Information processing device, information processing method, and program

Abstract

The present invention provides an information processing device that makes the creation of detailed specifications easy and hassle-free. [Solution] The information processing device comprises: a receiving unit that receives as input the components of the invention to be prepared, prior art documents relating to the invention, and matters described in the specification; a specifying unit that performs a specifying process to identify components of prior art documents similar to the components of the invention to be prepared; an extraction unit that extracts descriptions relating to the components identified in the specifying process from the prior art documents; an instruction unit that instructs a language model to describe matters based on the extracted descriptions; and an output unit that receives and outputs the output from the language model in response to the instructions.

Description

Technical Field

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] It is known to generate documents using large language models (LLMs). Patent Document 1 discloses a technique that assists in generating specifications using an LLM.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problem to be Solved by the Invention

[0004] In related specification generation systems, when an LLM is made to write a specification, the content of the specification may be insufficiently detailed. This is because the LLM has not undergone specialized training specific to a particular domain, and when it is required to write text that handles knowledge in a specific domain such as patents, the content may lack sufficient expertise and become insufficiently detailed. To solve this issue, it is conceivable to make the LLM learn the specialized expertise, but in that case, retraining may be required. That is, in specification creation using an LLM, extra labor may be required to create a detailed specification.

[0005] An object of the present disclosure is to provide an information processing apparatus that can solve the above-described problems.

Means for Solving the Problem

[0006] This disclosure provides an information processing device comprising: a receiving means for receiving components of an invention to be prepared, prior art documents relating to the invention, and matters described in the specification as input; a identifying means for performing an identifying process to identify components of the prior art documents that are similar to the components of the invention to be prepared; an extraction means for extracting descriptions relating to the components identified in the identifying process from the prior art documents; an instruction means for giving instructions to a language model to describe the matters described based on the extracted descriptions; and an output means for receiving and outputting an output from the language model in response to the instructions.

[0007] Furthermore, this disclosure provides an information processing method that takes as input elements of an invention to be written, prior art documents relating to the invention, and matters described in the specification, performs a specification process to identify elements of the prior art documents that are similar to the elements of the invention to be written, extracts descriptions relating to the elements identified in the specification process from the prior art documents, gives instructions to a language model to describe the matters based on the extracted descriptions, and receives and outputs the output from the language model in response to the instructions.

[0008] Furthermore, this disclosure provides a program that causes a computer to perform the following processes: a process of receiving as input the components of an invention to be written, prior art documents relating to the invention, and the matters described in the specification; a process of performing a specific process to identify components of the prior art documents that are similar to the components of the invention to be written; a process of extracting from the prior art documents the descriptions relating to the components identified in the specific process; a process of giving instructions to a language model to describe the matters based on the extracted descriptions; and a process of receiving and outputting the output from the language model in response to the instructions. [Effects of the Invention]

[0009] According to this disclosure, the preparation of a detailed specification can be made easily and without effort. [Brief explanation of the drawing]

[0010] [Figure 1] This is a block diagram showing the hardware configuration of the information processing device relating to this disclosure. [Figure 2] This block diagram shows the functional configuration of the information processing device related to this disclosure. [Figure 3] This is a flowchart showing the operation of the information processing device related to this disclosure. [Figure 4] This is a schematic block diagram showing the functional configuration of the information processing device related to this disclosure. [Figure 5] This flowchart shows an example of the processing procedure performed by the information processing device related to this disclosure. [Figure 6] This block diagram shows the functional configuration of the information processing device related to this disclosure. [Figure 7] This is a flowchart showing the operation of the information processing device related to this disclosure. [Figure 8] This block diagram shows the functional configuration of the information processing device related to this disclosure. [Figure 9] This is a flowchart showing the processing procedure of the information processing device related to this disclosure. [Figure 10] This is a schematic block diagram showing the functional configuration of the information processing device related to this disclosure. [Figure 11] This is a flowchart showing the operation of the information processing device related to this disclosure. [Modes for carrying out the invention]

[0011] First, the hardware configuration of the information processing device will be described with reference to Figure 1. Figure 1 is a block diagram showing the hardware configuration of the information processing device according to an embodiment of this disclosure.

[0012] Figure 1 shows an example of the hardware configuration of an information processing device. The information processing device is implemented by computer 10.

[0013] The computer 10 includes a processor 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a storage device 14 such as a hard disk that stores programs, an input / output interface 15 for inputting and outputting data, and a communication interface 16 for network connection. The respective components are connected via a bus 17.

[0014] The processor 11 operates an operating system to control the entire computer. Examples of the processor 11 include a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a GPU (Graphics Processing Unit). The processor 11 loads programs stored in, for example, the ROM 12, the storage device 14, or the like. Then, the processor 11 executes each process coded in the program. The processor 11 may execute processes or instructions in the illustrated flowchart based on the program.

[0015] The ROM 12 stores application programs, programs according to each embodiment, and the like. Further, the RAM 13 is used as a work area of the processor 11.

[0016] Examples of the storage device 14 include a semiconductor memory such as a flash memory, an HDD (Hard Disk Drive), and the like. The storage device 14 stores, for example, an OS (Operating System) program, application programs, programs according to each embodiment, and the like.

[0017] The input / output interface 15 is connected to a peripheral device (not shown). The connection method may be a wired network or a wireless network.

[0018] The communication interface 16 is connected to a communication network (not shown), such as a LAN (Local Area Network) or WAN (Wide Area Network), via a wireless or wired network. The communication network may consist of multiple communication networks. This allows the computer to connect to an external large-scale language model server (LLM) via the communication network. The computer 10 may have components other than those shown in Figure 1. For example, the computer 10 may have a drive device. For example, the processor 11 may be mounted on a drive device or the like and read programs and data stored on a non-temporary tangible recording medium into the RAM 13.

[0019] The information processing device may be configured to include some of the components described in Figure 1. For example, the information processing device may be configured to include only the processor 11, ROM 12, and RAM 13 from the components described above. In this case, the storage device 14, input / output interface 15, and communication interface 16 may be provided as external devices to the information processing device. Furthermore, some of the arithmetic functions of the information processing device may be implemented by an external server or cloud service.

[0020] First, we will describe the first embodiment.

[0021] Figure 2 is a block diagram showing the functional configuration of an information processing device. As shown in Figure 2, the information processing device 100 includes a receiving unit 110, a specific unit 120, an extraction unit 130, an instruction unit 140, and an output unit 150. The information processing device 100 is also connected to a language model 101. In the example shown in Figure 1, the language model 101 is located outside the information processing device 100 and is connected to the information processing device 100 via a communication network. However, the information processing device 100 may also have the language model 101 built-in, and the specific configuration of the connection is not limited.

[0022] The reception unit 110 is a reception means that receives as input the components of the invention to be prepared, prior art information, and description items. For example, if a user intends to prepare a patent application specification for an invention called "image processing device," the reception unit 110 will receive components such as "image processing unit," "display control unit," and "operation reception unit," information on related prior art documents, and description items such as "problems the invention aims to solve" and "means for solving the problems." The reception unit 110 may also receive instructions for revisions to the specification output by the output unit 150.

[0023] The reception unit 110 receives the above-mentioned information through the user interface. The reception unit 110 may also use a voice recognition system. In this case, the user can input the components of the invention, prior art information, and descriptions by voice. The reception unit 110 may also use a handwriting recognition system. In this case, the user can input information by handwriting on a touchscreen or tablet. The reception unit 110 may also accept input via email or an API (Application Programming Interface).

[0024] The identification unit 120 is an identification means that performs an identification process to identify components of prior art that are similar to the components of the invention received by the reception unit 110. The identification unit 120 identifies components of prior art that are semantically similar to the input components, for example, by using methods such as vector search or semantic search. For example, for the component "image processing unit", it identifies similar components such as "image processing unit" and "video processing module" from the prior art. In vector search, for example, the components are vectorized and the similarity is determined using cosine similarity. Alternatively, a general keyword search may be used in the identification unit 120.

[0025] The identification unit 120 may perform the identification process using a machine learning model different from vector search or semantic search. In this case, the machine learning model is trained on a large amount of patent documents and identifies components of prior art that are similar to the input components.

[0026] Prior art refers to, for example, patent documents. However, prior art is not limited to patent documents and may include various other technical documents.

[0027] The extraction unit 130 is an extraction means that extracts descriptions of the components identified by the identification unit 120 from prior art. The extraction unit 130 performs the extraction process using a language model 101. The language model 101 is, for example, a Large Language Model (LLM). The LLM understands the given context and extracts relevant information. For example, the LLM extracts descriptions of an "image processing unit," such as its function, operation, and relationship with other components. The type of language model 101 is not particularly limited. Multiple different language models may be used as the language model 101. The extraction unit 130 may also use a rule-based natural language processing system as the language model 101. In this case, relevant descriptions can be extracted from prior art based on predefined rules. The extraction unit 130 may use multiple methods in combination to extract the descriptions.

[0028] The instruction unit 140 is an instruction means that instructs the language model 101 to write the required information based on the information extracted by the extraction unit 130. For example, the instruction unit 140 may instruct the language model 101 to "describe the operation of the image processing unit in approximately 500 characters. In doing so, be sure to mention the 'image data compression process' and 'noise reduction function' extracted from prior literature."

[0029] Instructions from the instruction unit 140 are represented as prompts (instructional statements) to be input to the language model 101. Instructions from the instruction unit 140 may consist of one prompt or multiple prompts. Furthermore, instructions from the instruction unit 140 are not limited to prompts. For example, tokenized information equivalent to the prompts described above may be input to the language model 101 as an instruction. Instructions from the instruction unit 140 may be not only text-based instructions but also structured data formats (e.g., JSON, XML).

[0030] The output unit 150 is an output means that receives and outputs output from the language model 101. The output text is a document that is a patent application specification or part thereof, written according to the instructions of the instruction unit 140. For example, the text "Problems to be Solved by the Invention" output from the output unit 150 may be displayed on a screen or printed by a printer. The output method from the output unit 150 is not limited to this. Furthermore, the user can make correction or additional instructions as needed based on the output content.

[0031] Next, the operation of the information processing device 100 will be explained with reference to Figure 3. Figure 3 is a flowchart showing the operation (information processing method) of the information processing device 100.

[0032] First, the reception unit 110 receives as input the constituent elements of the invention to be prepared, the prior art documents relating to the invention, and the matters to be described in the specification (step S101).

[0033] Next, the identification unit 120 identifies components of prior art that are similar to the components of the received invention (step S102).

[0034] Next, the extraction unit 130 extracts descriptions from the prior art that relate to the identified components of the prior art (step S103).

[0035] Subsequently, the instruction unit 140 instructs the language model 101 to record the information based on the extracted information (step S104).

[0036] Finally, the output unit 150 receives the output of the language model 101 and outputs it (step S105).

[0037] According to the information processing device 100 of this embodiment, since the description is identified based on past specifications, new training for the language model is unnecessary, and it can contribute to the creation of detailed specifications without effort. For example, the identification unit 120 identifies components of prior art similar to the specified components, the extraction unit 130 extracts the description, and the instruction unit 140 gives instructions to the language model 101, thereby enabling the creation of a detailed specification for a specific domain without retraining. This eliminates the tendency for general-purpose LLMs to lack expertise, resulting in superficial descriptions, and makes it easy to create detailed specifications that effectively utilize knowledge from past specifications without effort. In other words, the information processing device 100 of this embodiment facilitates improved quality and reduced time for specification creation by efficiently utilizing knowledge of prior art. Although this embodiment describes the creation of a patent specification as an example, the scope of application of the described method is not limited to the creation of patent specifications.

[0038] Alternatively, the instruction unit 140 may use a template-based text generation system instead of giving instructions to the language model 101. In this case, the extracted descriptions can be applied to a pre-prepared template to generate the descriptions.

[0039] The second embodiment will be described below.

[0040] Figure 4 is a schematic block diagram showing the functional configuration of the information processing device. In this figure, the information processing device 200 comprises a reception unit 210, a specification unit 220, an extraction unit 230, an instruction unit 240, an output unit 250, and a setting unit 260. The reception unit 210, specification unit 220, extraction unit 230, instruction unit 240, and output unit 250 have the same basic functions as the reception unit 110, specification unit 120, extraction unit 130, instruction unit 140, and output unit 150 of the first embodiment. Also, similar to the example of the information processing device 100, the information processing device 200 is connected to the language model 201.

[0041] The reception unit 210 accepts as input the components of the invention to be prepared, prior art information, and description items. For example, if a user inputs "image recognition device" as a component, specifies "Japanese Patent Publication No. 2020-1xxxxx" as the relevant prior art, and selects "problems to be solved by the invention" as the description items, the reception unit 210 accepts this information.

[0042] The identification unit 220 performs an identification process to identify components of prior art that are similar to the components of the invention received by the reception unit 210. For example, in response to the input "image recognition device," the identification unit 220 uses vector search or semantic search to identify similar components in the prior art, such as "image processing unit" or "object recognition unit."

[0043] The extraction unit 230 extracts descriptions from prior art that relate to the components identified by the identification unit 220. For example, the extraction unit 230 extracts a description relating to the "image processing unit" such as, "It takes image data captured by a camera as input and performs preprocessing such as edge detection and noise reduction."

[0044] The instruction unit 240 instructs the language model 201 to describe the information based on the description extracted by the extraction unit 230. The language model 201 is, for example, an LLM. For example, the instruction unit 240 instructs the language model 201 to generate a description of "the problem that the invention aims to solve," taking into account the extracted description and adding new features of the invention.

[0045] The output unit 250 receives and outputs the output from the language model 201 in response to the instructions from the instruction unit 240. For example, the text "Problems the invention aims to solve" output from the output unit 250 may be displayed on a screen or printed by a printer.

[0046] The setting unit 260 sets the weight of prior art based on the type of prior art. The weight of the prior art represents the degree of importance and relevance of the prior art that can be identified when extracting the description. The type includes, for example, the applicant, inventor, technical field, and filing date. However, the examples of types are not limited to these, and any matter that may be related to the weight may be used as a type.

[0047] The setting unit 260 may, for example, assign a high weight (e.g., 0.8) to the company's own patent applications, a medium weight (e.g., 0.5) to other companies' patent applications, and a low weight (e.g., 0.3) to non-patent literature. The setting unit 260 may also set the weights according to, for example, the degree of agreement between the invention and the technical field. Furthermore, the setting unit 260 may also set the weights according to, for example, the degree of agreement between the inventor and the invention for which the specification is being prepared. The weights are expressed as numerical values ​​between, for example, 0 and 1, but are not limited to this and may be determined as appropriate. If appropriate information about the type of prior art cannot be obtained, the setting unit 260 may set predetermined weights.

[0048] The instruction unit 240 modifies the instructions for describing the extracted descriptions based on the weights set by the setting unit 260. For example, if a high weight is set for a prior art document that is the company's own patent, the instruction unit 240 adds an instruction to the language model 201 such as "Prioritize referring to the description of the company's patent and use similar expressions." Also, if there is only one identified prior art document, the instruction unit 240 may modify the instructions for describing the document according to the weight set for that identified prior art document. For example, the instruction unit 240 may modify the instructions to refer to the document if the weight is greater than a predetermined criterion.

[0049] The output unit 250 may output the response to the instructions from the language model 201 and the weights set for prior art.

[0050] With this configuration, the information processing device 200 can more appropriately adjust the content of the generated specification according to the type and importance of the prior art. For example, it becomes possible to efficiently create specifications that emphasize the company's own prior art or that use expressions specific to a particular technical field.

[0051] Next, with reference to Figure 5, the operation of the information processing device 200 in the second embodiment will be described. Figure 5 is a flowchart showing an example of the processing procedure performed by the information processing device 200.

[0052] First, the reception unit 210 receives the components of the invention, prior art information, and description from the user (step S201). For example, the reception unit 210 receives input from the user of a component called "image recognition device," a prior art document called "Japanese Patent Publication No. 2020-1xxxxx," and a description called "problems to be solved by the invention."

[0053] Next, the identification unit 220 identifies components of prior art that are similar to the components of the received invention (step S202). For example, the identification unit 220 uses vector search or semantic search to identify components such as an "image processing unit" or "object recognition unit" that are similar to an "image recognition device".

[0054] Next, the setting unit 260 sets weights based on the type of prior art (step S203). For example, if "Japanese Patent Publication No. 2020-123456" is the company's patent application, the setting unit 260 sets the weight to 0.8. This weighting quantifies the importance and relevance of the prior art and is used in subsequent processing. The weights set are not limited to this.

[0055] Subsequently, the extraction unit 230 extracts descriptions relating to the components of the identified prior art (step S204). For example, as a description relating to the "image processing unit," it extracts a description such as, "It takes image data captured by a camera as input and performs preprocessing such as edge detection and noise reduction."

[0056] Next, the instruction unit 240 modifies the instructions to the language model 201 based on the set weights (step S205). For example, if a high weight is set for prior art that is the company's own patent, the instruction unit 240 adds an instruction to the language model 201 such as, "Prioritize referring to the description in our own patent and use similar expressions."

[0057] Then, the instruction unit 240 instructs the language model 201 to write the required information based on the modified instructions (step S206). For example, regarding "the problem the invention aims to solve," it instructs the language model 201 to generate a description that incorporates the new features of the invention while referring to the extracted description.

[0058] When the language model 201 generates a description based on the instructions, the output unit 250 receives the output from the language model 201 and outputs it (step S207). For example, the generated "Problem to be Solved by the Invention" text may be displayed on a screen or printed by a printer.

[0059] Thus, the information processing device 200 of the second embodiment can efficiently generate more appropriate specification descriptions by weighting prior art according to its importance. According to the information processing device 200, for example, it is possible to efficiently create specifications that emphasize the company's own prior art or that use expressions specific to a particular technical field.

[0060] According to the information processing device 200 of this embodiment, a more appropriate specification can be efficiently generated by setting weights based on the type of prior art and adjusting the instructions to the language model 201 according to those weights. For example, by setting high weights for the company's own patents and highly relevant prior art, it becomes possible to create a specification that places more emphasis on the descriptions in those documents. This makes it possible to generate a higher quality specification that not only refers to the descriptions in prior art but also takes into account the importance and relevance of the prior art.

[0061] The third embodiment will be described below.

[0062] Figure 6 is a block diagram showing the functional configuration of the information processing device. As shown in Figure 6, the information processing device 300 of this embodiment includes a reception unit 310, a specific unit 320, an extraction unit 330, an instruction unit 340, and an output unit 350. The reception unit 310, specific unit 320, extraction unit 330, instruction unit 340, and output unit 350 have the same basic functions as the reception unit 110, specific unit 120, extraction unit 130, instruction unit 140, and output unit 150 of the first embodiment. Also, similar to the example of the information processing device 100, the information processing device 300 is connected to a language model 301.

[0063] The reception unit 310 accepts the components of the invention to be prepared, prior art information, and description as input. For example, if a user inputs "facial recognition camera" as a component, specifies "high-precision facial recognition system" and "personal authentication method by image processing" as related prior art, and selects "mode for carrying out the invention" as the description, the reception unit 310 accepts this information.

[0064] The identification unit 320 performs an identification process to identify components of prior art that are similar to the components of the invention received by the receiving unit 310. In this embodiment, the identification unit 320 performs the identification process multiple times with different parameters. In this embodiment, the parameters are parameters related to the search method used by the identification unit 320 for identification, for example, parameters related to the algorithm selected by the identification unit 320. In a more detailed example, the parameters may be values ​​that indicate the weights of the two search methods when the identification unit 320 uses semantic search and keyword search to identify components of prior art. For example, the identification unit 320 performs three searches with different parameters and identifies relevant components from all the obtained prior art. In this way, the identification unit 320 aims to identify more relevant components of prior art by performing the identification process multiple times with different parameters.

[0065] Furthermore, increasing the number of specific processes can lengthen processing time, potentially increasing user waiting times. This is especially true when there is a large amount of prior art or when the invention has complex components.

[0066] In this case, the identification unit 320 may set an upper limit on the processing time, and if that limit is reached, it may identify relevant components from the prior art obtained at that time. For example, if a time limit of 30 seconds is set for each specific process, and the total of three specific processes exceeds 90 seconds, the identification unit 320 may identify relevant components from the prior art obtained at that time.

[0067] The extraction unit 330 extracts descriptions relating to the components of the prior art identified by the identification unit 320. In this embodiment, if there are multiple prior art documents, the extraction unit 330 extracts descriptions common to those documents. Furthermore, the extraction unit 330 excludes descriptions relating to specific data or numerical values ​​from the extraction target.

[0068] Descriptions common to multiple prior art documents are assumed to be general or highly important technical content in the technical field to which the subject of the specification belongs. Therefore, the extraction unit 330 extracts descriptions common to multiple prior art documents. Furthermore, specific data or numerical values ​​may be unsuitable for explaining the invention subject to the specification based on descriptions where the underlying conditions differ from document to document. Therefore, the extraction unit 330 excludes descriptions related to specific data or numerical values ​​from the extraction process.

[0069] For example, with respect to the invention of a "facial recognition camera" for which a specification is to be prepared, the extraction unit 330 extracts the description "extracts facial feature points and compares them with registered data," which is common to both the prior art document concerning a "high-precision facial recognition system" and the prior art document concerning a "personal authentication method by image processing." On the other hand, with respect to a similar invention for which a specification is to be prepared, the extraction unit 330 excludes descriptions related to specific numerical values ​​in the prior art document, such as "the camera resolution is 12 million pixels or more."

[0070] Furthermore, when the extraction unit 330 extracts descriptions common to multiple prior art documents, if only one highly relevant prior art document is found, or if there are very few common descriptions, the extraction unit 330 may not be able to extract sufficient descriptions.

[0071] In this case, if the amount of common information is below a certain level, the extraction unit 330 may extract information deemed important from individual prior art documents. For example, if the amount of common information is less than 100 characters, the extraction unit 330 may extract additional information from the prior art document with the highest similarity, resulting in a total of approximately 300 characters of information. The number of characters extracted is not limited to this and can be determined appropriately depending on the required amount of information and the similarity of the prior art documents.

[0072] Furthermore, by excluding descriptions of specific data or numerical values ​​from the extraction unit 330, important information may be lost. For example, qualitative expressions such as "high precision" or "high speed" may not adequately describe the features of the invention.

[0073] Therefore, when excluding specific numerical values ​​from the extraction target, the extraction unit 330 may extract expressions that indicate the trend or relative magnitude of those numerical values. For example, instead of excluding the statement "The camera resolution is 12 million pixels or higher," the extraction unit 330 may extract the expression "Use a high-resolution camera."

[0074] The instruction unit 340 instructs the language model 301 to describe the information based on the description extracted by the extraction unit 330. For example, the instruction unit 340 might give an instruction such as, "Regarding 'facial recognition cameras,' please generate a general description of 'modes for carrying out the invention' that does not include specific numerical values ​​or data, based on descriptions that have been commonly extracted from multiple prior art documents."

[0075] The output unit 350 receives and outputs the output in response to instructions from the language model 301. For example, it displays the generated description of "Modes for Carrying Out the Invention" on the display.

[0076] Next, the operation of the information processing device 300 according to this embodiment will be described. Figure 7 is a flowchart showing the operation of the information processing device 300 according to this embodiment.

[0077] First, the reception unit 310 receives the constituent elements of the invention to be prepared, prior art information, and description (step S301).

[0078] Next, the identification unit 320 performs identification processing multiple times with different parameters to identify components of similar prior art (step S302).

[0079] Next, the extraction unit 330 extracts descriptions from the prior art that relate to the identified components of the prior art (step S303). In step S303, if there are multiple prior art documents, the extraction unit 330 may extract common descriptions. Also, in step S303, the extraction unit 330 may exclude descriptions relating to specific data or numerical values ​​from the extraction.

[0080] Subsequently, the instruction unit 340 instructs the language model 301 to record the information based on the extracted information (step S304).

[0081] The output unit 350 receives the output generated by the language model 301 based on the instructions and outputs it (step S305).

[0082] As described above, the information processing device 300 of this embodiment can generate a more general and reusable specification description by extracting common descriptions from multiple prior art documents and excluding specific data and numerical values. This eliminates the tendency for general-purpose LLMs to lack specialization, resulting in a lack of detail, while avoiding overly specific descriptions, and enables the creation of a specification that covers a wide range of embodiments.

[0083] Furthermore, by performing the identification process multiple times with different parameters, it is possible to identify more relevant prior art elements and improve the accuracy of the extraction. In addition, by extracting only common descriptions, it is possible to focus on important information that is consistent across multiple prior art documents, thereby generating more reliable descriptions.

[0084] These measures enable the information processing device 300 of this embodiment to generate specification descriptions more efficiently and effectively. By limiting processing time, it is possible to collect as much relevant information as possible while appropriately managing the user's waiting time. Furthermore, by implementing measures for cases where there is little common description, it is possible to always ensure a certain amount of useful description. In addition, by replacing numerical data with expressions that show trends, it becomes possible to appropriately express the features of the invention while avoiding specific numerical values.

[0085] The fourth embodiment will be described below.

[0086] Figure 8 is a block diagram showing the functional configuration of the information processing device. As shown in Figure 8, the information processing device 400 includes a receiving unit 410, a specific unit 420, an extraction unit 430, an instruction unit 440, and an output unit 450. Also, similar to the example of the information processing device 100, the information processing device 400 is connected to a language model 401.

[0087] The reception unit 410 accepts as input the components of the invention to be prepared, prior art information, and description items. For example, if a user wants to prepare a specification for an invention called "image processing device," the reception unit 410 accepts components such as "image processing unit," "display control unit," and "operation reception unit," information on relevant prior art documents, and description items such as "problems the invention aims to solve" and "means for solving the problems."

[0088] The identification unit 420 performs an identification process to identify components of prior art that are similar to the components of the invention received by the receiving unit 410. The identification unit 420 uses methods such as vector search and semantic search to identify components of prior art that are semantically similar to the input components. For example, for the component "image processing unit," it identifies similar components such as "image processing unit" and "video processing module" from the prior art.

[0089] The extraction unit 430 extracts descriptions of the components identified by the identification unit 420 from prior art. For example, a language model 401 may be used for this extraction process.

[0090] The instruction unit 440 instructs the language model 401 to record information based on the information extracted by the extraction unit 430.

[0091] The instruction unit 440 gives instructions to determine whether adding the information based on the information extracted by the extraction unit 430 would cause a contradiction with the content of the invention for which the specification is to be prepared, and to modify the information based on the information extracted

[0092] For example, the instruction unit 440 may give instructions to the language model 401 such as, "If the extracted description of the 'image processing unit' is added to the description of the 'image processing unit' of the present invention, determine whether this contradicts the objective of the invention, which is 'realization of real-time processing'. If there is a contradiction, revise the description to resolve the contradiction."

[0093] Furthermore, the instruction unit 440 gives instructions to determine whether the component of the invention to be specified has a function that the corresponding component of the prior art has that is not present in the component of the invention to be specified; if it is determined that the component has this function, it gives instructions to determine whether adding this function would create a contradiction in the invention to be specified; and if it is determined that no contradiction would occur, it gives instructions to add this function. In other words, for example, the instruction unit 440 determines whether an additional function of a component of the prior art can be added to the invention, and instructs the language model 401 to add it if there is no contradiction.

[0094] In this case, the instruction unit 440 may, for example, give instructions to the language model 401 such as, "If the 'noise reduction function' of the 'image processing unit' in the prior art is added to the 'image processing unit' of the present invention, please determine whether there are any problems with consistency with the purpose of the invention and other components. If there are no problems, please generate a description of the 'image processing unit' with this function added."

[0095] The output unit 450 receives and outputs the output corresponding to the instruction from the language model 401.

[0096] Next, the operation of the information processing device 400 will be explained with reference to Figure 9. Figure 9 is a flowchart showing an example of the operation of the information processing device 400.

[0097] First, the reception unit 410 receives as input the constituent elements of the invention to be prepared, the prior art documents relating to the invention, and the matters to be described in the specification (step S401).

[0098] Next, the identification unit 420 identifies components of prior art that are similar to the components of the received invention (step S402).

[0099] Next, the extraction unit 430 extracts descriptions from the prior art that relate to the identified components of the prior art (step S403).

[0100] Subsequently, the instruction unit 440 instructs the language model 401 to describe the information based on the extracted description (step S404). In step S404, the instruction unit 440 gives instructions to determine whether describing the information based on the description extracted by the extraction unit 430 would cause a contradiction with the content of the invention for which the specification is to be prepared, and to modify the description of the information based on the extracted description based on the determination result. The instruction unit 440 also gives instructions to determine whether the components of the prior art corresponding to the components of the invention for which the specification is to be prepared have a function that is not present in the components of the invention for which the specification is to be prepared, and if it is determined that this function exists, it gives instructions to determine whether adding this function would cause a contradiction with the invention for which the specification is to be prepared, and if it is determined that no contradiction would occur, it gives instructions to add this function.

[0101] Finally, the output unit 450 receives the output from the language model 401 and outputs it (step S405).

[0102] With this configuration, the information processing device 400 of this embodiment can effectively utilize the descriptions in prior art while supporting the detailed description of the specification in a manner that does not impair the essence of the invention. In particular, the information processing device 400 can create a higher quality specification by determining inconsistencies and evaluating additional functions.

[0103] The inconsistency detection function and additional function evaluation function, which are features of this embodiment, eliminate the need for users to individually check the consistency between the description in prior art and the content of the invention, as well as the appropriateness of each additional function, significantly reducing the burden of drafting the specification. At the same time, since the LLM automatically makes these judgments, it is possible to prevent minor inconsistencies that users tend to overlook and to avoid missing useful additional functions, resulting in the efficient creation of higher-quality specifications.

[0104] The fifth embodiment will be described below.

[0105] Figure 10 is a schematic block diagram showing the functional configuration of the information processing device. In Figure 10, the information processing device 500 comprises a receiving unit 510, a specific unit 520, an extraction unit 530, an instruction unit 540, and an output unit 550. The receiving unit 510, specific unit 520, extraction unit 530, instruction unit 540, and output unit 550 have the same basic functions as the receiving unit 110, specific unit 120, extraction unit 130, instruction unit 140, and output unit 150 of the first embodiment. Also, similar to the example of the information processing device 100, the information processing device 500 is connected to the language model 501.

[0106] The information processing device 500 uses the language model 501 to create a specification and repeatedly modifies and improves that specification.

[0107] The reception unit 510 receives as input the components of the invention to be prepared, prior art information, and description of the invention. For example, if a user wants to prepare a specification for an invention called "image processing device," the reception unit 510 receives components such as "image processing unit" and "display control unit," relevant prior art documents, and description of the invention such as "problems to be solved by the invention" and "effects of the invention." The reception unit 510 may also receive instructions for revisions to the specification output by the output unit 550.

[0108] The identification unit 520 performs an identification process to identify components of prior art that are similar to the components of the invention received by the receiving unit 510. The identification unit 520 uses techniques such as vector search or semantic search to identify components of prior art that are semantically similar to the input components. For example, for the component "image processing unit," it identifies "image processing unit" and "video processing module" in the prior art as similar components. As shown in the example of the identification unit 120 above, the specific identification method of the identification unit 520 is not limited to this.

[0109] The extraction unit 530 extracts descriptions from prior art that relate to the components identified by the identification unit 520. For example, the extraction unit 530 extracts a description relating to the "image processing unit" such as, "It takes image data captured by a camera as input and performs preprocessing such as edge detection and noise reduction."

[0110] The instruction unit 540 instructs the language model 501 to write down the information based on the description extracted by the extraction unit 530. For example, it might instruct, "Please write a description of the operation of the image processing unit in about 200 characters. When doing so, be sure to mention the 'image data compression process' and the 'noise reduction function' that were extracted from prior literature."

[0111] Furthermore, the instruction unit 540 may repeatedly instruct the creation and modification of the specification based on the content of the specification generated by the language model 501. For example, the instruction unit 540 performs the following processing:

[0112] The instruction unit 540 instructs the language model 501 to create an initial specification based on the output received from the language model 501. Based on the created specification, the identification unit 520 again identifies prior art elements similar to those in the created specification. The extraction unit 530 extracts descriptions relating to the newly identified prior art elements. Based on the extracted new descriptions, the instruction unit 540 instructs the language model 501 to revise the descriptions in the previously created specification.

[0113] For example, if the description of the "image processing unit" is insufficient in the initial draft of the specification, the instruction unit 540 will give instructions to the language model 501 such as, "Please provide a more detailed description of the image processing unit. In particular, please mention the 'real-time processing function' newly extracted from prior art and explain the advantages of the present invention."

[0114] The output unit 550 receives and outputs the output from the language model 501 in response to the instructions. For example, the output unit 550 outputs the contents of the specification generated by the language model 501 and details of the modified parts.

[0115] With this configuration, the information processing device 500 can repeatedly modify the content after the initial specification has been drafted. This makes it possible to efficiently create more detailed and comprehensive specifications. As a result, a reduction in the time required to draft the specification and an improvement in the quality of the resulting specification can be expected.

[0116] Next, we will explain the operation of the information processing device 500. Figure 11 is a flowchart showing an example of the operation of the information processing device 500.

[0117] The reception unit 510 receives the components of the invention, prior art information, and description from the user (step S501).

[0118] The identification unit 520 identifies components of prior art that are similar to the components of the received invention (step S502).

[0119] The extraction unit 530 extracts descriptions relating to the components of the identified prior art (step S503).

[0120] The instruction unit 540 instructs the language model 501 to record the information based on the information extracted in step S503 (step S504).

[0121] The output unit 550 receives the output from the language model 501 and outputs it (step S505).

[0122] The instruction unit 540 instructs the language model 501 to prepare the specification (step S506).

[0123] The specific section 520 identifies components of prior art that are similar to the components described in the prepared specification (step S507).

[0124] The extraction unit 530 extracts the description relating to the newly identified components (step S508).

[0125] The instruction unit 540 instructs the language model 501 to revise the specification based on the newly extracted description (step S509).

[0126] The output unit 550 receives and outputs the revised specification, which is the output of the language model 501 (step S510).

[0127] The user reviews the output specification and determines whether further revisions are necessary. The need for revisions is determined based, for example, on the completeness of the specification and the importance of newly extracted information. The reception unit 510 receives a revision instruction for the output specification (step S511). If revisions are necessary (step S511: Yes), the process returns to, for example, step S507 and is repeated. If no revisions are necessary (step S511: No), the process ends. Note that if no revisions are necessary, this includes, for example, when the reception unit 510 receives input indicating that no revisions are necessary, or when a predetermined time has elapsed without the reception unit 510 receiving a revision instruction.

[0128] As described above, the information processing device 500 of this embodiment can improve the specification created using the language model 501, making it possible to efficiently create higher-quality specifications. The information processing device 500 can solve the problem of insufficient detail due to the lack of expertise in general-purpose LLMs. In other words, by supplementing important information overlooked during the initial specification creation and providing more appropriate expressions in subsequent processing, it becomes possible to create higher-quality specifications.

[0129] According to this embodiment, the specification created using the language model 501 can be iteratively improved. This makes it possible to repeatedly improve the content even after the initial specification is created, and to efficiently create a higher-quality specification. For example, the quality of the specification can be improved by repeatedly performing the following processes: the identification unit 520 identifies similar components of prior art again based on the components of the created specification; the extraction unit 530 extracts descriptions related to the newly identified components; and the instruction unit 540 instructs the language model 501 to revise the specification based on the new descriptions.

[0130] Furthermore, this iterative process can improve the superficiality often resulting from a lack of general-purpose expertise in LLMs. By supplementing important information overlooked during the initial drafting and providing more appropriate wording in subsequent processing, it becomes possible to create higher-quality specifications. For example, it becomes easier to add or modify technical details not mentioned in the initial drafting, or to use more appropriate terminology, through iterative processing.

[0131] Furthermore, the information processing device 500 makes it easier to reduce the time required to prepare the specification and to improve the quality of the prepared specification. By using the language model 501, it becomes possible to extract relevant information from prior art and prepare and improve the specification based on that information, thereby reducing the time required to prepare the specification. In addition, by using the language model 501, it becomes possible to ensure consistency in the writing style and use of technical terms throughout the specification, making it easier to achieve uniform quality.

[0132] Thus, the information processing device 500 of this embodiment can achieve efficient specification creation by utilizing the language model 501 and modifying the specification through iterative processing.

[0133] In this embodiment, the information processing device 500 uses the language model 501 to create a specification and makes iterative revisions. This allows for the extraction of appropriate similar components and useful descriptions even when the prior art information provided by the user is insufficient or when there is little prior art on a particular technical field.

[0134] The instruction unit 540 may also generate a message prompting the user to provide additional prior art information. For example, the message might say, "Relevant prior art is insufficient. Please provide more prior art information." The output unit 550 outputs this message.

[0135] Furthermore, if no similar components are found for a particular component, the instruction unit 540 may instruct the language model 501 to query for general technical information about that component and to supplement the description in the specification based on that information. For example, the instruction unit 540 may instruct the language model 501 to "No similar components were found for the image processing unit. Please generate a general description of the image processing technology in about 200 characters and incorporate it into the specification."

[0136] Furthermore, the identification unit 520 may gradually broaden the search range for similar components. For example, the identification unit 520 may initially seek a more precise similarity to identify components of prior art, and if sufficient results are not obtained, it may lower the similarity threshold to identify components of prior art. This allows the identification unit 520 to broadly identify potentially related components even if no directly similar components are found.

[0137] In this way, the information processing device 500 of this embodiment makes it easy to create a specification that maintains a certain level of quality even when prior art information is insufficient. Furthermore, by encouraging the user to provide additional information, it can support the collection of information necessary for creating a more comprehensive specification.

[0138] This disclosure may also be implemented by recording a program for realizing the above-described functions on a computer-readable recording medium, loading the program recorded on this recording medium into a computer system, and executing it. The term "computer system" as used herein includes hardware such as an operating system and peripheral devices.

[0139] Furthermore, "computer-readable recording media" refers to portable media such as flexible disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computer systems. In addition, "computer-readable recording media" may also include those that dynamically hold programs for a short period of time, such as communication lines used when transmitting programs over networks such as the Internet or communication lines such as telephone lines, and those that hold programs for a certain period of time, such as volatile memory inside computer systems that act as servers or clients in such cases.

[0140] Furthermore, the above program may be intended to implement some of the functions described above. It may also be a so-called differential file (differential program) that can implement the aforementioned functions in combination with programs already recorded in the computer system.

[0141] Furthermore, the above program may be for the purpose of realizing some of the functions described above, or it may be for the purpose of realizing the above functions in combination with a program already recorded in the computer system, or it may be for the purpose of realizing using a programmable logic device such as an FPGA (Field Programmable Gate Array).

[0142] While embodiments of this disclosure have been described in detail above with reference to the drawings, the specific configurations are not limited to these embodiments and include designs and other elements that do not depart from the gist of this disclosure. Furthermore, the contents described in each embodiment can be used in combination with each other.

[0143] Some or all of the above embodiments may also be described as follows, but are not limited to the following:

[0144] (Note 1) A receiving means that accepts as input the components of the invention to be prepared, the prior art documents relating to the invention, and the matters described in the specification, A specifying means that performs a specifying process to identify components of the prior art that are similar to the components of the invention that is the subject of the specification preparation, An extraction means for extracting descriptions relating to the components identified in the specified process from the aforementioned prior art, An instruction means for giving instructions to a language model to describe the information based on the extracted description, Output means that receives and outputs the output corresponding to the instruction from the language model, An information processing device equipped with the following features. (Note 2) The system includes a setting means for setting the weight of prior art based on the type of prior art, The indicating means modifies the indication of the description of the description item based on the extracted description, based on the weight. The information processing device described in Appendix 1. (Note 3) The aforementioned identification means performs the identification process multiple times with different parameters. The information processing device described in Appendix 1 or 2. (Note 4) The extraction means, when the receiving means receives multiple prior art documents, extracts descriptions common to the multiple prior art documents. The information processing device described in Appendix 1 or 2. (Note 5) The extraction means excludes descriptions of specific data or numerical values ​​from the extraction target. The information processing device described in Appendix 1 or 2. (Note 6) The aforementioned instruction means is An instruction to determine whether the description of the matters based on the extracted description will result in a contradiction with the content of the invention for which the specification is to be prepared, Based on the results of the above determination, an instruction is given to change the description of the information based on the extracted description, The information processing device described in Appendix 1 or 2. (Note 7) The aforementioned instruction means is An instruction to determine whether the component of the invention to be specified has a function that the corresponding component of the prior art does not have, If it is determined that the invention has the aforementioned function, an instruction is given to determine whether adding the aforementioned function would result in a contradiction with the invention for which the specification is being prepared, If it is determined that the aforementioned contradiction does not occur, the instruction to add the aforementioned function will be given. The information processing device described in Appendix 6. (Note 8) The instruction means instructs the creation of a specification based on the received output, The aforementioned identifying means identifies components of the prior art that are similar to the components described in the prepared specification, The extraction means extracts descriptions relating to the newly identified components of the prior art, The instruction means instructs the language model to modify the description of the specification based on the newly extracted description. The information processing device described in Appendix 1. (Note 9) The system accepts the constituent elements of the invention to be specified, the prior art documents relating to the invention, and the information to be included in the specification as input. A process is performed to identify the components of the prior art that are similar to the components of the invention for which the specification is to be prepared. From the aforementioned prior art, we extract descriptions relating to the components identified in the aforementioned specific processing, The language model is instructed to record the information based on the extracted information. The language model receives and outputs the output corresponding to the instruction. Information processing methods. (Note 10) On the computer, A process that accepts as input the components of the invention to be prepared, the prior art documents relating to the invention, and the information described in the specification, A process of performing a specific operation to identify components of the prior art that are similar to the components of the invention that are the subject of the specification preparation, A process to extract descriptions of the components identified in the specified process from the aforementioned prior art, A process to instruct the language model to record the information based on the extracted information, A process that receives and outputs the output corresponding to the instruction from the language model, A program that executes the command.

[0145] Furthermore, some or all of the configurations described in Appendices 2 to 8, which are subordinate to Appendice 1 above, may also be subordinate to Appendices 9 and 10 in the same way as those described in Appendices 2 to 8. Moreover, not limited to Appendices 1, 9, and 10, some or all of the configurations described as appendices may also be subordinate to various hardware, software, various recording means for recording software, or systems, without departing from the embodiments described above. [Explanation of Symbols]

[0146] 100 Information Processing Devices 110 Reception Department 120 Specific section 130 Extraction part 140 Instruction section 150 Output section 160 Setting section

Claims

1. A receiving means that accepts as input the components of the invention to be prepared, the prior art documents relating to the invention, and the matters described in the specification, A specifying means that performs a specifying process to identify components of the prior art that are similar to the components of the invention that is the subject of the specification preparation, An extraction means for extracting descriptions relating to the components identified in the specified process from the aforementioned prior art, An instruction means for giving instructions to a language model to describe the information based on the extracted description, Output means that receives and outputs the output corresponding to the instruction from the language model, An information processing device equipped with the following features.

2. The system includes a setting means for setting the weight of prior art based on the type of prior art, The indicating means modifies the indication of the description of the description item based on the extracted description, based on the weight. The information processing apparatus according to claim 1.

3. The aforementioned identification means performs the identification process multiple times with different parameters. The information processing apparatus according to claim 1 or 2.

4. The extraction means, when the receiving means receives multiple prior art documents, extracts descriptions common to the multiple prior art documents. The information processing apparatus according to claim 1 or 2.

5. The extraction means excludes descriptions of specific data or numerical values ​​from the extraction target. The information processing apparatus according to claim 1 or 2.

6. The aforementioned instruction means is An instruction to determine whether the description of the matters based on the extracted description will result in a contradiction with the content of the invention for which the specification is to be prepared, Based on the results of the above determination, an instruction is given to change the description of the information based on the extracted description, The information processing apparatus according to claim 1 or 2.

7. The aforementioned instruction means is An instruction to determine whether the component of the invention to be specified has a function that the corresponding component of the prior art does not have, If it is determined that the invention has the aforementioned function, an instruction is given to determine whether adding the aforementioned function would result in a contradiction with the invention for which the specification is being prepared, If it is determined that the aforementioned contradiction does not occur, the instruction to add the aforementioned function will be given. The information processing apparatus according to claim 6.

8. The instruction means instructs the creation of a specification based on the received output, The aforementioned identifying means identifies components of the prior art that are similar to the components described in the prepared specification, The extraction means extracts descriptions relating to the newly identified components of the prior art, The instruction means instructs the language model to modify the description of the specification based on the newly extracted description. The information processing apparatus according to claim 1.

9. The system accepts the constituent elements of the invention to be specified, the prior art documents relating to the invention, and the information to be included in the specification as input. A process is performed to identify the components of the prior art that are similar to the components of the invention for which the specification is to be prepared. From the aforementioned prior art, we extract descriptions relating to the components identified in the aforementioned specific processing, The language model is instructed to record the information based on the extracted information. The language model receives and outputs the output corresponding to the instruction. Information processing methods.

10. On the computer, A process that accepts as input the components of the invention to be prepared, the prior art documents relating to the invention, and the information described in the specification, A process of performing a specific operation to identify components of the prior art that are similar to the components of the invention that are the subject of the specification preparation, A process to extract descriptions of the components identified in the specified process from the aforementioned prior art, A process to instruct the language model to record the information based on the extracted information, A process that receives and outputs the output corresponding to the instruction from the language model, A program that executes the command.

Images