Solution generation device, program, and recording medium

Abstract

We provide a solution generation device, program, and recording medium that uses generation AI to generate solutions to problems and evaluate the patentability of those solutions. [Solution] The solution generation device 1 includes a reception unit that receives input information from a user, a prompt creation unit that creates a prompt based on the input information received by the reception unit, a generation result acquisition unit that inputs the prompt into a trained generation model and acquires the generation result, a similar patent document information acquisition unit that acquires information on similar patent documents based on the generation result, and a presentation unit that presents the generation result and information on similar patent documents to the user.

Description

[Technical Field]

[0001] The present invention relates to a solution generation device, a program, and a recording medium. [Background technology]

[0002] In recent years, advances in artificial intelligence (AI) technology have led to increased automation and efficiency in a variety of fields. In particular, generative AI, which can generate new content based on given conditions, is expected to be used in many fields.

[0003] In addition, in the fields of research and development and product development, the creation of new ideas has become an important issue. In order to create novel inventions, it is necessary to research and analyze existing technologies and find original solutions that are differentiated from them.

[0004] Patent Document 1 discloses a data analysis device that accepts constraints that formally describe information used by a data analyst when selecting attributes to use in analysis from among explanatory variables, explanatory variables, and a target variable, and performs data analysis on attribute selection patterns that satisfy the constraints.

[0005] Patent Document 2 discloses an analysis device that receives relational data containing multiple objects and the relationships between them as input, estimates latent vectors that characterize the structure of the relational data, and associates objects between different data sets. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Publication No. 2018-088087 [Patent Document 2] Japanese Patent Publication No. 2020-021343 Summary of the Invention [Problem to be solved by the invention]

[0007] However, although Patent Document 1 focuses on automating attribute selection in data analysis, it does not take into consideration the creation of new ideas.

[0008] Furthermore, in Patent Document 2, although it is possible to associate objects between different data sets, it is not possible to generate solutions.

[0009] Therefore, an object of the present invention is to provide a solution generation device, a program, and a recording medium that can generate solutions to problems using a generation AI. [Means for solving the problem]

[0010] In order to solve the above problem, the solution generation device of the present invention is characterized by comprising: a reception unit that receives input information from a user; a prompt creation unit that creates a prompt based on the input information received by the reception unit; a generation result acquisition unit that inputs the prompt into a trained generation model and acquires the generation result; a similar patent document information acquisition unit that acquires information on similar patent documents based on the generation result; and a presentation unit that presents the generation result and information on similar patent documents to the user.

[0011] In this way, the present invention makes it possible to use generation AI to generate solutions for inventions and automatically obtain information on similar patent documents necessary to evaluate the patentability of the solutions, thereby enabling rapid determination of the patentability of the generated solutions.

[0012] In addition, the generated results and information on similar patent documents are automatically presented, allowing users to efficiently determine the patentability of the generated solution. [Effects of the Invention]

[0013] According to the present invention, it is possible to provide a solution generation device, program, and recording medium that can generate solutions to problems using generation AI and evaluate the patentability of those solutions. [Brief explanation of the drawings]

[0014] [Figure 1] 1 is a diagram showing the overall configuration of an example of a solution generation device according to the present invention; [Figure 2] FIG. 1 is a schematic diagram illustrating an example of a solution generation and patent search process according to the present invention. [Figure 3] 1 is a functional block diagram showing an example of a solution generation device according to the present invention; [Figure 4] 1 is a flowchart showing an example of information processing in the solution generation device according to the present invention. [Figure 5] FIG. 2 is a screen diagram showing an example of a user interface section of the solution generation device according to the present invention. [Figure 6] FIG. 1 is a diagram showing an example of a screen presenting the generated results and patent information in the solution generation device according to the present invention. [Figure 7] FIG. 10 is a schematic diagram illustrating an example of a solution generation and patent search process according to a second embodiment of the present invention. [Figure 8] FIG. 10 is a functional block diagram showing an example of a solution generation device according to a third embodiment of the present invention. [Figure 9] FIG. 10 is a diagram showing the overall configuration of an example of a solution generation system according to a fourth embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0015] The following describes embodiments of the present invention. The present invention is not limited to the embodiments described below, and can be modified, added, modified, deleted, or otherwise altered within the scope of what a person skilled in the art can conceive. Any embodiment that achieves the functions and effects of the present invention is within the scope of the present invention.

[0016] A solution generation device according to an embodiment of the present invention is characterized by comprising a reception unit that receives input information from a user, a prompt creation unit that creates a prompt based on the input information received by the reception unit, a generation result acquisition unit that inputs the prompt into a trained generation model and acquires the generation result, a similar patent document information acquisition unit that acquires information on similar patent documents based on the generation result, and a presentation unit that presents the generation result and information on the similar patent documents to the user.

[0017] Furthermore, when the prompt creation unit receives a problem as input information, it is preferable that the prompt creation unit analyzes the problem and creates a prompt for generating a solution to the problem.

[0018] It is also preferable to further include a search unit that searches for and acquires solutions to problems from web information on the Internet.

[0019] It is also preferable that the prompt creation unit further creates a patent search formula generation prompt that generates a patent search formula for obtaining information on similar patent documents based on the solution, and the similar patent document information acquisition unit inputs the patent search formula generation prompt into a trained generation model to obtain a patent search formula and inputs the patent search formula into a patent database to obtain information on similar patent documents.

[0020] It is also preferable that the prompt creation unit further creates a search prompt for searching for and acquiring information on similar patent documents based on the solution, and the similar patent document information acquisition unit inputs the search prompt into the trained generation model to acquire information on similar patent documents as a generation result.

[0021] Furthermore, the similar patent document information acquisition unit searches using words that represent technical features extracted from the solution and groups of words that are semantically related to the words, thereby: The similar patent document information acquisition unit acquiresIt is preferable to acquire information on similar patent documents. The similar patent document information acquisition unit evaluates the similarity of each document in the patent document database using a group of words representing technical components extracted from the generated results including the solution and a group of semantically related words. This similarity evaluation can be performed, for example, by scoring the word group based on the frequency of appearance between the group of words and terms described in the patent document (e.g., TF-IDF, etc.) or semantic distance (e.g., vector distance using a natural language processing model such as Word2Vec or BERT), and ranking the documents based on the score. Such evaluation methods are well known to those skilled in the art of natural language processing and information retrieval at the time of filing.

[0022] It is also preferable that the prompt creation unit further creates an evaluation prompt for obtaining an evaluation on the novelty and inventive step of the solution, and that the solution generation device further includes a patentability evaluation presentation unit that inputs the evaluation prompt into the trained generative model to obtain an evaluation and presents the evaluation to the user.

[0023] Furthermore, the present invention can be implemented as a program that causes a computer to function as a reception means for receiving input information from a user, a prompt creation means for creating a prompt based on the input information received by the reception means, a generation result acquisition means for inputting the prompt into a trained generation model and acquiring the generation result, a similar patent document information acquisition means for acquiring information on similar patent documents based on the generation result, and a presentation means for presenting the generation result and information on similar patent documents to the user.

[0024] The program can also be implemented as a computer-readable recording medium on which the program is recorded.

[0025] In addition, in a computer equipped with a processing unit and a storage device, the storage device has a model storage area for storing a trained generative model and a patent database storage area for storing a patent document database, and the processing unit is characterized by having a reception unit for receiving input information from a user, a prompt creation unit for creating a prompt based on the input information received by the reception unit, a generation result acquisition unit for inputting the prompt into the trained generative model to perform calculation processing and obtain a generation result related to a solution, a similar patent document information acquisition unit for acquiring information on similar patent documents from the patent database storage area based on the generation result, and a presentation unit for presenting the generation result and information on similar patent documents to the user. [Example]

[0026] The present invention will be explained in more detail below by showing examples, but the present invention is not limited to these examples.

[0027] Example 1 The configuration of a solution generation device 1 according to a first embodiment of the present invention will be described with reference to Fig. 1. The solution generation device 1 is connected to a first information processing device 10 and a second information processing device 20, which are external devices. The first information processing device 10 includes a trained generative model 12, and the second information processing device 20 includes a patent database 22.

[0028] The first information processing device 10 is an information processing device that realizes a generative AI and includes a trained generative model 12.

[0029] A trained generative model 12 is an AI model that has been pre-trained using large amounts of data and is capable of generating new content based on prompts.

[0030] The second information processing device 20 is an information processing device that realizes a patent database 22, which is a database of patent documents.

[0031] The solution generation device 1 comprises a processor 2, a RAM / ROM 3, a user interface section 4, a storage section 5 and an interface section 7. The storage section 5 is provided with a program storage area 6.

[0032] The processor 2 is a central processing unit that executes the arithmetic processing of each part of the solution generation device 1.

[0033] The RAM / ROM 3 is composed of a volatile memory device (RAM) that performs temporary storage and a non-volatile memory device (ROM).

[0034] The user interface unit 4 may be, for example, a touch panel provided on a liquid crystal display screen as an input interface, or may be an input device such as a keyboard or a mouse. It may also be a device that accepts audio information as input information, such as a microphone. Furthermore, the user interface unit 4 may be, as an output interface, a liquid crystal panel that displays images, or a device that provides audio information as output information, such as a speaker.

[0035] The storage unit 5 is a device that stores programs and data, and is configured with, for example, an HDD or SSD.

[0036] The program storage area 6 is a storage area included in the storage unit 5, and stores programs, data, etc. for performing information processing.

[0037] The interface unit 7 is a part that inputs and outputs data to and from an external information processing device, and performs wireless or wired communication.

[0038] The solution generation and patent search process according to the first embodiment of the present invention will be described with reference to Fig. 2. The solution generation device 1 creates, for example, a prompt P1 based on information input by a user and inputs it to the trained generative model 12. The trained generative model 12 outputs the generation results including the solutions to the solution generation device 1 according to the instructions of the prompt P1.

[0039] The solution generation device 1 creates a patent search formula based on the solution. Specifically, the solution generation device 1 creates a prompt to extract important keywords representing technical features from the solution and inputs this to the trained generative model 12. The solution generation device 1 classifies the important keywords output from the trained generative model 12 into terms related to purpose and effect, terms related to means and configuration, and terms related to the field of application. Next, the solution generation device 1 creates a prompt to obtain synonyms and similar words for each keyword and inputs this to the trained generative model 12. The solution generation device 1 also creates a prompt to automatically assign an International Patent Classification (IPC) from the extracted keywords and inputs this to the trained generative model 12. For the assigned IPC, the solution generation device 1 creates a prompt to identify related higher and lower classifications from the classification hierarchical structure and inputs this to the trained generative model 12. If necessary, the solution generation device 1 creates a prompt to assign Japan-specific classifications such as FI and F-terms and inputs this to the trained generative model 12. The solution generation device 1 executes the actual search by inputting the generated patent search formula into the patent database 22, and obtains, for example, the title of the invention, abstract, representative claim, representative drawing, applicant / inventor information, application date / publication date information, etc. In order to calculate the similarity of the obtained patent documents, the solution generation device 1 creates a prompt for similarity calculation and inputs it into the trained generative model 12. The solution generation device 1 ranks the patent documents based on their similarity, sorts them in descending order of similarity, and presents them to the user from the user interface unit 4.

[0040] The functional blocks of the solution generation device 1 will be described with reference to Fig. 3. The solution generation device 1 comprises a reception unit 30, a prompt creation unit 40, a generation result acquisition unit 50, a search unit 55, a similar patent document information acquisition unit 60, and a presentation unit 70.

[0041] The receiving unit 30 receives input information from the user, particularly input information such as tasks and conditions, and transmits the input information to the prompt generating unit 40.

[0042] The prompt creation unit 40 creates prompts based on the input information received by the reception unit 30. In particular, when a problem is received, it analyzes the problem and creates a prompt for generating a solution to the problem. It also creates a patent search query generation prompt for generating a patent search query for obtaining information on similar patent documents based on the solution, and a search prompt for searching for and obtaining information on similar patent documents.

[0043] The generation result acquisition unit 50 inputs the prompts created by the prompt creation unit 40 into the trained generation model 12 and acquires the generation results. The generation results include solutions to problems, patent search formulas, and the like.

[0044] The search unit 55 searches for and acquires solutions to problems from web information on the Internet. Specifically, the search unit 55 extracts technical terms contained in the input problem and keywords indicating related technical fields, and uses these keywords to perform a web search via a search engine API (for example, Google Custom Search API or Bing Web Search API). The search unit 55 selects reliable information sources from the web pages obtained as search results, for example, based on the following criteria: (1) Web pages belonging to academic or research institution domains (.ac.jp, .edu, etc.) (2) Literature registered in technical literature databases and academic paper databases (J-STAGE, CiNii, etc.) (3) Articles in industry magazines and technical information sites (4) Corporate technical documents and white papers (5) Presentation materials for technical conferences and seminars Next, the search unit 55 extracts the following information from the selected information sources using natural language processing techniques: (1) Description of specific solutions to the problem (2) Comparison with conventional technology and explanation of its advantages (3) Implementation examples and experimental results (4) Description of technical effect (5) Information on related technology trends and market needs The search unit 55 performs the following process to evaluate the freshness and relevance of the extracted information: (1) Judgment of freshness based on the publication date of information (2) Calculation of relevance score for the input task (3) Checking consistency of descriptions across multiple sources (4) Technical feasibility assessment This makes it possible to acquire highly reliable solution information that complements the results generated by the trained generative model 12. Furthermore, the search unit 55 can also collect newly released technical information by periodically performing searches. The acquired information is combined with the generated results and presented to the user by the presentation unit 70. Note that information collection by the search unit 55 is carried out in compliance with copyright law and the terms of use of each website.

[0045] The similar patent document information acquisition unit 60 acquires information on similar patent documents based on the acquired solution. Information on similar patent documents may be acquired by performing a search using words expressing technical features extracted from the generated results and a group of words semantically related to the extracted words.

[0046] The presentation unit 70 presents the generated solution and information on related patent documents to the user. Specifically, the presentation unit 70 displays the generated solution and information on related patent documents via the user interface unit 4.

[0047] The flow of information processing in the solution generation device 1 will be described with reference to Fig. 4. First, in step ST1, the receiving unit 30 receives input information from the user. The input information includes, for example, a technical problem to be solved and conditions related to the problem. Next, in step ST2, the prompt creating unit 40 analyzes the received input information and creates a prompt for the trained generative model 12. This prompt is configured to appropriately reflect the characteristics and requirements of the input problem.

[0048] In step ST3, the generation result acquisition unit 50 inputs the created prompt into the trained generative model 12 and acquires a solution as a generation result. In the following step ST4, the search unit 55 searches for and acquires related technical information and solutions from web information on the Internet. This makes it possible to obtain information that complements the generation result by the trained generative model 12.

[0049] In step ST5, the prompt creation unit 40 creates a patent search formula generation prompt for searching patent documents based on the obtained solution. In step ST6, this patent search formula generation prompt is input into the trained generation model 12 to obtain an optimal patent search formula. In step ST7, the obtained patent search formula is input into the patent database 22, and in step ST8, the similar patent document information acquisition unit 60 acquires information on related patent documents. When searching for similar patent documents, a search may also be performed using words that represent the technical features of the generated solution or their related words.

[0050] In the final step ST9, the presentation unit 70 presents the generated solution, supplementary information from Web information, and information on similar patent documents in a format that is easy for the user to understand.

[0051] The user interface of the solution generation device 1 will be described in more detail with reference to Figures 5 and 6. Screen IMG1 in Figure 5 has an input field TXT1 where the user inputs issues and conditions. This input field TXT1 is configured as a text area where the user can freely write the technical issues they are facing and the requirements for the solution. Below the input field TXT1, a button BT1 is provided for instructing the trained generative model 12 to generate a solution. This button BT1 is labeled "Generate Ideas," and clicking or touching it starts the process of generating a solution based on the input information. Screen IMG1 also has a preview area (not shown) for checking whether the input information is appropriate, a history display area (not shown) for referencing past input history, and so on.

[0052] Screen IMG2 in Figure 6 displays the generated solution TXT2 and information on similar patent documents TXT3 retrieved from the patent database 22. Solution TXT2 lists specific technical means for solving the problem in bullet points, with important keywords highlighted in each item. Information on similar patent documents TXT3 lists publication numbers, invention titles, abstracts, and similarities with the solution, and users can click on each document to view detailed information. There is also an editing area (not shown) where users can enter correction instructions to further improve the solution. These interface elements allow users to efficiently generate and evaluate solutions.

[0053] <Example 2> A second embodiment of the present invention will be described with reference to FIG. 7. In this embodiment, information on similar patent documents is directly acquired from the trained generative model 12 using prompt P2. Specifically, the solution generation device 1 creates a search prompt based on the solution generation result and inputs it into the trained generative model 12 to acquire information on similar patent documents. This method allows information on similar patent documents to be acquired directly from the trained generative model 12 without going through the patent database 22, enabling more efficient search processing. Furthermore, the trained generative model 12 can generate information such as the publication number, invention title, abstract, and similarity to the solution of the similar patent document based on the input search prompt. The generated information is presented to the user via the presentation unit 60 of the solution generation device 1. In this embodiment, the trained generative model 12 also functions as a patent database, further simplifying the overall system configuration.

[0054] The prompt generator 40 generates a search prompt based on the solution. The search prompt is a natural language instruction for instructing the trained generative model 12 to search for similar patent documents, and includes the following components: (1) Specify the purpose of the search "Search for patent documents similar to the solution below" "Search for patent documents containing the following technical features" A statement that clearly specifies the purpose of the search, such as (2) Description of the technical features of the solution -Specific description of the means to solve the problem Listing important components Description of technical effects A detailed description of the technical features, including (3) Specify the search range ·Search target technical field Search period Search target countries and regions Conditions to limit the search range, such as (4) Specify the output format Bibliographic information to be obtained (publication number, title of invention, applicant, etc.) Similarity calculation method ·Sort order Output format specifications such as

[0055] The prompt generator 40 generates search prompts according to, for example, the following algorithm: (1) Technical feature extraction step Extract key technical features from the solution ·Score the importance of each feature Extraction of keywords and related terms (2) Search context generation step Convert extracted features into natural-sounding sentences Clarifying the technical context · Clearly express search intent (3) Adding search conditions -Added refinement criteria by technical field -Setting the search range according to time · Geographical search range setting (4) Steps for describing output requirements -Specify necessary bibliographical information Specify the similarity calculation method · Specify the sort order of the results

[0056] For example, the following search prompt might be generated for a solution regarding image processing: "Search for patent documents similar to the following solutions: Technical features: - Extracting feature points from the input image - A process for generating feature values ​​by combining the position information and brightness information of feature points - The process of recognizing objects from their features using machine learning models Search range: - Technical fields: image processing, pattern recognition - Period: 2015 onwards - Countries: Japan, USA, Europe Output requirements: - Publication number, title of invention, abstract, representative drawing - Calculate the similarity score (0-100) with the above technical features - Sort by descending similarity score - Output top 20 results

[0057] The following optimizations are also performed during the search prompt generation process: (1) Appropriate selection of synonyms and related words (2) Weighting keywords to reduce search noise (3) Adding contextual information to improve search accuracy (4) Adjust the search scope appropriately This enables the trained generative model 12 to perform highly accurate patent document searches based on search prompts.

[0058] Example 3 A third embodiment of the present invention will be described with reference to FIG. 8. In this embodiment, a patentability evaluation presentation unit 170 is added. The patentability evaluation presentation unit 170 creates evaluation prompts for evaluating the novelty and inventive step of the generated solution and inputs them to the trained generative model 12 via the interface unit 7. The trained generative model 12 outputs evaluation results regarding the novelty and inventive step of the solution as the generated results. For example, for novelty, the evaluation output may be "The differences from similar patent documents are clear, and there is novelty." For inventive step, the evaluation output may be "An advancement from the prior art is recognized, and there is inventive step." The patentability evaluation presentation unit 170 presents these evaluation results to the user. Furthermore, the patentability evaluation presentation unit 170 can also present more detailed evaluation results, such as specific differences from similar patent documents that serve as the basis for the evaluation, and an explanation of the technical effect that supports the inventive step. This allows the user to more accurately determine the patentability of the generated solution. Specifically, in the novelty evaluation, the solution is analyzed from the following perspectives: (1) Comparative analysis of components Extract each component of the solution and compare it in detail with the corresponding components of similar patent documents Clarifying commonalities and differences Identification of novel components (2) Combinatorial analysis - Novelty evaluation of the method of combining components -Analysis of interactions and relationships between components - Identifying new combinations not found in existing technologies (3) Parameter analysis Comparison of technical parameters such as numerical ranges, ratios, conditions, etc. - Identifying numerical ranges and conditions that differ from existing technologies

[0059] In assessing inventive step, the solution is analysed in terms of: (1) Identifying the issues - Clarifying the problem the solution is trying to solve Comparison with issues faced by conventional technologies -Analysis of differences in problem-solving approaches (2) Motivation analysis -Evaluation of the ease of combining multiple patent documents -Relevance analysis of technical fields Check for any obstructive factors (3) Analysis of effects Identifying the technical effect of the solution -Whether there are significant effects compared to conventional technologies Check for unexpected effects

[0060] Based on the above analysis, the trained generative model 12 outputs the following evaluation results: <Example of novelty evaluation> "Novelty due to the novel combination of components A, B, and C" "Novelty exists by defining a numerical range X that is not disclosed in the prior art" "The control method for component D is different from conventional technology and is therefore novel." <Example of inventive step evaluation> "There is no motivation to combine References 1 and 2, so there is an inventive step." "The approach to solving the problem is different from conventional technology and is inventive" "Unpredictable and significant effect Y was observed, and there is an inventive step."

[0061] The patentability evaluation presentation unit 170 presents these evaluation results to the user along with specific reasons. The evaluation results also include the following supplementary information: (1) Explanation of specific differences from similar patent documents (2) Specific examples of technical effects that support the inventive step (3) Proposal of amendments for patent protection (4) Points to note and suggestions for improvement regarding fulfillment of patent requirements

[0062] Furthermore, in order to improve the reliability of the evaluation results, the patentability evaluation presenting unit 170 also performs the following processes: (1) Cross-checking from multiple perspectives (2) Check the consistency of the evaluation results (3) Confirm the appropriateness of the cited references that form the basis of the evaluation (4) Check for consistency with industry standards and common technical knowledge This allows users to more accurately determine the patentability of the generated solutions and consider patenting strategies.

[0063] Example 4 A fourth embodiment of the present invention will be described with reference to Figure 9. In this embodiment, a trained generative model and a patent database are stored inside the device. Specifically, a model storage area 212 and a patent database storage area 222 are provided in the storage unit 5, and the trained generative model and the patent database are stored in these areas, respectively.

[0064] The model storage area 212 stores trained generative models required for generating solutions and evaluating patentability, and the patent database storage area 222 stores patent databases required for searching patent documents. This makes it possible for the device to generate solutions and evaluate patentability independently, without relying on the first information processing device 10 and the second information processing device 20.

[0065] The solution generation device 1 in this embodiment can execute all the functions explained in the first embodiment. With this configuration, network connection is not required, which improves security and makes it possible to avoid processing delays caused by communication with the outside. Furthermore, since processing can be completed within the device, the availability and reliability of the system are improved.

[0066] <Aspect 1> The solution generation device 1 comprises a reception unit 30 that receives input information from a user, a prompt creation unit 40 that creates a prompt based on the input information received by the reception unit 30, a generation result acquisition unit 50 that inputs the prompt into a trained generation model 12 and acquires the generation result, a similar patent document information acquisition unit 60 that acquires information on similar patent documents based on the generation result, and a presentation unit 70 that presents the generation result and information on the similar patent documents to the user.

[0067] According to this aspect, a solution can be efficiently generated from the information input by the user and presented to the user together with related patent document information.

[0068] <Aspect 2> The solution generation device 1 according to aspect 1 is characterized in that, when a problem is received as input information, the prompt generation section 40 analyzes the problem and generates a prompt for generating a solution to the problem.

[0069] According to this aspect, it is possible to create a prompt for generating an appropriate solution to a problem input by a user.

[0070] <Aspect 3> The solution generation device 1 according to aspect 2 further comprises a searching section 55 that searches for and acquires solutions to problems from web information on the Internet.

[0071] According to this aspect, it is possible to search for a solution by utilizing the latest technical information on the Internet. Note that the solution generation device does not need to include a search unit.

[0072] <Aspect 4> The solution generation device 1 described in aspect 2 is characterized in that the prompt creation unit 40 further creates a patent search formula generation prompt that generates a patent search formula for obtaining information on similar patent documents based on the solution, and the similar patent document information acquisition unit 60 inputs the patent search formula generation prompt into the trained generation model 12 to obtain a patent search formula and inputs the patent search formula into the patent database 22 to obtain information on similar patent documents.

[0073] According to this aspect, an optimal patent search formula can be created by utilizing a generative model, and similar patent documents can be efficiently obtained.

[0074] <Aspect 5> The solution generation device 1 described in aspect 2 is characterized in that the prompt creation unit 40 further creates a search prompt for searching and acquiring information on similar patent documents based on the solution, and the similar patent document information acquisition unit 60 inputs the search prompt into the trained generation model 12 to acquire information on similar patent documents as the generation result.

[0075] According to this aspect, it is possible to obtain information on similar patent documents by directly using a generative model.

[0076] <Aspect 6> The solution generation device 1 described in aspect 2 is characterized in that the similar patent document information acquisition unit 60 acquires information on similar patent documents by performing a search using words that represent technical features extracted from the solution and groups of words that are semantically related to the words.

[0077] According to this aspect, it is possible to perform a patent search that takes into account semantic relevance based on the technical features of the solution.

[0078] <Aspect 7> The solution generation device 1 described in aspect 2 is characterized in that the prompt creation unit further creates an evaluation prompt to obtain an evaluation on the novelty and inventive step of the solution, and further includes a patentability evaluation presentation unit 170 that inputs the evaluation prompt into the trained generative model 12 to obtain an evaluation and presents the evaluation to the user.

[0079] According to this aspect, the patentability of the generated solution can be automatically evaluated and presented to the user. Note that the solution generation device does not necessarily have to include the patentability evaluation presentation unit.

[0080] <Aspect 8> A program for causing a computer to function as a receiving means 130 that receives input information from a user, a prompt creation means 140 that creates a prompt based on the input information received by the receiving means 130, a generation result acquisition means 150 that inputs the prompt into a trained generation model 12 and acquires the generation result, a similar patent document information acquisition means 160 that acquires information on similar patent documents based on the generation result, and a presentation means 180 that presents the generation result and information on similar patent documents to the user.

[0081] According to this aspect, the present invention can be embodied as a computer program.

[0082] <Aspect 9> A computer-readable recording medium having the program according to embodiment 8 recorded thereon.

[0083] According to this aspect, a program for implementing the present invention can be provided via a recording medium.

[0084] <Aspect 10> A solution generation device 1 is a computer comprising an arithmetic processing unit 2 and a memory device 3, wherein the memory device 3 has a model memory area 212 for storing a trained generative model 12 and a patent database memory area 222 for storing a patent document database 22, and the arithmetic processing unit 2 comprises a reception unit 30 for receiving input information from a user, a prompt creation unit 40 for creating a prompt based on the input information received by the reception unit 30, a generation result acquisition unit 50 for inputting the prompt into the trained generative model 12 to perform arithmetic processing and acquiring a generation result related to a solution, a similar patent document information acquisition unit 60 for acquiring information on similar patent documents from the patent database memory area 222 based on the generation result, and a presentation unit 70 for presenting the generation result and information on similar patent documents to the user. [Explanation of symbols]

[0085] 1 Solution generator 2 processors 3 RAM / ROM 4. User Interface 5 Storage section 6 Program storage area 7 Interface section 10 First information processing device 12 Pre-trained generative models 20 Second information processing device 22 Patent Database 30 Reception 40 Prompt Creation Section 50 Generation result acquisition part 55 Search Department 60 Similar patent document information acquisition department 70 Presentation section 130 Reception 140 Prompt Creation Section 150 Generation result acquisition unit 155 Search Department 160 Similar patent document information acquisition department 170 Patentability Evaluation Presentation Department 180 Presentation section 212 Model Storage 222 Patent Database Storage Area TXT1 input field TXT2 solution TXT3 Search Results BT1 button P1 prompt P2 prompt IMG1 screen IMG2 screen

Claims

1. a reception unit that receives input information from a user; a prompt creation unit that creates a prompt based on the input information received by the reception unit; a generation result acquisition unit that inputs the prompt into a trained generation model and acquires a generation result including a solution to the input information; a similar patent document information acquisition unit that extracts technical components from the content of the solution acquired by the generation result acquisition unit, searches a patent document database using words that represent the technical components and words related to the words, acquires information on related patent documents, evaluates their similarity to the solution, and ranks them based on the similarity; a presentation unit that presents the generated results and the ordered information of the patent documents to a user; A solution generation device comprising:

2. When the prompt creation unit receives a technical problem as the input information, the prompt creation unit analyzes the technical problem and creates the prompt for generating a solution to the technical problem. The solution generator of claim 1 .

3. The system further includes a search unit that searches for and acquires the solution to the technical problem from web information on the Internet. Solution generator according to claim 2.

4. The prompt creation unit further creates a patent search query creation prompt that creates a patent search query for acquiring information on similar patent documents based on the solution; The similar patent document information acquisition unit inputs the patent search formula generation prompt into the trained generation model to acquire the patent search formula, and inputs the patent search formula into a patent database to acquire information on the similar patent documents. Solution generator according to claim 2.

5. the prompt creation unit further creates a search prompt for searching and acquiring information on similar patent documents based on the solution; The similar patent document information acquisition unit inputs the search prompt into the trained generation model and acquires information on the similar patent documents as a generation result. Solution generator according to claim 2.

6. The solution generation device according to claim 2, characterized in that the similar patent document information acquisition unit acquires information on similar patent documents by performing a search using words expressing technical features extracted from the solutions and groups of words semantically related to the extracted words.

7. The prompt creation unit further creates an evaluation prompt for obtaining an evaluation of novelty and inventive step of the solution, and further includes a patentability evaluation presentation unit that inputs the evaluation prompt into the trained generative model to obtain the evaluation and presents the evaluation to the user. Solution generator according to claim 2.

8. a computer, a receiving means for receiving input information from a user, a prompt generating means for generating a prompt based on the input information received by the receiving means; a generation result obtaining means for inputting the prompt into a trained generation model and obtaining a generation result including a solution to the input information; a similar patent document information acquisition means for extracting technical components from the content of the solution acquired by the generation result acquisition means, searching a patent document database using words representing the technical components and a group of words related to the words, acquiring information on related patent documents, evaluating the similarity between the technical components and the solution, and ranking the related patent documents based on the similarity; a presentation means for presenting the generated results and the ordered information of the patent documents to a user; A program to function as a

9. A computer-readable recording medium on which the program according to claim 8 is recorded.

10. In a computer having a processing unit and a storage device, The storage device includes: a model storage area for storing a trained generative model; a patent database storage area for storing a patent document database; and The arithmetic processing device comprises: a reception unit that receives input information from a user; a prompt creation unit that creates a prompt based on the input information received by the reception unit; a generation result acquisition unit that inputs the prompt into the trained generation model to perform calculation processing and acquires a generation result including a solution for the input information; a similar patent document information acquisition unit that extracts technical components from the content of the solution acquired by the generation result acquisition unit, acquires information on related patent documents from the patent database storage area using words that represent the technical components and groups of words related to the words, evaluates the similarity between the technical components and the solution, and ranks the patent documents based on the similarity; a presentation unit that presents the generated results and the ordered information of the patent documents to a user; A solution generation device comprising:

Images