribute to cost reduction on the system developer's side. The patent document creation device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates an answer using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer using the large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention can provide information quickly and accurately by using information such as option data and FAQ IDs. It also reduces the burden on system developers, resulting in cost savings.

[0194] The present invention relates to a patent document drafting device that supports the creation of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents can be acquired, and a large-scale language model can be used to generate questions and obtain answers based on the acquired information. Furthermore, if an answer cannot be obtained, an answer can be generated using the large-scale language model. Furthermore, a patent document can be created based on the acquired information and answers. In this way, the creation of patent documents can be supported.

[0195] A question-answering system generates answers to questions from users by referencing past question histories. In this process, weights are assigned to the basis documents in the answer history, and by referencing the basis documents in descending order of weight, a more appropriate answer can be generated. Furthermore, because the weights of the basis documents in the answer history are automatically set based on the past question histories, the user does not need to set them manually, enabling more efficient question-answering. The speech recognition device includes a speech recognition unit that receives speech data input from a speech input unit, analyzes the speech data, and converts it into text data; an intention identification unit that analyzes the text data and identifies an intention; a response generation unit that generates a response based on the intention identified by the intention identification unit; and a speech synthesis unit that converts the response generated by the response generation unit into speech data and outputs it. A speech recognition device analyzes input speech data, converts it into text data, identifies the user's intent from the text data, and generates an appropriate response. The intent identification unit uses natural language processing techniques such as syntactic analysis and semantic analysis, while the response generation unit may use techniques such as template matching and machine learning. Furthermore, the output response is generated as natural speech using speech synthesis technology, allowing users to communicate smoothly and naturally. The image processing device includes a convolution operation unit that receives image data input from an image input unit and performs a convolution operation on the image data, a pooling processing unit that performs a pooling process on the image data processed by the convolution operation unit, a fully connected layer that processes the image data processed by the pooling processing unit in a fully connected layer, and a classifier that outputs a classification result from the image data processed by the fully connected layer. Image processing devices use convolutional neural network (CNN) algorithms, such as convolution and pooling operations, to perform image classification tasks.

[0196] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. A weighting setting unit can also change the importance of question-related items extracted by matching with headlines and question-related items extracted by matching with text data included in each item of recorded document information.

[0197] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. If a first user answer is not acquired, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The setting unit can accept changes to prompts or settings of new prompts. In claim 1, "the large-scale language model" is defined in some way. For example, the large-scale language model is considered to be constructed using language processing technology such as "GPT-3."

[0198] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents can be acquired, and questions can be generated and answers can be obtained using a large-scale language model. Furthermore, if an answer cannot be acquired, an answer can be generated using the large-scale language model. Furthermore, patent documents can be drafted based on the acquired information and answers. The present invention enables efficient and accurate drafting of patent documents. A question analysis device that estimates the intent of an input question and determines the priority of answer candidates based on the estimated intent. The weight setting unit calculates a weight for each word included in the input question from the frequency of appearance of the word and semantic information of the word, and determines the priority of the answer candidates based on the calculated weight for the word. The present invention relates to a question analysis device that estimates the intent of an input question and determines the priority of answer candidates based on the estimated intent. According to the present invention, it is possible to calculate a weight for each word from the input question and determine the priority of answer candidates based on the weight. It is also possible to determine the priority of answer candidates based on the estimated intent. Furthermore, by displaying answer candidates in descending order of priority, users can quickly and accurately access the information they are looking for. The present invention enables efficient and accurate information searches.

[0199] a patent document drafting device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The patent document drafting device is configured to, when basis document information such as "3291.txt" is input, refer to a weight setting list that indicates a weight value of "1.5", and notify the answer formation unit after being notified of the weight value of "1.5".

[0200] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it acquires patent information, generates questions using the large-scale language model, and acquires answers from the user. If no answers are available, it generates answers using the large-scale language model. Then, it creates a patent document based on the patent information and answers. In addition, the patent document drafting device of the present invention can learn the user's revision tendencies and set new prompts based on the content and frequency of revisions made by the revision unit.

[0201] A text data processing device comprising a weighting setting unit that weights items extracted based on a heading and items extracted based on text data of a structural unit located at the end of a hierarchical structure for specific text data, wherein the weighting setting unit sets the importance of the items extracted based on the heading to be higher than the importance of the items extracted based on text data of a structural unit located at the end of the hierarchical structure.

[0202] The present invention relates to a patent information collection device comprising a setting unit, a collection unit, a question generation unit, an answer acquisition unit, and an answer generation unit. The setting unit sets record items and a collection method. The collection unit collects patent information based on the set record items and collection method. The question generation unit generates a question based on the patent information. The answer acquisition unit acquires an answer from a user. The answer generation unit automatically generates an answer when there is no answer from the user. The present invention enables efficient collection of patent information and question answering. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention is a patent document creation support device that utilizes large-scale language model technology. Specific components include an acquisition unit, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit collects patent information, the question generation unit automatically generates questions for detailed explanations using large-scale language model technology, the answer acquisition unit functions as an interactive interface with the user, the answer generation unit functions as an automatic response system, and the creation unit automatically creates patent documents using the large-scale language model, the collected information, and the content of the dialogue.

[0203] The present invention relates to an apparatus and method for supporting the creation of a record document. According to the present invention, detailed settings can be made for creating a record document, such as the contents of the record document (record items), the method for collecting information to be recorded in the record document, and the questions to be asked to the user (question items). Furthermore, the record document can be automatically created based on the answers given by the user. This can significantly reduce the time and effort required to create a record document. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. This invention provides a patent document creation support device and method that utilizes large-scale language model technology. Specifically, questions such as points requiring detailed explanation or points of doubt are automatically generated from patent information, and patent documents are automatically created from the answers to those questions. Furthermore, if necessary, more accurate information collection and analysis can be achieved by manually entering information or selecting appropriate information from the automatically generated information. These functions are expected to have the effect of improving the efficiency and quality of work that was previously done manually.

[0204] The present invention relates to a patent document drafting device equipped with a weighting setting unit that calculates the frequency of occurrence of words contained in character information within unit information and sets a weight for each word based on the calculated frequency of occurrence of the words. The weighting setting unit calculates the frequency of occurrence of words contained in character information within unit information and sets a weight for each word based on the calculated frequency of occurrence of the words. This allows the patent document drafting device to extract more appropriate keywords when drafting a patent document. Furthermore, the present invention allows the patent document drafting device to extract more accurate keywords when drafting a patent document. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to patents described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit; and a weighting setting unit. The weighting setting unit performs weighting processing using a method called "TF-IDF." The present invention relates to a patent document preparation device equipped with a weighting setting unit that performs weighting processing using a method called "TF-IDF." The weighting processing is performed using a method called "TF-IDF." This enables the patent document preparation device to extract keywords that increase the importance of each keyword using a method called "TF-IDF." The present invention also enables the patent document preparation device to extract keywords that increase the importance of each keyword using a method called "TF-IDF."

[0205] The patent document creation device can acquire information about the patent to be described in the patent document, generate a question using a large-scale language model, and acquire an answer. If an answer is not available, the device can generate an answer using the large-scale language model. Furthermore, the device can create a patent document based on the acquired information and answer. In this way, the patent document creation device can assist in the creation of patent documents.

[0206] The present invention provides a patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. When a user taps (clicks) a question content in the content, the terminal device requests, from an information processing server, question posting content to answer the question indicated in the question content tapped by the user.

[0207] The dialogue management server device of the present invention comprises a dialogue management unit for conducting dialogue with a user terminal device, a transmission unit for transmitting the response content generated by the dialogue management unit to the user terminal device, a receiving unit for receiving information transmitted from the user terminal device, and a control unit for changing the response content generated by the dialogue management unit based on the information received by the receiving unit. The user terminal device also includes a display unit that displays the stamps, a transmission unit that transmits a stamp selected from the stamps, and a display unit that displays the response content transmitted from the dialogue management server device.

[0208] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained, the device also generates an answer using the large-scale language model. The device then creates a patent document based on the acquired information and the answer. In this way, the creation of patent documents can be supported.

[0209] The dialogue system according to the present invention is executed between a user terminal device and an operator terminal device of an operator. It may also be executed via a communication device other than the user terminal device or the operator terminal device. When a user makes an inquiry to an operator, the operator uses a large-scale language model to generate a response to the user's inquiry. Furthermore, when a user responds to an inquiry from the operator, the response content is generated using the large-scale language model. This allows the operator to collect necessary information and the user to obtain necessary information. The patent document creation device described in claim 1 comprises an acquisition unit that acquires patent information related to patents described in patent documents, a question generation unit that generates a question based on the patent information using a large-scale language model, an answer acquisition unit that acquires an answer to the question, an answer generation unit that generates an answer to the question when no answer is obtained from the answer acquisition unit, and a creation unit that creates a patent document from the patent information, the question, and the answer using a large-scale language model. With this configuration, it is possible to extract only necessary information from a large number of patent documents and create patent documents easily and quickly. The patent document drafting device described in claim 1 can automatically generate questions and answers using a large-scale language model and create patent documents from that information. With this configuration, there is no need to manually create questions, answers, and other documents, making it possible to create patent documents efficiently and accurately.

[0210] The patent document drafting device according to the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it acquires information about patents to be described in patent documents, and generates questions based on that information using a large-scale language model. If an answer cannot be obtained, it also generates an answer using the large-scale language model. Then, it creates a patent document based on the acquired information and the answer. In some cases, a question may be sent directly from a user system. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires information about patents to be described in the patent document; a display unit that displays questions generated using the large-scale language model based on the information; an input unit that inputs answers to the questions; a determination unit that determines whether the input answers are correct, and if correct, displays the next question, or if incorrect, displays the same question again; and a creation unit that creates the patent document using the large-scale language model when all questions have been asked. The user enters an answer to the displayed question in the answer input field and sends it by clicking the "Submit" button. The judgment unit then determines whether the answer is correct, and if it is correct, the next question is displayed. If it is incorrect, the same question is displayed again. Once all questions have been answered, a patent document is created using a large-scale language model.

[0211] The terminal device comprises a communication unit, an input unit, a memory unit, an output unit, and a control unit. The communication unit communicates with external devices. The input unit accepts input from a user. The memory unit stores programs and data. The output unit outputs information to the user. The control unit controls each unit. In the patent document drafting device according to claim 1, the large-scale language model is BERT. 1. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model (BERT); an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model (BERT) if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model (BERT) based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The "large-scale language model" in claim 1 is BERT.

[0212] The patent document drafting device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, the device generates an answer using the large-scale language model. The patent document is then drafted using the large-scale language model. This device can be provided with a screen that is displayed in common on both the user terminal device and the operator terminal device.

[0213] The user terminal device and operator terminal device according to the present invention are devices that output dialogue content, such as information output devices such as laptop or notebook computers. The user terminal device and operator terminal device are equipped with communication means capable of communicating with a patent document drafting device. The patent document drafting device supports the drafting of patent documents using a large-scale language model, and includes: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit.

[0214] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The terminal device also includes an operation unit, a display unit, and an output unit.

[0215] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, an answer is generated using the large-scale language model. The patent document is created using the large-scale language model based on the patent information and the answer from the user or an automatically generated answer. Instructions for correcting the patent document may be provided to the user by providing a guide so that the user can input each item that makes up the patent document. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires information related to patents to be described in the patent document; a question generation unit that generates a question using the large-scale language model from the information acquired by the acquisition unit; an answer generation unit that generates an answer using the large-scale language model when no answer is obtained from a user to the question generated by the question generation unit; a storage unit that saves the answer acquired by the answer acquisition unit when an answer is obtained from a user to the question generated by the question generation unit; and a creation unit that creates a patent document consisting of the answer saved in the storage unit or the answer generated by the answer generation unit, the question generated by the question generation unit, and the information acquired by the acquisition unit. The device of the present invention uses a large-scale language model to assist in the creation of patent documents. The device begins with an information gathering section, in which the device collects necessary information (e.g., inventor's name and type of invention). Next, the device automatically generates a question (e.g., "What is this invention?") from this information. If the user does not respond to this question, the device outputs an automatic response such as "This invention is..." If the user responds, the response is saved. Finally, the device constructs the entire patent document from the collected information and automatic and manual responses.

[0216] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. The patent document drafting device of the present invention includes an acquisition unit, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit acquires information about patents. The question generation unit generates a first question using the large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. Although the word "prompt" is used in the embodiment of the present invention, the same function can be achieved without using the word "prompt."

[0217] The patent document creation device includes an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. In the patent information input field, it is possible to input questions and answers sent from the server, or to instruct the server to request an answer. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0218] The patent document drafting device according to the present invention prepares a patent document by acquiring information about the patent to be described in the patent document, generating a question using a large-scale language model, and acquiring an answer from the user. If an answer cannot be acquired, the device generates an answer using the large-scale language model. The answer entered in the answer input field is sent to the server in response to the selection of the "Send Answer" button.

[0219] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, an answer is generated using the large-scale language model. The patent document is created using the large-scale language model based on the patent information and the answer from the user or an automatically generated answer. Instructions for correcting the patent document may be provided to the user by providing a guide so that the user can input each item that makes up the patent document. a storage unit that stores the answer acquired by the answer acquisition unit when an answer is received from the user to the question generated by the question generation unit; a re-question processing unit that generates a question again using the large-scale language model when the answer stored in the storage unit is insufficient; an additional processing unit that, when an answer is received from the user to the question generated again by the re-question processing unit, uses the answer stored in the storage unit in combination with the answer received from the user to the question generated again by the re-question processing unit, and continues processing again by the re-question processing unit when the answer is insufficient; and a creation unit that creates a final patent document from the results processed by the additional processing unit. The device of the present invention is composed of the following components. First, an acquisition unit acquires information related to patent documents and generates a question from that information using a large-scale language model. If no answer is obtained from the user at this time, an answer is generated automatically. The question at this time and the automatically generated answer are also saved. If necessary, a new question is generated again using the large-scale language model. In this case, an answer may also be generated automatically.

[0220] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in the patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. Here, the "first question" and the "second question" may, for example, request a more detailed explanation from the user about the meaning of a word contained in the patent document or about what is achieved by a sentence contained in the patent document.

[0221] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, the device generates an answer using the large-scale language model. The patent document is then created based on the acquired information and the answer. At this time, the patent document display field on the operation screen displays the portion corresponding to the item selected in the item selection field. a first automatic answer to the first question generated by the answer generation unit; a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit; a template selection unit that selects and displays a template document from a group of template documents that includes at least a portion of the patent document created by the creation unit; a display control unit that displays the template document selected by the template selection unit in a format that includes at least a portion of the template document; and an output control unit that outputs the template document selected by the template selection unit in a format that includes at least a portion of the template document.

[0222] In response to a correction request from a user, the patent document drafting device can display a screen that allows the user to select the items that the user wants to correct and the correction contents.

[0223] The patent document display field is configured to display the "claims" of the patent document. In claim 1, the questions generated using a large-scale language model are automatically generated based on patent information. In claim 1, the questions generated using a large-scale language model are automatically generated based on patent information.

[0224] The patent document drafting device includes an acquisition unit that acquires patent information related to a patent to be described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired from the user, and a drafting unit that creates a patent document using a large-scale language model based on the patent information and the first user answer or the first automatic answer. Each component may be physically distributed across multiple computers or may be implemented as a single computer.

[0225] The patent document drafting support system includes at least a server (patent document drafting device) 100, a user terminal, and a large-scale language model server. The server 100 includes an acquisition unit that acquires information about patents, a question generation unit that generates questions using a large-scale language model, an answer acquisition unit that acquires answers, an answer generation unit that generates answers when there are no answers, and a creation unit that creates patent documents based on the patent information and answers from users. The user terminal transmits questions and answers to the server 100. The large-scale language model server provides the large-scale language model to the server 100. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a display unit that creates and displays the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention provides a system and method for supporting the creation of "patent documents" by utilizing "large-scale language model" technology. The "patent document creation device" of the present invention is composed of components such as an "acquisition unit," "question generation unit," "answer acquisition unit," "answer generation unit," and "display unit." The "method" of the present invention provides a method for supporting the creation of "patent documents" by utilizing "large-scale language model" technology. The "method" of the present invention is composed of a series of steps: an "acquisition step," an "question step," an "answer step," and a "display step."

[0226] The present invention relates to a patent document drafting device. According to the present invention, it is possible to support the drafting of patent documents using a large-scale language model. Specifically, patent information is acquired, a question is generated using a large-scale language model, and an answer is obtained from a user. If an answer cannot be acquired, an answer is generated using the large-scale language model. A patent document is then created based on the patent information and the answer. The patent document created in this way is sent to a user terminal, and the patent document display field on the operation screen is updated. The present invention makes it easier to create patent documents, enabling efficient patent document creation. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires information about patents described in the patent document; an answer acceptance unit that accepts answers to questions generated using the large-scale language model based on the information; an automatic response unit that automatically answers the questions generated using the large-scale language model when there is no answer from the answer acceptance unit; a creation unit that creates the patent document based on the information and the answer from the answer acceptance unit or the automatic response result from the automatic response unit; a storage device that stores a processing program that executes the acquisition unit and the creation unit; and a computer system that executes the processing program. The present invention relates to a patent document drafting device and method using a large-scale language model. Specifically, the device generates a question from acquired information using a large-scale language model, and if there is an answer or an automatic response result to the question, it can create and display a patent document from the answer or an automatic response result. If there is no automatic response result, an automatic response can be automatically performed and displayed. These functions are realized by a processing program executed on a computer system, a storage device, etc.

[0227] The present invention relates to a patent document creation support system. The patent document creation support system supports the creation of patent documents using a large-scale language model. The patent document creation support system includes an acquisition unit, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit acquires information about patents. The question generation unit generates a first question using the large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a computer system connected to a display device that can create and display the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. In the embodiment of the present invention, a "display device" and a "computer system" are added. These additional elements enable the "large-scale language model" to be used more effectively.

[0228] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the creation of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information, a question generation unit that generates a first question, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer when an answer cannot be acquired, and a creation unit that creates a patent document. The patent document drafting device of the present invention can support the creation of patent documents using a large-scale language model, making it possible to create patent documents with high accuracy and speed.

[0229] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained, the device generates an answer using the large-scale language model. The device then creates a patent document based on the acquired information and the answer. At this time, the patent document display field on the operation screen displays the part of the created patent document that corresponds to the item selected in the item selection field.

[0230] The patent document display field is configured to display the "claims" of the patent document. In claim 1, the questions generated using the large-scale language model are generated based on patent information, so it is expected that more accurate questions will be generated. Also, even if an answer cannot be obtained, a highly accurate answer can be obtained because an answer can be generated using the large-scale language model. Furthermore, since the patent document creation device creates patent documents using the large-scale language model, it is possible to create more accurate patent documents.

[0231] The patent document creation device includes an acquisition unit that acquires patent information, an answer acquisition unit that acquires answers to questions generated using a large-scale language model, an answer generation unit that generates an answer using the large-scale language model if there is no answer, a creation unit that creates a patent document, and a display processing unit. The display processing unit can display an operation screen. The operation screen includes a patent information input field, a patent document display field, and a prompt selection field. A user can input patent information in the patent information input field, select a question from the prompt selection field, and input an answer. In addition, the patent document created by the creation unit is displayed in the patent document display field. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit; and a display processing unit. In this device, an appropriate question can be selected from a "prompt selection field."

[0232] The patent document drafting device according to this embodiment is configured as shown in FIG. 1. The patent document drafting device 1 includes an acquisition unit 2, a question generation unit 3, an answer acquisition unit 4, an answer generation unit 5, and a creation unit 6. The acquisition unit 2 is a component for acquiring patent information, and can acquire patent information from, for example, a database of the Japan Patent Office. The question generation unit 3 is a component for generating a first question based on the patent information using the large-scale language model. The answer acquisition unit 4 is a component for receiving a first user answer from a user. The answer generation unit 5 is a component for generating a first automatic answer to the first question using the large-scale language model. The creation unit 6 is a component for creating a patent document using the large-scale language model. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a patent document drafting device that uses the large-scale language model to draft the patent document based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. 1 and 2, the present embodiment includes a non-factoid question information receiving device 10, which includes non-factoid question information receiving means 11, non-factoid question information analysis means 12, and non-factoid response means 13. The non-factoid question information receiving means 11 is a means for receiving non-factoid question information, such as voice input or keyboard input.

[0233] The search system of the present invention comprises a document analysis processing unit, a document index information storage unit, a question analysis unit, and a document search unit. The document analysis processing unit associates words obtained by morphological analysis with the frequency of appearance of the words and an index (document identifier) ​​for searching the searched document, and stores these as document index information. The question analysis unit analyzes a question from a client and obtains a word (or group of words) that is key to the question as a search term. The document search unit references the document index information storage unit and obtains a document identifier associated with the word corresponding to the search term. The patent document drafting device described in claim 1 supports the drafting of patent documents using a large-scale language model. Specifically, it is equipped with an acquisition unit that acquires patent information and a question generation unit that generates a first question, and if an answer is not obtained from the user, the answer generation unit can automatically generate an answer. It also has a creation unit that creates patent documents based on the large-scale language model, making it possible to draft patent documents with high accuracy and speed. The present invention aims to provide a search system that can quickly and accurately return answers to client inquiries. To this end, the present invention uses technologies such as morphological analysis and index creation to enable fast and accurate search processing. The present invention also provides a patent document creation device that uses a large-scale language model, which is expected to contribute to improving the efficiency of patent-related work.

[0234] The present invention relates to a dialogue processing device and a dialogue processing method that can accurately grasp the user's intention and provide an appropriate answer by generating a question summary for a question entered by a user using a large-scale language model and generating an answer based on the question summary. Specifically, the patent document drafting device includes an acquisition unit that acquires patent information, a question generation unit that generates a question summary using a large-scale language model, an answer acquisition unit, an answer generation unit, and a creation unit. In addition, if an appropriate answer is not obtained, log data is output by a log output processing unit. A sentence suitable for an example patent specification, related to the content of claim 1: In one embodiment of the present invention, a patent document drafting device is provided. The patent document drafting device includes an acquisition unit for acquiring patent information, a question generation unit for generating a question summary using a large-scale language model, an answer acquisition unit, an answer generation unit, and a creation unit. The patent document drafting device can provide an appropriate answer to a question entered by a user. Furthermore, if an appropriate answer is not obtained, log data is output by a log output processing unit.

[0235] The patent document drafting device of the present invention supports the creation of patent documents using a large-scale language model. Specifically, it includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if no answer is acquired from the user, and a creation unit that creates a patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The patent document drafting device of the present invention, with the above-described configuration, is capable of supporting the creation of patent documents with high accuracy and speed.

[0236] The patent document creation device according to the present invention comprises an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit for acquiring patent information collects information related to patents and provides it to the question generation unit. The question generation unit generates a first question using a large-scale language model and provides it to the answer acquisition unit. The answer acquisition unit receives a first user answer from a user and provides it to the creation unit. If there is no first user answer, the answer generation unit generates a first automatic answer and provides it to the creation unit. The creation unit creates a patent document using the large-scale language model.

[0237] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device comprises an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question cannot be acquired, and a creation unit that creates patent documents using the large-scale language model. By using a large-scale language model, the patent document drafting device of the present invention is capable of generating questions and answers with high accuracy, and is also capable of creating patent documents automatically. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: a display unit that displays at least one question summary and an annotated question summary selected from a question summary automatically generated by the large-scale language model from information about a patent described in the patent document and an annotated question summary automatically generated by the large-scale language model from the information; a receiving unit that receives an answer entered by a user to the at least one question summary and annotated question summary; a display unit that displays the received answer and an answer summary automatically generated by the large-scale language model from the at least one question summary and annotated question summary; and a display unit that displays the patent document automatically generated by the large-scale language model from the at least one question summary and annotated question summary and the answer summary. The present invention relates to a patent document drafting device and method, and more specifically, to a technology for supporting the drafting of patent documents using a large-scale language model. The present invention is capable of generating questions and answers with high accuracy, and also enables automated creation of patent documents. Specifically, the patent document drafting device of the present invention is composed of the following components: an acquisition unit; a question generation unit that generates a first question; an answer acquisition unit; an answer generation unit that automatically generates an answer to the first question even if an answer to the first question cannot be obtained; and a creation unit that automatically creates patent documents using a large-scale language model.

[0238] 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0239] The present invention relates to a patent document drafting device that supports the drafting of patent documents using a large-scale language model. According to the present invention, the device includes: an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. In an embodiment of the present invention, the use of the large-scale language model enables highly accurate question answering, and the inclusion of an automatic answering function enables efficient document drafting.

[0240] The annotated question search infrastructure according to the present invention comprises an annotated question search processor, an annotated question storage unit, a question summary query processor, and a dialogue processor. The annotated question search processor analyzes the syntax of an input question and extracts important phrases from the input question that indicate the question content. The annotated question storage unit stores annotated question summaries. The question summary query processor searches for annotated question summaries. The dialogue processor presents annotated question summaries in response to a dialogue with a user. In the step [Question Summary Determination Process by Annotated Question Search], the annotated question search processing unit of the annotated question search platform analyzes the syntax of the input question, extracts important phrases from the input question, and generates an annotated question summary using the important phrases as keys. The generated annotated question summary is presented in response to a dialogue with the user. The annotated information presentation device according to the present invention comprises an information presentation device main body and an information acquisition device, and the information acquisition device determines the information content to be displayed by the information presentation device main body based on information input by a user. The information presentation device main body determines the information content to be displayed based on the input information, and when the user operates the displayed content, additional information corresponding to the operation content is displayed. The information presentation device main body also estimates user attributes, preferences, etc. from the user's operation history, operation status, etc., and changes the displayed content, additional information, etc. based on the estimation results. In the "Presenting Annotated Information" step, the information content to be displayed is determined based on the information entered by the user. In addition, when the user performs an operation, additional information is displayed according to the operation (see figure), and the display content and additional information are also changed based on the results inferred from the user's attributes and preferences.

[0241] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires a first user answer from a user, an answer generation unit that generates a first automatic answer when a first user answer is not acquired, and a creation unit that creates a patent document using the large-scale language model. It is also possible to configure the device so that a user can select a predicate as an answer from a list of primary answers, and elements related to the selected predicate (such as a subject or modifier) ​​are extracted and presented to the user.

[0242] The patent document drafting device according to the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it comprises an acquisition unit for acquiring patent information, a question generation unit for generating a first question, an answer acquisition unit for acquiring an answer, an answer generation unit for generating an answer when an answer is not acquired, and a drafting unit for creating a patent document. This device may include a button or buttons for the user to select an answer, and the user may execute evaluation input by selecting one of the buttons.

[0243] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question based on the patent information using a large-scale language model. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question using the large-scale language model. The creation unit creates the patent document using the large-scale language model. In addition, if the answer acquisition unit is unable to acquire the first user answer, the first automatic answer can be included as reference information.

[0244] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. A first question is created based on the patent information using a large-scale language model, and an answer is acquired from a user. If no answer is available, an answer to the first question is generated using the large-scale language model. A patent document is created using the large-scale language model based on the first user answer or the first automatic answer.

[0245] A patent document preparation device in which a prompt for causing a large-scale language model to execute processing on a patent document displayed in a patent document display field is a user-selectable field. 1. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit; and a prompt selection field. The prompt selection field allows selection from multiple candidates suggested by the large-scale language model, and the selected candidate is the most appropriate one among the candidates suggested by the large-scale language model.

[0246] An interactive system comprising a display processing unit that displays a plurality of prompts in a prompt selection field, each of the prompts being an identifiable object. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and the patent document is created using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit, and is capable of being displayed by a display processing unit.

[0247] In the prompt selection field, for example, "raw", "stage", "number", and "time" may be prompts to be executed. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0248] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit that generates a first question using a large-scale language model, an answer acquisition unit that acquires an answer to the first question, an answer generation unit that generates an answer to the first question when an answer to the first question cannot be acquired, and a creation unit that creates a patent document based on the patent information and the first user answer or the first automatic answer. The operation screen includes at least a patent information display field, a question display field, and an answer input field.

[0249] The present invention relates to a patent document drafting device that supports the creation of patent documents using a large-scale language model. According to the present invention, patent information related to patents described in patent documents is acquired, a question is generated using a large-scale language model, and an answer is acquired from a user. If an answer cannot be acquired, an answer is generated using the large-scale language model. Then, a patent document is created using the large-scale language model based on the acquired information and answer. In this way, the creation of patent documents can be supported. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a patent document drafting device that uses the large-scale language model to draft the patent document based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. For example, if a question is asked, "What technical field does this invention belong to?", the answer will be generated in the form of, "This invention belongs to the XX technical field." Also, if a question is asked, "What is the XX technical field?", the answer will be generated in the form of, "The XX technical field is △△." The present invention relates to an apparatus and method for supporting the creation of patent documents using a large-scale language model, and more specifically to an advanced automatic response system and an operating method for the automatic response system that utilizes the latest technologies such as large-scale language models and artificial intelligence technology. In the present invention, for example, if a question is asked such as "What technical field does this invention belong to?", an answer is generated in the form of "This invention belongs to the XX technical field."

[0250] The patent document creation device acquires information about the patent to be described in the patent document and generates a question using a large-scale language model. If an answer cannot be obtained from the user, the device generates an answer using the large-scale language model. The patent document is then created based on the acquired information and the answer. At this time, the patent document display field on the operation screen displays the portion corresponding to the item selected in the item selection field. a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit; a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit; and an operation screen that displays a template display field that shows the content of a template selected in a template selection field from a group of templates that includes at least a portion of the patent document created by the creation unit. The operation screen displays a template display field that displays the content of the template selected in the template selection field. This enables users to easily, quickly, accurately, efficiently, effectively, comfortably, safely, and reliably acquire, learn, master, improve, and develop the desired information, knowledge, techniques, skills, experience, insights, understanding, judgments, decisions, etc.

[0251] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. When a user inputs patent information, the question generation unit generates a first question using a large-scale language model. If there is an answer from the user, the answer acquisition unit acquires the answer, and if there is not, the answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document based on the first user answer or the first automatic answer. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. When the "Create patent document with this content" button is selected on the operation screen, a confirmation dialog box will appear asking "Is this content OK?" If OK is selected, the document will be sent to the server.

[0252] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit acquires patent information related to patents. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer is not acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The operation screen includes a patent document display field.

[0253] The patent document creation device includes an acquisition unit that acquires patent information related to the patent to be described in the patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired from the user, and a creation unit that creates a patent document using a large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The operation screen includes an item selection field that allows each item constituting the patent document to be selected.

[0254] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. A user inputs patent information into a patent information input field. The question generation unit generates a first question using a large-scale language model. The answer acquisition unit receives a first user answer. The answer generation unit generates a first automatic answer using the large-scale language model. The creation unit creates a patent document using the large-scale language model. a first user answer to the first question obtained from the user, or a first automatic answer to the first question generated by the answer generation unit; and an operation screen including a patent information input field for receiving patent information from a user, and at least one operation element selected from the group consisting of a "Next" button, a "Back" button, a "Done" button, a "Cancel" button, a "Save" button, a "Print" button, a "Send by Email" button, a "PDF Output" button, a "Word Output" button, and an "Excel Output" button.

[0255] The patent document creation device includes an acquisition unit that acquires patent information, an answer acquisition unit that acquires answers to questions generated using a large-scale language model, an answer generation unit that generates an answer using a large-scale language model if no answer is available, a creation unit that creates a patent document, and a display processing unit. The display processing unit can display an operation screen. The operation screen includes a patent information input field, a patent document display field, and a prompt selection field. A user can input patent information in the patent information input field and select one of multiple prompts from the prompt selection field. Processing is performed using the large-scale language model on the selected prompt, and the results are displayed in the patent document display field. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit, and a display processing unit. The present invention aims to solve the technical problem in the "large-scale language model" technical field that "highly accurate sentence generation has been difficult with conventional technology" and to provide a "patent document drafting device" and a "sentence generation method" that are capable of "highly accurate sentence generation." The present invention is a technical means in the "large-scale language model" technical field, and by utilizing technical means such as a "large-scale language model" and "artificial intelligence," it is possible to provide a "sentence generation method" and a "patent document drafting device" that are capable of "highly accurate sentence generation."

[0256] The patent document creation device is composed of an acquisition unit that acquires information about patents to be described in patent documents, an answer acquisition unit that acquires answers to questions generated using a large-scale language model, an answer generation unit that generates answers using the large-scale language model when there is no answer to the question generated using the large-scale language model, and a creation unit that creates patent documents using the large-scale language model.The display processing unit displays an input screen for information necessary to create patent documents on the display unit, and displays patent documents created by the server, in accordance with information sent from the server.

[0257] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer to the first question. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. The display processing unit displays an operation screen. A sentence suitable for an example patent specification, related to the content of claim 1: In an embodiment of the present invention, a patent document drafting device is provided. The device includes an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model, and the answer acquisition unit acquires a first user answer from a user. If there is no answer from the user, the answer generation unit automatically generates a first automatic answer. The creation unit then creates a patent document using the large-scale language model, and the display processing unit displays the patent document on an operation screen to provide it to the user.

[0258] When acquiring patent information, the display processing unit displays an operation screen including an input field into which the patent information can be input on the user terminal. The question generation unit generates a first question requesting a detailed explanation of the patent based on the patent information. The answer acquisition unit acquires an answer to the first question from the user. If an answer to the first question is not acquired from the user, the answer generation unit generates an answer to the first question using a large-scale language model. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or first automatic answer.

[0259] An information processing device comprising: a display processing unit that displays a plurality of prompts as identifiable objects in a prompt selection field. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit.

[0260] The display processing unit controls the display of data on the display unit. In order to determine the data to be displayed on the display unit, the display processing unit determines the data to be displayed based on the operation content operated by the user. The display processing unit also determines the position and size of the data to be displayed on the display unit based on the operation content operated by the user. Furthermore, the display processing unit determines the color and shape of the data to be displayed on the display unit based on the operation content operated by the user. In this way, data can be displayed in an appropriate form depending on the operation content operated by the user.

[0261] The patent document creation device includes an acquisition unit that acquires information about the patent to be described in the patent document, a question generation unit that generates a question using a large-scale language model, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer using the large-scale language model if there is no answer, and a creation unit that creates the patent document using the large-scale language model. Furthermore, an object capable of identifying a prompt may be displayed so that it can be dragged (moved) in response to a user's selection, and the selection receiving unit can receive a user's action of superimposing (dragging) the object on the patent document display field as a selection of the prompt associated with the specified object.

[0262] The selection receiving unit receives from the user a selection of at least one of the multiple prompts displayed in the prompt selection field. For example, if the user selects "patent document creation," the question generating unit generates a question related to patent document creation. Also, if the user selects "patent search," the question generating unit generates a question related to patent search. In this way, appropriate questions can be generated according to the information the user is seeking.

[0263] The present invention relates to a patent document drafting device including a dialogue processing platform selection receiving unit, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The dialogue processing platform selection receiving unit receives designation of a question and answer system to be used from a user. The question generation unit generates a first question using a large-scale language model based on patent information. The answer acquisition unit acquires a first user answer from the user. The answer generation unit generates a first automatic answer using the large-scale language model if the first user answer cannot be acquired. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The present invention is a "patent document creation device" composed of a combination of elemental technologies such as a "large-scale language model," "patent document," "question," and "answer." This device is composed of five functional elements: an "acquisition unit," a "question generation unit," a "answer acquisition unit," a "answer generation unit," and a "creation unit." In this device, the "acquisition unit" first collects "patent information," the "question generation unit" uses the "large-scale language model" to automatically generate a "first question" such as "points requiring detailed explanation," and the "answer acquisition unit" collects a "first answer" (= input) such as "input from the user (= first user answer)."

[0264] The patent document drafting device according to the present invention supports the drafting of patent documents using a large-scale language model. It acquires information about the patents to be described in patent documents and generates questions based on that information. If an answer cannot be obtained, it generates an answer using the large-scale language model. Then, it creates a patent document based on the acquired information and the answer. In this way, it is possible to improve the efficiency of patent document drafting and support the drafting of high-quality patent documents.

[0265] The patent document creation device includes an acquisition unit that acquires information about the patent to be described in the patent document, a question generation unit that generates a question using a large-scale language model, an answer acquisition unit that acquires an answer, an answer generation unit that generates an answer using the large-scale language model if there is no answer, and a creation unit that creates the patent document using the large-scale language model. Furthermore, an object capable of identifying a prompt may be displayed so that it can be dragged (moved) in response to a user's selection, and the selection receiving unit can receive a user's action of superimposing (dragging) the object on the patent document display field as a selection of the prompt associated with the specified object.

[0266] The patent document creation device includes an acquisition unit that acquires patent information related to a patent to be described in a patent document, a question generation unit that generates a first question requesting a detailed explanation of the patent, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question if an answer to the first question is not acquired from the user, and a creation unit that creates a patent document using a large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. The creation unit combines the acquired {patent information} and {answer to the first question} (first user answer) with a prompt that uses {patent information} and {answer to the first question} in a prompt that is stored in advance in a prompt table, and inputs the combined information into the large-scale language model.

[0267] The patent document creation device is composed of an acquisition unit for acquiring patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The acquisition unit for acquiring patent information collects information on patents to be described in patent documents. The question generation unit generates a first question based on the patent information. The answer acquisition unit receives a first user answer from a user. The answer generation unit generates a first automatic answer if there is no first user answer. The creation unit creates a patent document based on the patent information and the first user answer or the first automatic answer. A feature of the present invention is that question generation and answer generation are performed using a large-scale language model.

[0268] A patent document preparation device in which a prompt for causing a large-scale language model to execute processing on a patent document displayed in a patent document display field is a user-selectable field. 1. A patent document creation device that supports the creation of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit; and a prompt selection field. The prompt selection field allows selection from multiple candidates suggested by the large-scale language model, and the selected candidate is the most appropriate one among the candidates suggested by the large-scale language model.

[0269] A method for identifying an identifiable object, comprising: a display unit that displays a prompt; an object display unit that displays an identifiable object; and a character display unit that displays characters that indicate the content of the prompt by superimposing them on the identifiable object.

[0270] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, it acquires patent information, generates questions using the large-scale language model, and acquires answers from the user. If no answers are available, it generates answers using the large-scale language model. Then, it creates a patent document based on the patent information and answers. In addition, the patent document drafting device of the present invention can learn the user's revision tendencies and set new prompts based on the content and frequency of revisions made by the revision unit.

[0271] An interactive system comprising a display processing unit that displays a plurality of prompts in a prompt selection field, each of the prompts being an identifiable object. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. For example, the system of the present invention can be applied even when technologies such as voice recognition technology and natural language processing technology advance. An interactive system that can be applied even as technologies such as voice recognition and natural language processing advance.

[0272] In the prompt selection field, for example, "raw", "stage", "number", and "time" may be prompts to be executed. 1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition unit that acquires patent information related to a patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question acquired from the user or a first automatic answer to the first question generated by the answer generation unit. In one embodiment of the present invention, prompt selection fields such as "raw", "step", "number", "time", etc. are provided. By selecting and inputting appropriate ones from these prompt selection fields, more accurate and rapid input becomes possible. Also, even when words such as "raw" and "step" are related in some way, how they are related is automatically determined, so the accuracy is likely to be improved.

[0273] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. Based on the patent information, a first question is generated using a large language model, and an answer is acquired from the user. If there is no answer, an answer to the first question is generated using the large language model. Based on the first user answer or the first automatic answer, a patent document is created using the large language model.

[0274] The patent document creation device is composed of an acquisition unit that acquires information regarding the patent described in the patent document, a question generation unit that generates questions using a large language model, an answer acquisition unit that acquires answers, an answer generation unit that generates answers using the large language model when there is no answer, and a creation unit that creates a patent document using the large language model. Also, an object capable of identifying a prompt may be displayed so as to be draggable (movable) according to the user's selection, and the selection reception unit can receive the user's superimposing operation (drag operation) of the object on the patent document display field as the selection of a prompt associated with a predetermined object.

[0275] A method for identifying a distinguishable object, comprising: a display unit that displays a prompt; an object display unit that displays a distinguishable object; and a character display unit that displays characters indicating the content of the prompt, superimposed on the distinguishable object.

[0276] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer. In addition, the display position of the user object corresponds to the location of the city, town, or village where the address of the searched answer candidate is located.

[0277] The patent document creation device includes an acquisition unit that acquires patent information, a question generation unit, an answer acquisition unit, an answer generation unit, and a creation unit. The question generation unit generates a first question using a large-scale language model based on the patent information. The answer acquisition unit acquires a first user answer from a user. The answer generation unit generates a first automatic answer using the large-scale language model when an answer to the first question is not acquired from the user. The creation unit creates a patent document using the large-scale language model based on the patent information and the first user answer or the first automatic answer.

[0278] The present invention relates to a method for creating a security maturity report including a security implementation progress report. According to the present invention, in order to create a security maturity report having the characteristics of a security implementation progress report, the method comprises the steps of creating a security implementation progress report, collecting information necessary for the security implementation progress report, analyzing the collected information and extracting information necessary for the security implementation progress report, organizing the extracted information and expressing it in a format necessary for the security implementation progress report, and creating the security implementation progress report based on the organized information. A method for creating a security maturity report, the method comprising the steps of: providing a security maturity report; and providing a security maturity report as a progress report of security implementation. 2. The method according to claim 1, wherein the collected information is collected from a system configuration diagram, a system design diagram, or the like. 3. The method according to claim 1, wherein the extracted information is extracted from a system configuration diagram, a system design diagram, or the like. 4. The method according to claim 1, wherein the organized information is expressed in the form of a table, a graph, or the like. 5. The method according to claim 1, wherein the system configuration diagram, system design diagram, etc. are created on a computer.

[0279] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document, a question generation unit that generates a first question requesting a detailed description of the patent using the large-scale language model based on the patent information, an answer acquisition unit that acquires an answer to the first question from a user, an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user, and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. This configuration eliminates the need for manual document drafting as in the past, enabling efficient and accurate document drafting.

[0280] The present invention relates to a patent document drafting device. The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model if an answer to the first question is not acquired from the user; and a creation unit that creates the patent document using the large-scale language model based on the patent information, a first user answer to the first question acquired from the user, or a first automatic answer to the first question generated by the answer generation unit. The patent document drafting device of the present invention is capable of creating highly accurate patent documents by using the large-scale language model.

[0281] The patent document drafting device of the present invention supports the drafting of patent documents using a large-scale language model. Specifically, the device includes an acquisition unit that acquires patent information related to a patent described in a patent document; a question generation unit that generates a first ques...

Claims

1. A patent document drafting device that supports the drafting of patent documents using a large-scale language model, an acquisition unit that acquires patent information related to the patent described in the patent document; a question generation unit that generates a first question requesting a detailed explanation of the patent based on the patent information using the large-scale language model; an answer acquisition unit that acquires an answer to the first question from a user; an answer generation unit that generates an answer to the first question using the large-scale language model when an answer to the first question is not obtained from the user; a creation unit that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user or a first automatic answer to the first question generated by the answer generation unit; A patent document creation device comprising:

2. the question generation unit generates a second question requesting a detailed explanation of the patent based on the first user answer or the first automatic answer, using the large-scale language model; the answer acquisition unit acquires an answer to the second question from the user; the answer generation unit generates an answer to the second question using the large-scale language model when an answer to the second question is not acquired from the user; and the creation unit creates the patent document using the large-scale language model based on the patent information, the first user answer or the first automatic answer, and a second user answer to the second question obtained from the user or a second automatic answer to the second question generated by the answer generation unit. The patent document drafting device according to claim 1.

3. a correction unit that corrects the patent document created by the creation unit using the large-scale language model; The patent document drafting device according to claim 1.

4. The creation unit creates the patent document for each of a plurality of items constituting the patent document. The patent document drafting device according to claim 1.

5. When using the large-scale language model, a predetermined prompt is input to the large-scale language model, a setting unit configured to accept at least one of a change by the user to the predetermined prompt to be input to the large-scale language model and a setting by the user of a new prompt; The patent document drafting device according to claim 1.

6. a display processing unit that displays, on a terminal device used by the user, an operation screen including an input field into which the patent information can be input, a display field that displays the patent document created by the creation unit, and a selection field in which a plurality of prompts for processing the patent document displayed in the display field using the large-scale language model can be selected; a selection receiving unit that receives, from the user, a selection of at least one of the plurality of prompts displayed in the selection field; Furthermore, the creation unit processes the patent document displayed in the display field using a large-scale language model according to the selected prompt in response to a selection operation of the prompt received by the selection receiving unit; The patent document drafting device according to claim 1.

7. the display processing unit displays the plurality of prompts as identifiable objects in the selection field; the selection receiving unit receives, by the user, an action of superimposing the object on the display field as a selection of the prompt associated with the predetermined object; 6. The patent document drafting apparatus according to claim 5.

8. A method for controlling a patent document drafting device that supports the drafting of patent documents using a large-scale language model, comprising: an acquisition step of acquiring patent information relating to the patents described in the patent documents; a query generation step of generating a first query requesting a detailed description of the patent based on the patent information using the large-scale language model; an answer acquisition step of acquiring an answer to the first question from a user; an answer generation step of generating an answer to the first question using the large-scale language model if an answer to the first question is not obtained from the user; a creation step of creating the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user or a first automatic answer to the first question generated in the answer generation step; A method for controlling a patent document preparation device.

9. A control program for a patent document drafting device that supports the drafting of patent documents using a large-scale language model, On the computer, an acquisition function for acquiring patent information relating to the patents described in the patent document; a question generation function that generates a first question requesting a detailed explanation of the patent using the large-scale language model based on the patent information; an answer acquisition function for acquiring an answer to the first question from the user; an answer generation function that generates an answer to the first question using the large-scale language model when an answer to the first question is not obtained from the user; a creation function that creates the patent document using the large-scale language model based on the patent information and a first user answer to the first question obtained from the user or a first automatic answer to the first question generated by the answer generation function; A control program for a patent document creation device that realizes the above.

Images