Analysis program, information processing device and analysis method

Abstract

To provide a technique for providing a novel value by using a large-scale language model for analysis of patent information.SOLUTION: An analysis program of patent information causes a computer to execute the steps of: acquiring technical information including technical ideas; acquiring patent information to be compared with the technical information; designating technical information and patent information with respect to a predetermined instruction sentence for evaluating relevance between the two pieces of information from a specific view point; and inputting the instruction sentence in which the technical information and the patent information are designated to the large-scale language model, to output an evaluation result representing the relevance between the technical information and the patent information in terms of a viewpoint based on a result obtained from the large-scale language model.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present disclosure relates to an analysis program, an information processing device, and an analysis method. [Background technology]

[0002] In recent years, various large-scale language models have been developed. A large-scale language model is a language model trained on a huge amount of text data, and is trained to be able to process various natural languages.

[0003] Non-Patent Document 1 discloses a method for improving the output accuracy of ChatGPT, which is an example of a large-scale language model. Specifically, Non-Patent Document 1 discloses that the accuracy of the output is improved by inputting the instruction "Let's think step by step" into ChatGPT. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] "The spell to make ChatGPT smarter", Nikkei Newspaper, [online], March 24, 2023, [Retrieved September 6, 2023], Internet<URL:https: / / www.nikkei.com / article / DGXZQOUC22BVO0S3A320C2000000 / > Summary of the Invention [Problem to be solved by the invention]

[0005] Large-scale language models can be applied to a variety of services, but there are still few examples of them being realized as concrete services. In this regard, there is a demand for using large-scale language models in patent information analysis to provide new value. [Means for solving the problem]

[0006] In one example of the present disclosure, a patent information analysis program is provided, which causes a computer to execute the steps of acquiring technical information including a technical idea, acquiring patent information to be compared with the technical information, designating the technical information and the patent information in a predetermined instruction sentence for evaluating the relevance of two pieces of information from a specific viewpoint, and outputting an evaluation result indicating the relevance between the technical information and the patent information from the viewpoint based on a result obtained from a large-scale language model by inputting the instruction sentence in which the technical information and the patent information are designated into the large-scale language model.

[0007] In one example of the present disclosure, the predetermined instruction sentence is stipulated to evaluate the relevance of the two pieces of information from a plurality of viewpoints, and in the outputting step, an evaluation result indicating the relevance between the technical information and the patent information is output for each of the plurality of viewpoints.

[0008] In one example of the present disclosure, the multiple perspectives include at least one of a technical field perspective, a problem perspective, an action function perspective, a content suggestion perspective, a parameter perspective, a numerical limitation perspective, and a technical term perspective.

[0009] In an example of the present disclosure, the analysis program further causes the computer to execute a step of outputting a comprehensive evaluation result that summarizes the other evaluation results of the viewpoints.

[0010] In one example of the present disclosure, each of the plurality of viewpoints is associated with a weight, and the comprehensive evaluation result is calculated by applying the weight to the other evaluation results of the viewpoints.

[0011] In one example of the present disclosure, the predetermined instruction sentence includes an instruction for causing the large-scale language model to output the relevance between the two pieces of information, and an instruction for causing the large-scale language model to output an evaluation reason for the relevance. The evaluation result includes the relevance between the technical information and the patent information, and the evaluation reason for the technical information and the patent information.

[0012] In one example of the present disclosure, the analysis program further causes the computer to execute a step of receiving an input regarding a concept level when comparing the patent information with the technical information. The step of specifying further includes reflecting the input regarding the concept level in the instruction sentence.

[0013] In one example of the present disclosure, the analysis program further causes the computer to execute the steps of: accepting input regarding a conceptual level when comparing the patent information with the technical information while the evaluation result is being displayed on a display; generating an instruction for re-evaluation that reflects the input regarding the conceptual level; and inputting the instruction for re-evaluation into the large-scale language model to output the results obtained from the large-scale language model.

[0014] In one example of the present disclosure, the analysis program further causes the computer to execute a step of calculating a similarity between the technical information and the patent information based on a predetermined algorithm. The evaluation result is output for each of the viewpoints based on the result obtained from the large-scale language model and the similarity.

[0015] In one example of the present disclosure, the multiple perspectives include a first perspective and a second perspective different from the first perspective. The large-scale language model includes a first language model and a second language model different from the first language model. The instruction sentence for evaluating the relevance between the technical information and the patent information from the first perspective is input to the first language model. The instruction sentence for evaluating the relevance between the technical information and the patent information from the second perspective is input to the second language model.

[0016] In one example of the present disclosure, the technical information includes a drawing, the patent information includes a drawing, and the viewpoint includes a viewpoint of the contents of the drawing.

[0017] In one example of the present disclosure, the instruction text includes a first instruction for evaluating the relevance of the two pieces of information from a specific perspective, and a second instruction for further evaluating the relevance of the two pieces of information from the specific perspective using a result of evaluating the relevance of the two pieces of information based on the first instruction.

[0018] In another example of the present disclosure, an information processing device capable of analyzing patent information is provided. The information processing device includes a control unit for operating the information processing device. The control unit executes a process of acquiring technical information including a technical idea, a process of acquiring patent information to be compared with the technical information, a process of designating the technical information and the patent information in response to a predetermined instruction sentence for evaluating the relevance of two pieces of information from a viewpoint, and a process of outputting an evaluation result indicating the relevance between the technical information and the patent information from the viewpoint based on a result obtained from a large-scale language model by inputting the instruction sentence in which the technical information and the patent information are designated into the large-scale language model.

[0019] In another example of the present disclosure, a method for analyzing patent information executed by a computer is provided, which includes the steps of: acquiring technical information including a technical idea; acquiring patent information to be compared with the technical information; specifying the technical information and the patent information in a predetermined instruction sentence for evaluating the relevance of two pieces of information from a plurality of viewpoints; and inputting the instruction sentence in which the technical information and the patent information are specified into a large-scale language model, based on the result obtained from the large-scale language model, outputting an evaluation result indicating the relevance between the technical information and the patent information for the viewpoints.

[0020] The above and other objects, features, aspects and advantages of the present invention will become apparent from the following detailed description of the invention taken in conjunction with the accompanying drawings. [Brief description of the drawings]

[0021] [Figure 1] FIG. 1 illustrates an example of a device configuration of an information processing system. [Diagram 2] FIG. 13 is a diagram for explaining a relevance evaluation function. [Diagram 3] FIG. 13 is a diagram for explaining a relevance assessment function according to a modified example. [Figure 4] FIG. 2 is a schematic diagram illustrating an example of a hardware configuration of an information processing device. [Diagram 5] FIG. 2 is a schematic diagram illustrating an example of a hardware configuration of a user terminal. [Figure 6] FIG. 2 is a diagram illustrating an example of a data flow between an information processing device, a user terminal, and a server. [Figure 7] FIG. 13 is a diagram showing an example of a search condition input screen. [Figure 8] FIG. 13 is a diagram showing an example of an input screen for review conditions. [Figure 9] FIG. 13 is a diagram showing an example of an instruction text. [Figure 10] FIG. 13 is a diagram showing answer information that is an example of an answer generated by a large-scale language model. [Figure 11] FIG. 13 is a diagram showing an example of an evaluation result screen. [Figure 12] FIG. 13 is a flowchart showing a process for evaluating the relevance between technical information and patent information. [Figure 13] FIG. 13 is a diagram showing an example of weights associated with different viewpoints. [Figure 14] FIG. 13 is a diagram illustrating a schematic of a relevance assessment process according to another example. [Figure 15] FIG. 13 is a diagram illustrating a schematic of a relevance assessment process according to another example. [Figure 16] FIG. 13 is a diagram showing another example of an instruction sentence. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0022] Hereinafter, each embodiment according to the present invention will be described with reference to the drawings. In the following description, the same parts and components are given the same reference numerals. Their names and functions are also the same. Therefore, detailed description thereof will not be repeated. Note that each embodiment and each modified example described below may be appropriately and selectively combined.

[0023] <A. Information Processing System 10> First, with reference to FIG. 1, the device configuration of the information processing system 10 will be described. FIG. 1 is a diagram showing an example of the device configuration of the information processing system 10.

[0024] As shown in FIG. 1, the information processing system 10 includes an information processing device 100, a user terminal 200, and a server 300. The information processing device 100, the user terminal 200, and the server 300 are configured to be communicable with each other through a network NW (for example, the Internet).

[0025] The information processing device 100 is a notebook or desktop PC (Personal Computer), a tablet terminal, a smartphone, or other computer equipped with a communication function. The number of information processing devices 100 constituting the information processing system 10 may be one or two or more. The information processing device 100 is, for example, operated by Company "A".

[0026] The user terminal 200 is, for example, a notebook or desktop PC, a tablet terminal, a smartphone, or other computer equipped with a communication function. The number of user terminals 200 constituting the information processing system 10 may be one or two or more. The information processing device 100 is, for example, owned by User "A", who is an ordinary user.

[0027] The server 300 is a notebook or desktop PC (Personal Computer), a tablet terminal, a smartphone, or other computer equipped with a communication function. The number of servers 300 constituting the information processing system 10 may be one or two or more. The server 300 is, for example, operated by Company "B".

[0028] The server 300 stores a large-scale language model 324. The large-scale language model 324 is a language model trained on a huge amount of text data of more than several billions, and is trained to be able to process various natural languages. The large-scale language model 324 is also called LLM (Large Language Models). The large-scale language model 324 is trained to generate an output according to an instruction sentence when the instruction sentence is input.

[0029] Examples of the large-scale language model 324 include the GPT series, such as GPT-3 (Generative Pretrained Transformer) and GPT-4, PaLM (Scaling Language Modeling with Pathways), LLaMA (Large Language Model Meta AI), and known LLM. In addition to the GPT series, various large-scale language models may be used, such as a Transformer-based large-scale language model, such as BERT (Bidirectional Encoder Representations from Transformers), a CNN (Convolutional Neural Network), an RNN (Recurrent Neural Network), and an LSTM (Long Short Term Memory).

[0030] For example, company "B" publishes an API (Application Programming Interface) for utilizing the functions of the large-scale language model 324. This allows designers and general users of company "A" to utilize the functions of the large-scale language model 324 through the API.

[0031] The various processes described in this specification may be implemented in the information processing device 100, the user terminal 200, the server 300, or other computers.

[0032] In the above description, an example where the information processing system 10 includes the server 300 has been described. However, the information processing system 10 may not include the server 300. In this case, the information processing system 10 is composed of one or more information processing devices 100 and one or more user terminals 200.

[0033] <B. Processing Overview> The information processing device 100 provides various analysis functions related to patent information to the user "A". As an example, the information processing device 100 provides a function for evaluating the relevance between technical information and patent information to the user "A".

[0034] "Patent information" refers to data related to patent documents. Patent information includes, for example, patent documents such as published patent gazettes, patent gazettes, published patent applications, re-published patents, and utility model gazettes. As an example, patent information includes bibliographic data, specifications, claims, drawings, and abstracts. Examples of bibliographic data include, for example, application numbers, publication numbers, patent registration numbers, filing dates, publication dates, registration dates, applicants, patent holders, invention titles, agents, and filing countries. Note that patent information may be non-patent documents such as published technical reports.

[0035] "Technical information" refers to data containing technical ideas. A technical idea is a technical means for solving a technical problem. Technical information is, for example, something in which technical means are described as character strings. Technical information includes, for example, invention information in which invention-specific matters are described.

[0036] Hereinafter, with reference to FIG. 2, an overview of the function for evaluating the relevance between technical information 123 and patent information 125 will be described. FIG. 2 is a diagram for explaining the relevance evaluation function.

[0037] The information processing device 100 acquires technical information 123 from the user terminal 200. The technical information 123 may be a character string input by the user "A" of the user terminal 200, or may be data extracted from patent information in response to the operation of the user "A".

[0038] Furthermore, the information processing device 100 acquires patent information 125 to be compared with the technical information 123. The source from which the patent information 125 is acquired is arbitrary. As one example, the patent information 125 is acquired from a patent database 124 (see FIG. 4) described below that is stored in a storage device in the information processing device 100. As another example, the patent information 125 is acquired from a patent database in a server managed by the Japan Patent Office. As yet another example, the patent information 125 is acquired from a patent database managed by another server.

[0039] The relevance evaluation between the technical information 123 and the patent information 125 is realized by utilizing the large-scale language model 324. More specifically, first, the information processing device 100 specifies the technical information 123 and the patent information 125 for a predetermined instruction sentence 128 for evaluating the relevance between the two pieces of information from a specific viewpoint.

[0040] The instruction statement 128 is, for example, registered in advance as a template in the information processing device 100. The instruction statement 128 also includes an argument portion 129A and an argument portion 129B. The technical information 123 is specified in the argument portion 129A. On the other hand, the patent information 125 is specified in the argument portion 129B.

[0041] Moreover, the instruction statement 128 specifies from what viewpoint the relevance between the technical information 123 and the patent information 125 is to be evaluated. In the example of FIG. 2, it specifies that the two pieces of information are to be evaluated from viewpoint "α". The viewpoint "α" may be specified in advance or may be specified as an argument. When the viewpoint "α" is specified as an argument, the viewpoint is arbitrarily specified by the user "A" of the user terminal 200.

[0042] The instruction sentence 128, in which the technical information 123 and the patent information 125 are specified, is input to the large-scale language model 324. Upon receiving the instruction sentence 128, the large-scale language model 324 generates a response according to the instruction sentence 128. The generated response is output to the information processing device 100.

[0043] Based on the answer obtained from the large-scale language model 324, the information processing device 100 outputs an evaluation result 130 indicating the relevance between the technical information 123 and the patent information 125 for the viewpoint "α". The relevance may be expressed by a numerical value indicating the degree of relevance, or may be expressed by an explanatory text. In the example of FIG. 2, the relevance is expressed by a numerical value.

[0044] The evaluation result 130 output by the information processing device 100 may be output to any destination. As an example, the output destination is the user terminal 200. The evaluation result 130 output to the user terminal 200 is displayed on the display of the user terminal 200, for example.

[0045] As described above, the information processing device 100 uses the large-scale language model 324 to evaluate the relevance between the technical information 123 and the patent information 125, thereby providing new value to the user.

[0046] The relevance assessment function is effective in the following cases, for example. Until now, patent research has mainly been carried out by humans reading and screening or reviewing thousands to tens of thousands of patent documents. In this case, if a user uses the relevance assessment function shown in FIG. 2, the user can easily determine the degree to which the technical information 123 being researched is related to the patent information 125. This allows the user to significantly reduce the time required for screening or reviewing.

[0047] In the example of FIG. 2, the instruction 128 is defined to evaluate the relevance from a single perspective, “α”, but the instruction 128 may be defined to evaluate the relevance from multiple perspectives.

[0048] Fig. 3 is a diagram for explaining the relevance evaluation function according to the modified example. The instruction statement 128 shown in Fig. 3 is specified to evaluate the relevance between two pieces of information from multiple viewpoints "α" and "β". When the instruction statement 128 is input, the large-scale language model 324 generates an answer according to the instruction statement 128.

[0049] The information processing apparatus 100 outputs an evaluation result 130 indicating the relevance between the technical information 123 and the patent information 125 based on the result obtained from the large language model 324. The evaluation result 130 indicates the relevance between the technical information 123 and the patent information 125 for each of the viewpoints "α" and "β". The evaluation result 130 is transmitted from the information processing apparatus 100 to the user terminal 200 and displayed on the display of the user terminal 200.

[0050] Thereby, the user can determine whether the technical information 123 is relevant to the patent information 125 based on the relevance from multiple viewpoints. As a result, the user can more easily determine the degree of relevance between the technical information 123 to be investigated and the patent information 125.

[0051] <C. Types of viewpoints> Continuing to refer to FIGS. 2 and 3, the viewpoints "α" and "β" defined in the instruction text 128 will be described.

[0052] Any viewpoint can be defined in the instruction text 128. Preferably, the matters that the user emphasizes when evaluating the relevance between the technical information 123 and the patent information 125 can be defined as viewpoints. As an example, the judgment criteria of "motivation for applying the secondary cited invention to the primary cited invention" described in the "Examination Criteria, Part III, Chapter 2, Section 2, Inventive Step" are defined as the viewpoints "α" and "β".

[0053] More specifically, the viewpoints "α" and "β" include at least one of the viewpoints of commonality in technical field, commonality in problems, commonality in functions, commonality in content suggestion, commonality in parameters, commonality in numerical limitations, and commonality in technical terms. Thereby, the commonality between the technical information 123 and the patent information 125 is evaluated on an evaluation axis that meets the judgment criteria of inventive step.

[0054] Note that the viewpoints defined in the instruction 128 are not limited to the above examples. As another example, the viewpoints defined in the instruction 128 may be viewpoints suitable for use exploration. The viewpoints may be viewpoints of the commonality of the applied use or product, or may be viewpoints of the commonality of the applied industrial field.

[0055] As yet another example, the viewpoints defined in the instruction 128 may be viewpoints of the commonality of the drawings. This specific example will be described later.

[0056] As yet another example, the viewpoints defined in the instruction 128 may be viewpoints for evaluating whether there are inhibiting factors regarding combining one of the inventions related to the technical information 123 and the inventions related to the patent information 125 with the other.

[0057] As yet another example, the viewpoints may be viewpoints for evaluating whether the difference between the inventions related to the technical information 123 and the inventions related to the patent information 125 is a design change.

[0058] <D. Hardware Configuration> Next, with reference to FIGS. 4 and 5, the hardware configurations of the information processing apparatus 100 and the user terminal 200 shown in FIG. 1 above will be described in order.

[0059] Note that since the hardware configuration of the server 300 shown in FIG. 1 is the same as that of the information processing apparatus 100, the description thereof will not be repeated.

[0060] (D1. Information Processing Apparatus 100) First, with reference to FIG. 4, the hardware configuration of the information processing apparatus 100 shown in FIG. 1 will be described. FIG. 4 is a schematic diagram showing an example of the hardware configuration of the information processing apparatus 100.

[0061] The information processing device 100 includes a control device 101, a read only memory (ROM) 102, a random access memory (RAM) 103, a communication interface 104, a display interface 105, an input interface 107, and an auxiliary storage device 120. These components are connected to a bus 110.

[0062] The control device 101 is configured, for example, by at least one integrated circuit. The integrated circuit may be configured, for example, by at least one central processing unit (CPU), at least one graphics processing unit (GPU), at least one application specific integrated circuit (ASIC), at least one field programmable gate array (FPGA), or a combination thereof.

[0063] The control device 101 controls the operation of the information processing device 100 by executing various programs such as an analysis program 122 and an operating system. Based on receiving an execution command for each program, the control device 101 reads the program from the auxiliary storage device 120 or the ROM 102 to the RAM 103. The RAM 103 functions as a working memory and temporarily stores various data required for executing the various programs.

[0064] A LAN (Local Area Network), an antenna, and the like are connected to the communication interface 104. The information processing device 100 exchanges data with external devices via the communication interface 104. The external devices include, for example, a user terminal 200, a server 300, and other communication devices.

[0065] A display 106 is connected to the display interface 105. The display interface 105 sends an image signal for displaying an image to the display 106 in accordance with a command from the control device 101 or the like. The display 106 is, for example, a liquid crystal display, an organic EL (Electro Luminescence) display, or other display device. The display 106 may be configured integrally with the information processing device 100, or may be configured separately from the information processing device 100.

[0066] An input device 108 is connected to the input interface 107. The input device 108 is, for example, a mouse, a keyboard, a touch panel, or other device capable of accepting a user's operation. Note that the input device 108 may be configured integrally with the information processing device 100, or may be configured separately from the information processing device 100.

[0067] The auxiliary storage device 120 is, for example, a hard disk, a flash memory, an SSD (Solid State Drive), or other storage medium. The auxiliary storage device 120 stores an analysis program 122, a patent database 124, the above-mentioned instruction statement 128, and the like. The patent database 124 includes a plurality of patent information 125. The storage location of the analysis program 122, the patent database 124, and the instruction statement 128 is not limited to the auxiliary storage device 120, and may be stored in a memory area (for example, a cache memory, etc.) of the control device 101, the ROM 102, the RAM 103, an external device, and the like.

[0068] The analysis program 122 may be provided not as a standalone program but as part of an arbitrary program. In this case, various processes defined in the analysis program 122 are realized in cooperation with an arbitrary program such as the analysis program 222 described later. Even a program that does not include some of such modules does not deviate from the spirit of the analysis program 122 according to this embodiment. Furthermore, some or all of the functions provided by the analysis program 122 may be realized by dedicated hardware. Furthermore, the information processing device 100 may be configured in a form similar to a so-called cloud service in which at least one server executes part of the processing of the analysis program 122.

[0069] (D2. User terminal 200) Next, the hardware configuration of the user terminal 200 shown in Fig. 1 will be described with reference to Fig. 5. Fig. 5 is a schematic diagram showing an example of the hardware configuration of the user terminal 200.

[0070] The user terminal 200 includes a control device 201, a ROM 202, a RAM 203, a communication interface 204, a display interface 205, an input interface 207, and an auxiliary storage device 220. These components are connected to a bus 210.

[0071] The control device 201 is configured, for example, by at least one integrated circuit. The integrated circuit may be configured, for example, by at least one CPU, at least one GPU, at least one ASIC, at least one FPGA, or a combination thereof.

[0072] The control device 201 controls the operation of the user terminal 200 by executing various programs such as an analysis program 222 and an operating system. Based on receiving an execution command for each program, the control device 201 reads the program from the auxiliary storage device 220 or the ROM 202 to the RAM 203. The RAM 203 functions as a working memory and temporarily stores various data required for executing the program.

[0073] A LAN, an antenna, and the like are connected to the communication interface 204. The user terminal 200 exchanges data with external devices via the communication interface 204. The external devices include, for example, the information processing device 100, the server 300, and other communication devices. The user terminal 200 may be configured to be able to download the analysis program 222 from the information processing device 100.

[0074] A display 206 is connected to the display interface 205. The display interface 205 sends an image signal for displaying an image to the display 206 in accordance with a command from the control device 201 or the like. The display 206 is, for example, a liquid crystal display, an organic EL display, or other display device. The display 206 may be configured integrally with the user terminal 200, or may be configured separately from the user terminal 200.

[0075] An input device 208 is connected to the input interface 207. The input device 208 is, for example, a mouse, a keyboard, a touch panel, or other device capable of accepting a user's operation. The input device 208 may be configured integrally with the user terminal 200, or may be configured separately from the user terminal 200.

[0076] The auxiliary storage device 220 is, for example, a hard disk, a flash memory, an SSD (Solid State Drive), or other storage medium. The auxiliary storage device 220 stores the analysis program 222 and the like. The storage location of the analysis program 222 is not limited to the auxiliary storage device 220, and may be stored in a storage area (for example, a cache memory, etc.) of the control device 201, the ROM 202, the RAM 203, an external device (for example, a server), or the like.

[0077] The analysis program 222 may be provided not as a standalone program but as part of an arbitrary program. In this case, various processes defined in the analysis program 222 are realized in cooperation with an arbitrary program such as the above-mentioned analysis program 122. Even a program that does not include some of these modules does not deviate from the spirit of the analysis program 222 according to this embodiment. Furthermore, some or all of the functions provided by the analysis program 222 may be realized by dedicated hardware. Furthermore, the user terminal 200 may be configured in a form similar to a so-called cloud service in which at least one server executes part of the processing of the analysis program 222.

[0078] <E.データフロー> Next, the operation of the information processing system 10 relating to the relevance evaluation between the technical information 123 and the patent information 125 will be described with reference to Fig. 6 to Fig. 11. Fig. 6 is a diagram showing an example of a data flow between the information processing device 100, the user terminal 200, and the server 300.

[0079] As described above, the information processing device 100 provides a function for evaluating the relevance between the technical information 123 and the patent information 125. In the following, the explanation will be given on the assumption that the relevance evaluation function is applied to the review of the patent information 125.

[0080] (E1. Step S110) In step S110, the user terminal 200 displays the input screen 400A shown in Fig. 7. Fig. 7 shows an example of the input screen 400A.

[0081] The input screen 400A accepts search conditions for searching the patent information 125 from the above-mentioned patent database 124 (see FIG. 4). By inputting the search conditions, the user can specify one or more pieces of patent information 125 to be compared with the technical information 123. The input screen 400A includes, for example, a selection field 410 and an input field 412.

[0082] A selection input of a search type is accepted in the selection field 410. For example, search types that can be specified in the selection field 410 include "full-text search", "field search", "command search", "semantic search", and "number search".

[0083] The input field 412 accepts input of various search conditions. Examples of search conditions that can be input into the input field 412 include the patent document number, the type of the number (e.g., application number, publication number, etc.), and the country in which the application was published.

[0084] (E2. Step S112) Next, in step S112, the user terminal 200 displays an input screen 400A shown in Fig. 8. Fig. 8 shows an input screen 400B as an example.

[0085] The input screen 400B receives settings for the review conditions. As an example, the input screen 400B includes an input field 420, a selection field 422, an input field 424, a selection field 426, and a start button 428.

[0086] The input field 420 accepts settings for review targets. As an example, the input field 420 accepts input of the maximum number of patent information 125 to be reviewed and review parts in the patent information 125. Examples of review parts that can be specified include the abstract, the title of the invention, the top claim in the claims (i.e., claim 1), the technical field, the background art, the problem that the invention is trying to solve, the effect of the invention, the means for solving the problem, and the specification.

[0087] A selection field 422 accepts the selection of a peer review type. Selectable peer review types include, for example, relevance evaluation, user instructions such as summarization and information extraction, user tag / evaluation, and the like.

[0088] The input field 424 accepts input of the above-mentioned technical information 123. The input field 424 accepts input of, for example, a sentence indicating a technical idea. As one example, the user inputs the contents of an invention made by the user or another company into the input field 424. As another example, the user inputs the specifications of the company's product into the input field 424.

[0089] The selection field 426 accepts the selection of the large-scale language model 324 to be used. The large-scale language model 324 selected in the selection field 426 is used for evaluating the relevance between the technical information 123 and the patent information 125.

[0090] Based on the user pressing the start review button 428, the user terminal 200 transmits the search conditions entered into the input screen 400A shown in FIG. 7 and the review conditions entered into the input screen 400B shown in FIG. 8 to the information processing device 100.

[0091] (E3. Step S120) Next, in step S120, the information processing device 100 searches for patent information 125 that matches the search conditions entered on the above-mentioned input screen 400A (see FIG. 7) from among the patent information 125 registered in the above-mentioned patent database 124. The information processing device 100 regards the patent information 125 that matches the search conditions as a population to be compared with the technical information 123.

[0092] (E4. Step S122) Next, in step S122, the information processing device 100 generates the above-mentioned instruction sentence 128 (see FIGS. 2 and 3) to be input to the large-scale language model 324.

[0093] 9 is a diagram showing an example of a directive 128. The directive 128 includes a plurality of argument parts 129A and a plurality of argument parts 129B. The information processing device 100 specifies technical information 123 for each of the plurality of argument parts 129A, and specifies patent information 125 for each of the plurality of argument parts 129B. The directive 128 is generated for each of the patent information 125 searched in step S120.

[0094] 9, the instruction sentence 128 preferably includes an instruction 131A for causing the large-scale language model 324 to output the relevance between the two pieces of information, and an instruction for causing the large-scale language model 324 to output the evaluation reason for the relevance. This increases the number of indicators for determining the relevance by the user.

[0095] The instruction 128 also includes an instruction 131C for synthesizing the evaluation results from each perspective and outputting the relevance of the two pieces of information to the large-scale language model 324, and an instruction 131D for outputting the reason for the overall evaluation to the large-scale language model 324.

[0096] The information processing device 100 transmits the instruction statement 128 generated in step S122 to the server 300.

[0097] (E5. Step S130) Next, in step S130, based on receiving the instruction statement 128 from the information processing device 100, the server 300 inputs the instruction statement 128 to the large-scale language model 324. As a result, the large-scale language model 324 generates a response according to the instruction statement 128.

[0098] Fig. 10 is a diagram showing answer information 326, which is an example of an answer generated by the large-scale language model 324. In the example of Fig. 10, the answer information 326 is shown in a table format, but the format of the answer information 326 is arbitrary. The output format of the large-scale language model 324 is, for example, predefined in the instruction statement 128, and the large-scale language model 324 outputs the answer information 326 according to the output format defined in the instruction statement 128.

[0099] The response information 326 includes evaluation results 327 of the relevance between the technical information 123 and the patent information 125 for each piece of patent information 125. The evaluation results 327 include a relevance score indicating the degree of relevance between the technical information 123 and the patent information 125, and an evaluation reason for the degree of relevance. The relevance score is included for each evaluation perspective instructed in the instruction statement 128. Similarly, the evaluation reason is included for each evaluation perspective instructed in the instruction statement 128. Furthermore, the evaluation results 327 include an overall evaluation score for each evaluation perspective, and the overall evaluation reason.

[0100] The evaluation result 327 can be uniquely identified by an identifier of the patent information 125. The identifier is defined by, for example, the application number of the patent information 125, the publication number of the patent information 125, the registered publication number of the patent information 125, the applicant information of the patent information 125, or the like.

[0101] The information processing device 100 transmits the response information 326 generated in step S130 to the information processing device 100.

[0102] (E6. Steps S140, S150) Next, in step S140, information processing device 100 generates an evaluation result screen based on response information 326 received from server 300. The evaluation result screen is written in a language such as HTML (HyperText Markup Language).

[0103] Next, in step S150, the user terminal 200 displays the evaluation result screen generated by the information processing device 100. Fig. 11 is a diagram showing an example of the evaluation result screen 400C. The evaluation result screen 400C is displayed on the display 206 of the user terminal 200, for example.

[0104] The evaluation result screen 400C includes a display area 430 that displays information related to the technical information 123, and a display area 440 that displays the evaluation result of the relevance between the technical information 123 and the patent information 125. The display area 440 includes a display area 442 and evaluation result display areas 444, 446. The display areas 440 are arranged according to the number of pieces of patent information 125 compared with the technical information 123.

[0105] The display area 430 displays information entered in the above-mentioned input screens 400A and 400B (see FIGS. 7 and 8), etc. As an example, the display area 430 displays "ID" (Identification), "Review Type", "Model", "Status", "Handling Time", "Review Subject", and "Research Perspective".

[0106] "ID" is an identifier for uniquely identifying the peer review evaluation results. "Peer review type" corresponds to the item set in the selection field 422 (see FIG. 8) described above. "Model" indicates the type of large-scale language model 324 used during relevance evaluation, and corresponds to the information set in the selection field 426 (see FIG. 8) described above. "Status" indicates whether the relevance evaluation process was completed successfully or not. "Handling time" indicates the time required from the start to the end of the relevance evaluation process. "Peer review subject" corresponds to the information entered in the input field 420 (see FIG. 8) described above. "Research perspective" indicates the content of the technical information 123, and corresponds to the item entered in the input field 420 (see FIG. 8) described above.

[0107] The display area 442 displays information related to the patent information 125 that was the subject of the relevance evaluation. For example, the display area 442 displays the "title of the invention", the "application number", the "publication number", the "status", and the "abstract".

[0108] In the "Summary" field, for example, an outline of the contents of the patent information 125 is displayed. As one example, the outline is generated from the results output from the large-scale language model 324 when the large-scale language model 324 is instructed to summarize the contents of the patent information 125. As another example, the outline is generated from the abstract included in the patent information 125.

[0109] In the display area 444, the evaluation result generated from the answer information 326 (see FIG. 10) of the large-scale language model 324 is displayed in a graph. In the example of FIG. 11, the evaluation result is shown in a radar chart of evaluation scores from each viewpoint. Preferably, the type of graph to be displayed in the display area 444 can be arbitrarily set by the user.

[0110] In the display area 446, evaluation results generated from the answer information 326 (see FIG. 10) of the large-scale language model 324 are displayed in a table format. In the example of FIG. 11, in the display area 446, the relevance scores for each viewpoint included in the answer information 326, the evaluation reasons for each viewpoint included in the answer information 326, the overall evaluation score included in the answer information 326, and the overall evaluation reason included in the answer information 326 are displayed.

[0111] In this way, the information processing device 100 outputs a comprehensive evaluation result that summarizes the evaluation results for each viewpoint. This allows the user to easily recognize the degree of association between the technical information 123 and the patent information 125.

[0112] Preferably, a "total score" is further displayed in the display area 446 as the overall evaluation result. The "total score" corresponds to the sum of the evaluation result scores for each viewpoint. Note that an "average score" of the evaluation result scores for each viewpoint may be displayed instead of the "total score."

[0113] The evaluation results displayed on the evaluation result screen 400C may be configured to be selectable for each evaluation perspective. As an example, when the user evaluates the relevance focusing on the "perspective of commonality of issues", the user deselects perspectives other than the "perspective of commonality of issues". As a result, only the evaluation results related to the "perspective of commonality of issues" are displayed on the evaluation result screen 400C. With this configuration, the user can evaluate the relevance between the technical information 123 and the patent information 125 according to his / her own purpose. The selection of the evaluation perspective may be reflected in the instruction statement 128 before the generation of the answer information 326, or may be reflected only in the display content of the evaluation result screen 400C.

[0114] The evaluation criteria displayed on the evaluation result screen 400C may be configured to be rearrangeable, which allows the user to display evaluation items with higher priority according to their own purpose at the top.

[0115] Furthermore, the multiple display areas 440 displayed on the evaluation result screen 400C are configured to be rearrangeable according to selected sorting conditions, such as ascending / descending order of evaluation score, ascending / descending order of overall evaluation score, and ascending / descending order of total score.

[0116] Furthermore, the contents displayed on the evaluation result screen 400C are not limited to the example of FIG. 11. More information (hereinafter, also referred to as "additional information") may be displayed on the evaluation result screen 400C. Examples of the additional information include a technical summary of the patent information 125, a problem summary of the patent information 125, characteristic keywords contained in the patent information 125, and a classification assigned to the patent information 125. The additional information is generated by the large-scale language model 324.

[0117] <I.フローチャート> Next, an example of a control flow of the information processing device 100 will be described with reference to Fig. 12. Fig. 12 is a diagram showing a flowchart relating to the evaluation of the relevance between the technical information 123 and the patent information 125.

[0118] 12 is realized by the control device 101 of the information processing device 100 executing the above-described analysis program 122. In another aspect, a part or all of the processing may be executed by circuit elements or other hardware.

[0119] In step S210, the control device 101 judges whether the relatedness assessment start button 428 (see FIG. 8) has been pressed. If the control device 101 judges that the relatedness assessment start button 428 has been pressed (YES in step S210), the control device 101 switches the control to step S212. If not (NO in step S210), the control device 101 executes the process of step S210 again.

[0120] In step S212, the control device 101 searches for patent information 125 that matches the search conditions inputted on the input screen 400A (see FIG. 7) from among the patent information 125 registered in the patent database 124.

[0121] In step S214, the control device 101 generates the instruction statement 128 described above to be input to the large-scale language model 324. The process of generating the instruction statement 128 is as described above, and therefore will not be described repeatedly. The control device 101 transmits the generated instruction statement 128 to the server 300.

[0122] The server 300 inputs the received instruction sentence 128 into the large-scale language model 324, and generates answer information 326 (see FIG. 10) according to the instruction sentence 128. The answer information 326 is transmitted to the information processing device 100.

[0123] In step S216, the control device 101 generates the above-mentioned evaluation result screen 400C (see FIG. 11) based on the response information 326 received from the server 300. Since the evaluation result screen 400C is as described above, the description thereof will not be repeated.

[0124] In step S218, the control device 101 outputs the evaluation result screen 400C generated in step S216 to the user terminal 200. As a result, the user terminal 200 displays the evaluation result screen 400C on the display 206.

[0125] <J1. Others> Next, other examples of the above embodiment will be described.

[0126] As described above, the comprehensive evaluation results such as the "Comprehensive Evaluation Score" and "Total Score" shown in FIG. 11 are generated based on the results of evaluating the relevance between the technical information 123 and the patent information 125 from each perspective. At this time, the evaluation results for each perspective may be weighted.

[0127] FIG. 13 is a diagram showing an example of weights associated with each perspective. In the example of FIG. 13, a weight "w1" is associated with the perspective "α". A weight "w2" is associated with the perspective "β". A weight "w3" is associated with the perspective "γ". A weight "w4" is associated with the perspective "δ". The weights "w1" to "w4" may be preset or arbitrarily set by the user.

[0128] Also, in FIG. 13, a relevance score "s1" is shown as the evaluation result from the perspective "α". A relevance score "s2" is shown as the evaluation result from the perspective "β". A relevance score "s3" is shown as the evaluation result from the perspective "γ". A relevance score "s4" is shown as the evaluation result from the perspective "δ".

[0129] The information processing device 100 calculates a comprehensive evaluation result in consideration of the weights "w1" to "w4" for the evaluation results "s1" to "s4" from different perspectives. As an example, the information processing device 100 calculates a weighted average by multiplying the evaluation results "s1" to "s4" by the weights "w1" to "w4", respectively. The information processing device 100 outputs the calculated weighted average as the comprehensive evaluation result.

[0130] <J2. Others> Next, another example of the above embodiment will be described.

[0131] In some cases, the technical information 123 and the patent information 125 may be compared in a broader concept than expected. In preparation for this case, the information processing apparatus 100 is configured to be able to re-evaluate the relevance between the technical information 123 and the patent information 125. Specifically, when the evaluation result screen 400C (see FIG. 11) is displayed on the display 206 of the user terminal 200, the information processing apparatus 100 receives an input regarding the concept level when comparing the patent information 125 with the technical information 123.

[0132] In one aspect, the "input regarding the concept level" is realized in an input field (not shown) within the evaluation result screen 400C. As an example, assume that as an evaluation result between the technical information 123 and the patent information 125, a result of "Both patents are related to an information processing system and its method." is displayed. In this case, the user may input, for example, as the "input regarding the concept level", a statement such as "[The concept such as 'information processing system and its method' is too broad, so please compare with a specific concept such as 'information processing system for appropriately evaluating strategies'.]" By inputting such an instruction statement, the "input regarding the concept level" is realized.

[0133] In another aspect, the "input regarding the concept level" is realized in a selection field (not shown) within the evaluation result screen 400C. The options in the selection field indicate the concept level at the time of evaluation. Examples of the options include the concept "broad", the concept "normal", and the concept "narrow".

[0134] An instruction statement for re-evaluating the relevance based on the first concept is associated with the concept "broad". An instruction statement for re-evaluating the relevance based on the second concept is associated with the concept "medium". The second concept is narrower than the first concept. An instruction statement for re-evaluating the relevance with the third concept is associated with the concept "narrow". The third concept is narrower than the second concept.

[0135] After that, the information processing apparatus 100 generates an instruction statement for re-evaluation that reflects the "input regarding the concept level". The method of reflection is arbitrary. As an example, the information processing apparatus 100 includes the "input regarding the concept level" in the instruction statement for re-evaluation.

[0136] After that, the information processing apparatus 100 inputs the generated instruction statement for re-evaluation into the large language model 324. Then, the information processing apparatus 100 outputs the result obtained from the large language model 324. Thereby, the user can perform the relevance evaluation while gradually narrowing down the concept level.

[0137] Note that the timing for receiving the "input regarding the concept level" is not limited to after the execution of the relevance evaluation, and it may also be before the execution of the relevance evaluation.

[0138] As an example, the "input regarding the concept level" may be received on the above-described input screens 400A, 400B (see FIGS. 7 and 8) for setting the conditions related to the relevance evaluation. The "input regarding the concept level" is reflected in the above-described instruction statement 128 (see FIG. 9, etc.). Thereby, the user can perform the relevance evaluation by specifying the concept level.

[0139] <J3. Others> Next, still other examples of the above embodiment will be described.

[0140] In the example of FIG. 11 described above, the evaluation score and the reason for evaluation output from the large language model 324 were output as they were on the evaluation result screen 400C. However, it is not necessarily the case that the output of the large language model 324 is output as it is on the evaluation result screen 400C.

[0141] As an example, the information processing apparatus 100 calculates the similarity between the technical information 123 and the patent information 125 based on a predetermined algorithm. Any technique can be adopted for the similarity calculation algorithm. As an example, the information processing apparatus 100 extracts feature vectors based on sentences, images, patent classifications, etc. from each of the technical information 123 and the patent information 125. Then, the information processing apparatus 100 calculates the similarity between the feature vector extracted from the technical information 123 and the feature vector extracted from the patent information 125. The similarity is, for example, the cosine similarity. To extract the feature vector, for example, indicators corresponding to the TF, TFIDF, BM25, and other frequencies of words, indicators obtained by models that generate word vectors such as Word2Vec and FastText, and indicators obtained by sentence encoder models such as BERT can be used.

[0142] Then, the information processing apparatus 100 combines the evaluation results by perspective obtained from the large language model 324 with the calculated similarity described above and outputs it as an evaluation result. The method of combining the evaluation results by perspective and the similarity is arbitrary. As an example, the information processing apparatus 100 combines the evaluation results by perspective obtained from the large language model 324 by adding the calculated similarity.

[0143] <J4. Others> Next, still another example of the above embodiment will be described.

[0144] In the above example, when the information processing apparatus 100 evaluates the relevance between the technical information 123 and the patent information 125 from each perspective, one large language model 324 is used. However, the information processing apparatus 100 may evaluate the relevance between the technical information 123 and the patent information 125 using different large language models 324 for each perspective.

[0145] FIG. 14 is a diagram schematically showing the process of relevance evaluation in this example. Different types of large language models 324A and 324B are shown in FIG. 14.

[0146] For the large language model 324A, an instruction statement 128A for evaluating the relevance between the technical information 123 and the patent information 125 from the perspective "α" is input. On the other hand, for the large language model 324B, an instruction statement 128B for evaluating the relevance between the technical information 123 and the patent information 125 from the perspective "β" is input. Thus, in this example, the large language models 324A and 324B to be used are determined according to different evaluation perspectives.

[0147] Typically, the large language model 324A is tuned to be suitable for the relevance evaluation from the perspective "α". Similarly, the large language model 324B is tuned to be suitable for the relevance evaluation from the perspective "β". For this tuning, for example, fine-tuning of the Chat GPT API or fine-tuning of open-source models such as Llama2 and Japanese Stable LM is used.

[0148] <J5. Others> Next, further other examples of the above embodiment will be described.

[0149] In the above example, the information processing device 100 evaluated the relevance between the text content of the technical information 123 and the text content of the patent information 125. In contrast, in this example, the information processing device 100 evaluates the relevance between the content of the drawings included in the technical information 123 and the drawings included in the patent information 125.

[0150] Figure 15 is a diagram schematically showing the process of relevance evaluation in this example. As shown in Figure 15, the technical information 123 includes one or more drawings 123F. Similarly, the patent information 125 includes one or more drawings 125F. Also, the instruction statement 128 is defined to evaluate the relevance of the two pieces of information from the perspective of the content of the drawings. The drawings 123F and 125F are, for example, image data. The format of the image is not particularly limited.

[0151] The information processing apparatus 100 designates the technical information 123 for the argument part 129A of the instruction statement 128 and designates the patent information 125 for the argument part 129B of the instruction statement 128. The instruction statement 128 for which the technical information 123 and the patent information 125 are designated is transmitted from the information processing apparatus 100 to the server 300 and input to the large language model 324 in the server 300.

[0152] When receiving the input of the instruction statement 128, the large language model 324 generates an answer according to the instruction statement 128. Note that the large language model 324 is learned to be able to interpret the content of the image. Thereby, the large language model 324 interprets the content of the drawing 123F included in the technical information 123 and also interprets the content of the drawing 125F included in the patent information 125. As the large language model capable of interpreting the content of the image, for example, CLIP (Contrastive Language-Image Pretraining), BLIP (Bootstrapping Language Image Pre-training for unified vision-language understanding and generation), ViLBERT (Vision-and-Language BERT), etc. can be used.

[0153] Based on each interpretation result obtained from the large language model 324, the information processing apparatus 100 outputs an evaluation result 130 indicating the relevance between the technical information 123 and the patent information 125 from the perspective of the content of the drawing. The relevance may be represented by a numerical value indicating the degree of relevance or may be represented by a descriptive text. In the example of FIG. 15, the relevance is represented by a numerical value.

[0154] When the above-mentioned relevance is output as a descriptive text, the descriptive text is indicated by words such as "The drawing 123F included in the technical information 123 has a structure of XXX and does not match the part of XXX compared with the drawing 125F included in the patent information 125."

[0155] <J6. Others> Next, a further example of the above embodiment will be described.

[0156] In this example, the information processing device 100 evaluates the relevance between the technical information 123 and the patent information 125 in stages from the same viewpoint, thereby improving the accuracy of the relevance evaluation.

[0157] 16 is a diagram showing an instruction statement 128C in this example. Instruction statement 128C includes instruction 132A (first instruction) and instruction 132B (second instruction). Instruction 132A is an instruction to have the relevance of two pieces of information evaluated from a specific perspective. Instruction 132B is an instruction to have the relevance of the two pieces of information further evaluated from the same perspective using the evaluation result of the relevance of the two pieces of information based on instruction 132A.

[0158] The information processing device 100 specifies the technical information 123 for the argument part 129A of the directive 128C, and specifies the patent information 125 for the argument part 129B of the directive 128C. The directive 128C in which the technical information 123 and the patent information 125 are specified is input to the large-scale language model 324. Upon receiving the input of the directive 128C, the large-scale language model 324 evaluates the relevance between the technical information 123 and the patent information 125 in multiple stages according to the directive 128C.

[0159] The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present invention is defined by the claims, not the above description, and is intended to include all modifications within the meaning and scope of the claims. [Explanation of symbols]

[0160] 10 Information processing system, 100 Information processing device, 101 Control device, 102 ROM, 103 RAM, 104 Communication interface, 105 Display interface, 106 Display, 107 Input interface, 108 Input device, 110 Bus, 120 Auxiliary storage device, 122 Analysis program, 123 Technical information, 123F Drawing, 124 Patent database, 125 Patent information, 125F Drawing, 128 Instruction, 128A Instruction, 128B Instruction, 128C Instruction, 129A Argument part, 129B Argument part, 130 Evaluation result, 132A Instruction, 132B Instruction, 200 User terminal, 201 Control device, 202 ROM, 203 RAM, 204 Communication interface, 205 Display interface, 206 Display, 207 Input interface, 208 Input device, 210 Bus, 220 auxiliary storage device, 222 analysis program, 300 server, 324 large-scale language model, 324A large-scale language model, 324B large-scale language model, 326 answer information, 400A input screen, 400B input screen, 400C evaluation result screen, 410 selection field, 412 input field, 420 input field, 422 selection field, 424 input field, 426 selection field, 428 start button, 430 display area, 440 display area, 442 display area, 444 display area, 446 display area, NW network.

Claims

1. A patent information analysis program, The analysis program is configured to: acquiring technical information including a technical idea; obtaining patent information to be compared with the technical information; A step of designating the technical information and the patent information with respect to a predetermined instruction statement for evaluating the relevance of the two pieces of information from a specific viewpoint; and outputting an evaluation result indicating the relevance between the technical information and the patent information for the viewpoint based on the result obtained from a large-scale language model by inputting the instruction sentence in which the technical information and the patent information are specified into the large-scale language model.

2. The predetermined instruction sentence is defined to evaluate the relevance of two pieces of information from a plurality of viewpoints, 2. The analysis program according to claim 1, wherein in said outputting step, an evaluation result indicating a relevance between said technical information and said patent information is output for each of said plurality of viewpoints.

3. The analysis program according to claim 2, wherein the multiple perspectives include at least one of a technical field perspective, a problem perspective, an action / function perspective, a content suggestion perspective, a parameter perspective, a numerical limitation perspective, and a technical term perspective.

4. The analysis program further includes:

4. The analysis program according to claim 2, further comprising a step of outputting a comprehensive evaluation result that summarizes the different evaluation results of the viewpoints.

5. Each of the plurality of perspectives has an associated weight; The analysis program according to claim 4 , wherein the comprehensive evaluation result is calculated by weighting the other evaluation results of the viewpoints.

6. The predetermined instruction sentence is: An instruction to output a degree of relevance of the two pieces of information to the large-scale language model; and an instruction to cause the large-scale language model to output a reason for the assessment of the relevance; The evaluation results are as follows: the degree of association between the technical information and the patent information; The analysis program according to any one of claims 1 to 3, further comprising the evaluation reasons relating to the technical information and the patent information.

7. The analysis program further includes: receiving an input regarding a conceptual level at which the patent information is compared with the technical information; 4. The analysis program according to claim 1, wherein the step of specifying further includes reflecting an input regarding the conceptual level in the directive statement.

8. The analysis program further includes: receiving an input regarding a conceptual level at which the patent information is compared with the technical information while the evaluation result is being displayed on a display; generating instructions for re-evaluation that reflect the input regarding the concept level; and a step of inputting the instruction sentence for reevaluation into the large-scale language model and outputting a result obtained from the large-scale language model.

9. The analysis program further includes: Executing a step of calculating a similarity between the technical information and the patent information based on a predetermined algorithm; 4. The analysis program according to claim 1, wherein the evaluation results are output for each of the viewpoints based on the results obtained from the large-scale language model and the similarity.

10. The above aspects include: The first viewpoint, and a second aspect different from the first aspect, The large-scale language model is a first language model; a second language model different from the first language model; The instruction sentence for evaluating the relevance between the technical information and the patent information from the first perspective is input to the first language model; The analysis program according to claim 2 or 3, wherein the instruction sentence for evaluating the relevance between the technical information and the patent information from the second perspective is input to the second language model.

11. The technical information includes drawings, The patent information includes drawings, The analysis program according to any one of claims 1 to 3, wherein the viewpoint includes a viewpoint of the contents of a drawing.

12. The instruction: A first instruction for evaluating the relevance of the two pieces of information from a specific viewpoint; and a second instruction for further evaluating the relevance of the two pieces of information from the specific perspective using a result of the evaluation of the relevance of the two pieces of information based on the first instruction.

13. An information processing device capable of analyzing patent information, A control unit for operating the information processing device, The control unit is A process of obtaining technical information including technical ideas; A process of obtaining patent information to be compared with the technical information; A process of designating the technical information and the patent information in response to a predetermined instruction for evaluating the relevance of the two pieces of information; and executing a process of inputting the instruction sentence in which the technical information and the patent information are specified into a large-scale language model, and outputting an evaluation result indicating the relevance between the technical information and the patent information for the perspective based on the result obtained from the large-scale language model.

14. 1. A computer-implemented method for analyzing patent information, comprising: acquiring technical information including a technical idea; obtaining patent information to be compared with the technical information; A step of designating the technical information and the patent information in response to a predetermined instruction for evaluating the relevance of the two pieces of information from a plurality of viewpoints; and outputting an evaluation result indicating the relevance between the technical information and the patent information for the perspective based on a result obtained from a large-scale language model by inputting the instruction sentence in which the technical information and the patent information are specified into the large-scale language model.

Images