Ideation assistance system and ideation assistance method

Abstract

An ideation assistance system according to the present invention comprises a processor and a storage device. The storage device holds one or more ideas. The processor executes an idea divergence process for increasing the number of ideas held in the storage device and an idea convergence process for reducing the number of ideas which have been increased in the idea divergence process. In the idea convergence process, the processor calculates, for each of a plurality of ideas, a degree of importance, which decreases as a possibility of replacement with another idea included in the plurality of ideas increases, and selects an idea having a low degree of importance as an idea to be deleted so as to reduce the number of ideas. For ideas which have not been selected as ideas to be deleted among the plurality of ideas held in the storage device, the processor determines, on the basis of the degree of mutual similarity, whether the ideas have been converged. When it is determined that the ideas have been converged, the processor executes the idea divergence process again.

Description

Idea support system and idea support method Incorporation by Reference

[0001] This application claims priority from Japanese Patent Application No. 2023-169394, filed on September 29, 2023, the contents of which are incorporated herein by reference.

[0002] The present invention relates to a technology for supporting ideas.

[0003] People engaged in research and development and planning are often required to come up with new ideas that differ from existing ones. In such cases, they often deepen their own thinking and gather related information in order to come up with new ideas. One method for expanding ideas is called brainstorming, in which multiple people come up with ideas one after another and then expand on them. Since it is difficult for one person to expand the scope of ideas alone, this method uses the ideas of others as a springboard to expand ideas. Another method known as the KJ method is known, in which each person participating in a discussion writes down their ideas, and then organizes them by rearranging them to form a single idea.

[0004] These methods often require multiple people to cooperate, based on the premise that it is difficult to expand ideas on one person. However, this has the problem of being inefficient, as it is difficult to find time for multiple people. Also, if there are differences in prior knowledge or experience among participants, it can be difficult to hold a good discussion. Therefore, idea support systems have been developed to promote the spontaneous expansion of ideas, such as by presenting information randomly, so that even one person can expand their ideas.

[0005] For example, Japanese Patent Application Laid-Open No. 2007-149087 (Patent Document 1) discloses a method for creating sentences that express ideas more clearly by diagramming a draft that expresses the user's ideas and presenting the relationships between the components in an easy-to-understand manner.

[0006] Furthermore, Japanese Patent Application Laid-Open Publication No. 2013-80315 (Patent Document 2) discloses a mechanism that allows users to deliberately edit drafts that express ideas that they have gained through discussions with others, and suggests that the user discuss the ideas again with others, thereby encouraging repeated discussions and improvements to the ideas.

[0007] Patent Document 1: JP 2007-149087 A Patent Document 2: JP 2013-80315 A

[0008] The invention described in Patent Document 1 can encourage people to notice and come up with new ideas by receiving their ideas as sentences and presenting the results of analyzing them, as well as presenting related words, etc. However, unless the information presented is new to the user, it will not be possible to expand the user's ideas.

[0009] The invention described in Patent Document 2 uses a system to assist people in the process of organizing ideas they have gained through discussion, while at the same time encouraging them to start the discussion again, thereby contributing to the creation of more refined ideas.

[0010] However, in both inventions, the process of generating ideas itself is carried out outside the scope of the system, and the need for discussion with others to refine the ideas remains unchanged. In order to make it easier to come up with new ideas, it is necessary to provide idea support that is complete within the idea support system.

[0011] The present invention was made with this in mind. The present invention comprehensively supports the process of expanding a user's ideas and refining them. To achieve this, the system assumes that ideas are generated through a cycle of divergence (expanding ideas) and convergence (refining ideas), and switches the functions provided depending on the phase. This allows the process of expanding ideas to be comprehensively carried out by considering the system of the present invention.

[0012] In order to solve at least one of the above problems, the present invention provides a method for solving a problem comprising a processor and a storage device, wherein the storage device holds one or more ideas, and the processor executes an idea divergence process for increasing the number of ideas held in the storage device, and an idea convergence process for reducing the number of ideas after the increase by the idea divergence process, wherein in the idea convergence process, the processor calculates an importance for each of a plurality of ideas, the importance of which becomes lower the more likely the idea is to be substituted by other ideas included in the plurality of ideas, and reduces the number of ideas by selecting ideas with lower importance as candidates for deletion, and determines whether the ideas have converged based on their mutual similarity for ideas among the plurality of ideas held in the storage device that have not been selected as candidates for deletion, and if it is determined that the ideas have converged, executes the idea divergence process again.

[0013] According to one aspect of the present invention, when a person has an initial idea and tries to develop it, they can refine it into a higher quality idea by repeatedly expanding and summarizing the idea based on that initial idea.

[0014] Other problems, configurations and effects will become apparent from the following description of the embodiments.

[0015] FIG. 1 is a block diagram showing an example of the configuration of an idea support system according to a first embodiment of the present invention. FIG. 2 is a block diagram showing an example of the configuration of a physical implementation of the idea support system according to the first embodiment of the present invention. FIG. 3 is a flowchart showing an example of an operation procedure performed by the idea support system according to the first embodiment of the present invention. FIG. 4 is a state transition diagram showing an example of an operation procedure performed by the idea support system according to the first embodiment of the present invention. FIG. 5 is a flowchart showing an example of details of an initial idea estimation process performed by the idea support system according to the first embodiment of the present invention. FIG. 6 is an explanatory diagram showing an example of data of a category determination decision tree and a category-specific keyword list referenced by the idea support system according to the first embodiment of the present invention. FIG. 7 is an explanatory diagram showing an example of a category determination decision tree referenced by the idea support system according to the first embodiment of the present invention. FIG. 8 is an explanatory diagram showing an example of items in a user idea table held by the idea support system according to the first embodiment of the present invention. FIG. 9 is an explanatory diagram showing an example of an initial idea input screen presented to a user in an initial idea estimation process by the idea support system according to the first embodiment of the present invention. FIG. 10 is a flowchart showing an example of details of idea divergence process performed by the idea support system according to the first embodiment of the present invention. FIG. 11 is an explanatory diagram showing an example of an idea divergence screen presented to a user in an idea divergence process by the idea support system according to the first embodiment of the present invention. FIG. 12 is an explanatory diagram showing an example of idea literature data included in an idea literature database held by the idea support system according to the first embodiment of the present invention. FIG. 1 is an explanatory diagram showing an example of a screen presented to a user when idea proposals are generated in the idea divergence processing by the idea support system of Example 1 of the present invention. FIG. 2 is a flowchart showing an example of details of the divergence determination processing by the idea support system of Example 1 of the present invention. FIG. 3 is a flowchart showing an example of details of the idea convergence processing by the idea support system of Example 1 of the present invention. FIG. 4 is an explanatory diagram showing an example of a screen presented by the idea support system of Example 1 of the present invention when a user idea is selected in the divergence convergence processing. FIG. 5 is an explanatory diagram showing an example of a screen presented by the idea support system of Example 1 of the present invention when a user idea is deleted in the divergence convergence processing. FIG. 6 is a flowchart showing an example of details of the importance calculation processing by the idea support system of Example 1 of the present invention. FIG. 7 is a flowchart showing an example of details of the convergence determination processing by the idea support system of Example 1 of the present invention.1 is a flowchart showing an example of details of a process for determining the end of idea convergence processing by the idea support system of Example 1 of the present invention. FIG. 2 is a flowchart showing an example of details of an idea result summarizing process by the idea support system of Example 1 of the present invention. FIG. 3 is an explanatory diagram showing an example of a screen presented to a user in idea result summarizing processing by the idea support system of Example 1 of the present invention. FIG. 4 is a block diagram showing an example of a configuration of an idea support system of Example 2 of the present invention. FIG. 5 is an explanatory diagram showing an example of an initial idea input screen presented to a user in initial idea estimation processing by the idea support system of Example 2 of the present invention. FIG. 6 is an explanatory diagram showing an example of a termination condition table held by the idea support system of Example 2 of the present invention. FIG. 7 is a flowchart showing an example of details of a process for determining the end of idea convergence processing by the idea support system of Example 2 of the present invention. FIG. 8 is a block diagram showing an example of a configuration of an idea support system of Example 3 of the present invention. FIG. 9 is an explanatory diagram showing an example of an idea divergence screen 2800 presented to a user in idea divergence processing by the idea support system of Example 3 of the present invention.

[0016] FIG. 1 is a block diagram showing an example of the configuration of an idea support system according to a first embodiment of the present invention.

[0017] The idea support system 100 of this embodiment is a system having a user idea receiving unit 101 that receives user input, an idea divergence support unit 102 that presents information that encourages the user to diverge their ideas, an idea convergence support unit 103 that presents information that encourages the user to diverge their ideas, an idea result presentation unit 104 that supports the user in consolidating generated ideas and reaching a conclusion, a user idea candidate database 105 that saves information that expresses ideas generated by the user, a keyword extraction unit 106 that extracts related keywords from the information that expresses an idea, an idea similarity evaluation unit 107 that has the function of comparing two pieces of information that express an idea and evaluating their similarity, an information generation unit 108 that has the function of generating information about ideas, a related information search unit 109 that has the function of searching external data based on given information, and an idea literature database 110 that records existing idea information.

[0018] FIG. 2 is a block diagram showing an example of a configuration of a physical implementation of the idea generation support system according to the first embodiment of the present invention.

[0019] The idea generation support system 100 can be implemented using a general-purpose computer. Specifically, the idea generation support system 100 includes a processor 201 with high computing performance, a DRAM 202, which is a volatile temporary storage area capable of high-speed read / write, a storage device 203, which is a permanent storage area using a hard disk drive (HDD) or flash memory, an input device 204, such as a mouse or keyboard, for operation, a monitor 205, which displays operations to the user, and a communication interface 206, such as a serial port, for communicating with the outside world. The user idea reception unit 101, idea divergence support unit 102, idea convergence support unit 103, idea result presentation unit 104, keyword extraction unit 106, idea similarity evaluation unit 107, information generation unit 108, and related information search unit 109 can be realized by the processor 201 executing programs stored in the storage device 203. The user idea candidate database 105 and idea literature database 110 can be implemented by the processor 201 executing programs that store data in the storage device 213.

[0020] FIG. 3 is a flowchart showing an example of an operation procedure of the idea support system according to the first embodiment of the present invention.

[0021] 3 shows the interactions between the user 300 and the idea support system 100 and the operations within each element. When the user 300 starts the idea support system 100, the idea support system 100 executes an initial idea estimation process (step S301) to identify the user's initial idea. When the user 300 inputs an initial draft in response, the idea support system 100 executes an idea divergence process (step S302) to encourage the user 300 to come up with a new idea.

[0022] If the user 300 can expand his / her ideas based on this, the idea generation support system 100 executes an idea convergence process (step S303) to encourage the user 300 to summarize his / her ideas. If the user 300 summarizes his / her ideas based on this, the idea generation support system 100 executes an idea divergence process (step S302) again.

[0023] This procedure is repeated, and when the user finally comes up with an idea, the idea generation support system 100 executes an idea result summary process (step S304) to finalize the result of the user's idea as a summary of the study results. Details of the respective processing procedures of the initial idea estimation process (step S301), idea divergence process (step S302), idea convergence process (step S303), and idea result summary process (step S304) will be described later.

[0024] FIG. 4 is a state transition diagram showing an example of an operation procedure of the idea support system according to the first embodiment of the present invention.

[0025] Specifically, Figure 4 shows a state transition diagram relating to the processing procedure shown in Figure 3. When the idea generation support system 100 is started, it executes an initial idea estimation process (step S301), and then executes an idea divergence process (step S302). This is expected to broaden the user's ideas. Next, a divergence convergence process (step S303) is executed, and the user's ideas are consolidated to a certain extent. At this time, the idea generation support system 100 determines the progress of the user's idea generation, and if it determines that the progress of the idea generation is insufficient, it executes the idea divergence process (step S302) and the divergence convergence process (step S303) again. If it determines that the progress of the idea generation is sufficient, it proceeds to an idea generation result summary process (step S304), and then ends the operation.

[0026] FIG. 5 is a flowchart showing an example of details of the initial idea estimation process (step S301) performed by the idea generation support system according to the first embodiment of the present invention.

[0027] In the idea support system 100, the user idea reception unit 101 first receives an instruction from the user to start consideration (step S501).Then, the keyword extraction unit 106 acquires a category determination decision tree stored in the idea literature database 110, and generates a keyword list 506 by asking the user questions based on the tree (step S502).

[0028] FIG. 6 is an explanatory diagram showing an example of data of a category determination decision tree 610 and a category-specific keyword list 620 referred to by the idea generation support system according to the first embodiment of the present invention.

[0029] The category determination decision tree 610 is data for asking questions to the user, and is information in which multiple nodes (branches) are connected in a tree shape. The data for one record in the category determination decision tree 610 corresponds to one node, and has a node ID 611 to identify the node, a type 612 indicating the type of node, content 613 recording the content such as words that the node means, and a right node ID 614 and a left node ID 615 indicating the branch destination.

[0030] Here, a node that does not have a branch destination is called a leaf, and whether a node is a leaf or a branch can be determined by its type 612. If the node is a leaf, its stored content 613 is a category ID that indicates the category of the document. The category keyword list 620 holds data corresponding to this category ID, and includes the above-mentioned category ID 621 and the corresponding keyword list 622.

[0031] FIG. 7 is an explanatory diagram showing an example of a category determination decision tree 610 to be referred to by the idea generation support system according to the first embodiment of the present invention.

[0032] A method for creating the keyword list 506 using the category determination decision tree 610 and the category-specific keyword list 620 will be described with reference to Figure 7. As mentioned above, the category determination decision tree 610 has a structure in which nodes are connected in a tree shape, and Figure 7 shows this tree structure. In the figure, rounded rectangles indicate nodes that represent branches, and circles indicate leaf nodes.

[0033] In the figure, a node 701 labeled "database" is placed at the top, which means that the first question to be asked of user 300 is "Is the word "database" related to the idea of ​​user 300?" The node corresponding to the first question can be identified by selecting a record having a predetermined node ID 611 from category determination decision tree 610, and can be simply implemented by treating the node with the smallest node ID value, for example, the numerical value 1, as the top node.

[0034] If the answer is "irrelevant" after the user idea receiving unit 101 asks the user 300 the above question, the user idea receiving unit 101 refers to the left node ID 615 to identify the next node. The node 702 in the figure with the word "aggregation" written on it indicates this node. On the other hand, if the answer is "related," the user idea receiving unit 101 refers to the right node ID 614 to identify the next node. This corresponds to the node 703 in the figure with the word "search."

[0035] Similarly, by asking questions such as "Is the word "tally" related to the user's idea?" based on the word at the next node, it is possible to go down this tree structure. Eventually, a node with no branches, such as node 704 marked "H" in the figure, is reached. This "H" is the content of content 613 and corresponds to category ID 621. Therefore, by referring to category-specific keyword list 620, keyword list 622 can be obtained, and this becomes keyword list 506 presented to user 300.

[0036] Such a data structure is a well-known data structure called a decision tree, and can be automatically constructed by well-known decision tree machine learning processing. Specifically, if literature related to an idea is collected in advance, its category and a list of words appearing in the literature are extracted, and a well-known decision tree learning algorithm is applied to perform a discriminant analysis using the presence or absence of a word as an explanatory variable and the category as a target variable, the result will be data for the category determination decision tree 610. Furthermore, if literature is aggregated by category and a list of the words appearing in the literature is created, a category-specific keyword list 620 can also be easily created.

[0037] Returning to Figure 5, the following explanation will be given. The user idea receiving unit 101 of the idea support system 100 requests the user 300 to specify keywords related to the current user idea based on the keyword list 506 (step S503), and the information generating unit 108 generates a user idea proposal based on the specified keywords and a known generation AI (step S504). The user idea receiving unit 101 presents this user idea proposal to the user 300 and checks its validity (step S505). Once the validity is confirmed, the result is stored in the user idea table 800.

[0038] FIG. 8 is an explanatory diagram showing an example of items contained in a user idea table 800 held by the idea generation support system according to the first embodiment of the present invention.

[0039] The user idea table 800 includes an idea ID (01) that can uniquely identify an idea, a generation time 802 that indicates the time when the idea was generated, idea expression data 803 that indicates the generated user idea proposal, and a deletion flag 804 that indicates whether the data has been deleted. When the validity of the idea can be confirmed (step S505), a unique number is issued as appropriate, such as a consecutive number, as the idea ID 801, the time when this process was executed is recorded as the generation time 802, the generated idea is stored in the idea expression data 803, and the deletion flag 804 is set to False, thereby creating a record.

[0040] FIG. 9 is an explanatory diagram showing an example of an initial idea input screen 901 presented to the user 300 in the initial idea estimation process by the idea support system according to the first embodiment of the present invention.

[0041] On the initial idea input screen 901, a keyword list 506 is presented in an upper portion 902 of the screen, allowing the user 300 to select words that correspond to their own thoughts. In the process leading up to creating the keyword list 506 (step S502), keywords are presented here one by one, but after the keyword list 506 is created, the keywords are listed here collectively, and a button 903 for generating a draft is also displayed. By pressing this button, a user idea proposal based on the words is generated and presented in a lower portion 904 of the screen. Once the user 300 has confirmed the validity of the proposed proposal, they press a button 905, completing the validation check (step S505).

[0042] While an example of generating a draft sentence in step S504 is described here, anything can be generated based on keywords and expresses an idea. Typically, a known deep learning model for generating text data (such as GPT (Generative Pretrained Transformer)) can be used, and images and sounds can also be added (see FIG. 24 in Example 2 below for an example of adding images). Alternatively, if mathematical formulas or chemical formulas can be generated, adding them can more accurately express the idea.

[0043] As an example of generating sentences using GPT, GPT is capable of generating candidate strings that follow an input string, so if an input such as "■Keywords: idea generation support, artificial intelligence, database, similar search →When using artificial intelligence to support idea generation, a database is referenced to check whether similar ideas exist. ■Keywords: temperature, total, average →" is given, a sentence related to "temperature, total, average" will be generated next, making it easy to generate sentences based on keywords. By replacing the "temperature, total, average" part of this example with a string of words selected from keyword list 506, a sentence corresponding to any word can be generated.

[0044] Next, the idea divergence process (step S302) will be described with reference to FIG.

[0045] FIG. 10 is a flowchart showing an example of details of the idea divergence process (step S302) performed by the idea generation support system according to the first embodiment of the present invention.

[0046] The idea generation support unit 102 of the idea generation support system 100 first displays the user idea data in the user idea candidate database 105 and prompts the user 300 to select one (step S1001).

[0047] FIG. 11 is an explanatory diagram showing an example of an idea divergence screen 1101 presented to the user 300 in the idea divergence process by the idea support system according to the first embodiment of the present invention.

[0048] The idea divergence screen 1101 lists the user ideas 800 stored in the user idea candidate database 105 whose deletion flag 804 is set to False. The user 300 can select one of these ideas and use it as a starting point for expanding on the idea.

[0049] 10, when one user idea 800 is selected, the keyword extractor 106 of the idea generation support system 100 searches the idea literature database 110 for literature similar to the selected idea, and generates a keyword list related to the literature (step S1002).

[0050] FIG. 12 is an explanatory diagram showing an example of idea literature data 1200 included in the idea literature database 110 held by the idea generation support system according to the first embodiment of the present invention.

[0051] The idea literature data 1200 is data stored in advance on past literature that can be used as reference for ideas, and includes a literature ID 1201 that is an identifier for the literature, keywords 1202 related to the literature, idea expression data 1203 that expresses the content of the literature, and multivariate values ​​1204 that are the idea expression data converted into multivariates.

[0052] Of these, the multivariate values ​​1204 are set so that the smaller the Euclidean distance (the square root of the sum of the squares of the differences between the components) between the multivariate values, the more similar the idea expression data 1203. In the case of text, such multivariate values ​​can be generated by inputting the data into a well-known embedding method, for example, a deep learning model called Bert (Bidirectional Encoder Representations from Transformers), and can also be generated in a similar manner for other representations, such as images.

[0053] In generating a keyword list (step S1002), the keyword extraction unit 106 converts the selected user idea candidates 800 into multivariates in the same manner as described above, calculates the distance between this multivariate and the multivariate value 1204 of each document included in the idea document data 1200, selects a predetermined number of data items, for example 10 items, in order of smallest distance, and creates a keyword list by collecting keywords (1202).

[0054] Next, the idea generation support unit 102 requests the user 300 to specify a keyword based on this keyword list (step S1003), and the information generation unit 108 generates a user idea proposal (step S1004).

[0055] FIG. 13 is an explanatory diagram showing an example of a screen presented to the user 300 when the idea generation support system according to the first embodiment of the present invention generates ideas in the idea divergence process.

[0056] Buttons related to keywords in the keyword list are arranged in the center of the screen 1301, and the user 300 can select one of these to use or not use a keyword. Once the selection is complete, the user 300 can press a draft generation button 1302 to generate a draft idea (e.g., a draft sentence) in a text box 1303.

[0057] The user 300 checks whether the generated draft idea is appropriate (step S1005). Specifically, if the user 300 determines that the draft idea is inappropriate, the user 300 presses the draft generation button 1302 again. At this time, the user 300 may change the keyword selection in the center 1301 of the screen. Each time the draft generation button 1302 is pressed, a draft is regenerated (step S1004). On the other hand, if the user 300 is satisfied with the draft idea, the user presses the OK button 1304. The idea generation support unit 102 checks which button was pressed (step S1005), and if the OK button 1304 was pressed, the generated draft idea is saved in the user idea table 800 of the user idea candidate database 105 (step S1006).

[0058] In step S1004, similarly to step S504 described above, what is generated does not have to be a draft, as long as it can be generated based on keywords and expresses an idea.

[0059] As an example of generating a proposal using GPT, if an input such as "■ Original text: When using artificial intelligence to support idea generation, a database is referenced to check for similar ideas. ■ Keywords: Distributed database ■ Revision: When using artificial intelligence to support idea generation, a database is referenced to check for similar ideas, and using a distributed database speeds up the similarity search. ■ Original text: When compiling temperature data, the average temperature is calculated in advance, sent to the thermometer, and only the difference is sent, thereby reducing the amount of communication data. ■ Keywords: Forecast ■ Revision:" is given, and a proposal combining the "original text" and keywords can be expected to be generated, such as "When compiling temperature data, the average temperature and a forecast value based on that are calculated in advance, sent to the thermometer, and only the difference is sent, thereby reducing the amount of communication data." By replacing the original text and keywords in this example with user idea candidate 800 and a word string from the keyword list, ideas can be expanded to incorporate the selected keywords.

[0060] Finally, the idea divergence support unit 102 determines whether the user idea candidate database 105 has diverged (step S1007).

[0061] FIG. 14 is a flowchart showing an example of details of the divergence determination process (step S1007) performed by the idea generation support system according to the first embodiment of the present invention.

[0062] First, the idea divergence support unit 102 obtains the idea expression data 803 of records whose deletion flags 804 are set to False from the user idea table 800 (step S1401) and sends it to the idea similarity evaluation unit 107. The idea similarity evaluation unit 107 converts the received data into multivariates (step S1402). Next, the idea similarity evaluation unit 107 calculates the distance between the obtained multivariates (step S1403). This makes it possible to obtain the similarity between the idea expression data 803 of the user ideas 800.

[0063] The idea divergence support unit 102 evaluates whether the divergence is sufficient by using this feature. As an example, the variance of the Euclidean distance between the multivariates obtained in step S1402 is calculated, and whether the divergence is sufficient is determined based on whether the variance exceeds a predetermined threshold (step S1404).

[0064] Alternatively, if divergence of a predetermined number of ideas is expected, a predetermined number of clusters may be constructed based on the Euclidean distance between multivariates, and the maximum distance within a cluster may be determined to be large. This method has the advantage of eliminating the need to continually determine that divergence exists when, for example, two widely separated ideas remain. Furthermore, the distance can also be determined by the similarity between the idea expression data 803, such as by using the well-known cosine similarity.

[0065] Next, the idea convergence process (S303) will be described with reference to FIG.

[0066] FIG. 15 is a flowchart showing an example of details of the idea convergence process (step S303) performed by the idea generation support system according to the first embodiment of the present invention.

[0067] First, the idea convergence support unit 103 acquires, from among the user ideas 800 stored in the user idea candidate database 105, those for which the deletion flag 804 is False, and presents a list to the user 300 to select two (step S1501). Specifically, the idea expression data 803 of records for which the deletion flag 804 is False is displayed as a list, and two of them are selected. The idea convergence support unit 103 passes the two selected user idea candidates to the information generation unit 108. The information generation unit 108 generates a new user idea candidate by combining the two user idea candidates acquired from the idea convergence support unit 103, and presents the new user idea candidate to the user 300 (step S1502).

[0068] FIG. 16 is an explanatory diagram showing an example of a screen presented when a user selects an idea in the divergence-convergence process by the idea generation support system according to the first embodiment of the present invention.

[0069] A list of user ideas 800 is displayed in the upper part of the screen 1601, and a selected user idea 800 is surrounded by a bold frame. By pressing a proposal generation button 1603, a user idea proposal is created by combining these two ideas, and is displayed as a generated proposal in the lower part of the screen 1602. This is performed by the idea convergence support unit 103 to confirm the validity of the generated proposal with the user 300 (step S1503). If the validity of the generated proposal is confirmed (for example, if the user presses the new registration button 1604), the idea convergence support unit 103 newly registers the generated proposal in the user idea candidate database 105 (step S1504).

[0070] Here, as an example of sentence generation using GPT, if an input such as "■ Original text 1: When using artificial intelligence to support idea generation, a database is consulted to check whether similar ideas exist.■ Original text 2: When using artificial intelligence to support idea generation, new ideas are generated by combining information from multiple documents.■ Result: When using artificial intelligence to support idea generation, new ideas are generated by combining information from multiple documents, and a database is consulted to check whether similar ideas exist.■ Original text 1: When using a human sensor to count the number of people entering a building, false positives are reduced by comparing it with security camera footage taken near the entrance at the same time.■ Original text 2: Security camera footage taken near the building entrance is corrected to take into account changes in illuminance due to time of day.■ Result:" is given, it is expected that a sentence combining multiple sentences will be generated, such as "When using a human sensor to count the number of people entering a building, people are detected after correction is made to take into account changes in illuminance due to time of day, and false positives are reduced by comparing it with security camera footage taken near the entrance at the same time." If the original text in this example is replaced with the user idea 800 selected in step S1501, the idea can be expanded to incorporate the selected keyword.

[0071] Continuing the explanation, returning to Fig. 15, the idea convergence support unit 103 then calculates the importance of each of the user ideas 800, and proposes deletion of those whose importance is lower than a predetermined threshold (step S1505).

[0072] Here, the importance may be calculated, for example, based on the possibility that an idea included in the user idea 800 can be substituted with another idea. For example, if the user idea 800 includes two ideas that are completely (or almost) identical, it can be said that there is almost no problem in deleting one of them. Similarly, the higher the similarity between the two ideas, the higher the possibility that one of them can be substituted with the other (i.e., the less problems will occur if one of them is deleted). For this reason, the importance is calculated so that, for example, the higher the possibility that an idea can be substituted, the lower the importance. Details of the importance calculation will be described later (see step S1806 in FIG. 18, etc.).

[0073] FIG. 17 is an explanatory diagram showing an example of a screen presented when a user idea is deleted in the divergence-convergence process by the idea generation support system according to the first embodiment of the present invention.

[0074] A list of user ideas 800 is displayed in the upper part of the screen 1701, and ideas with an importance lower than a predetermined value are displayed with a dashed frame. Considering that the importance calculation is not always optimal, the system waits for the user to press button 1702 to confirm before deleting an idea, but if the importance calculation is sufficiently reliable, the idea may be deleted automatically. The idea convergence support unit 103 changes the deletion flag 804 to True for the user idea 800 that is the target of deletion.

[0075] FIG. 18 is a flowchart showing an example of details of the importance calculation process (step S1505) performed by the idea generation support system according to the first embodiment of the present invention.

[0076] First, the idea convergence support unit 103 acquires data for which the deletion flag 804 is set to False from the user idea table 800 (step S1801). Next, the idea similarity evaluation unit 107 converts each of the data acquired in step S1801 into multivariates (step S1802) and calculates the similarity between the data (step S1803). Here, the similarity may be, for example, Euclidean distance or cosine similarity between the multivariates.

[0077] Next, the idea convergence support unit 103 generates clusters using a known clustering method based on the distances between the multivariates (S1804). Examples of clustering methods that can be used include k-means and hierarchical clustering. The idea convergence support unit 103 then assigns a level of importance to the user idea 800 using the similarity and the clustering results.

[0078] Here, the importance may be calculated by, for example, first assigning a predetermined value (step S1805) and then subtracting a value based on a condition (step S1806). Examples of the condition may reflect the number of other ideas in the same cluster being greater than a predetermined value, the similarity with other nearby ideas being below a predetermined threshold, or the creation time being older than a predetermined value.

[0079] The importance of the ideas used in the idea combination (step S1502) (i.e., the two ideas selected in step S1501) may be reduced, making the ideas used in the idea combination more likely to be deleted.

[0080] Continuing the explanation, returning to Fig. 15, the idea convergence support unit 103 next determines whether the user's ideas are converged in order to determine whether the idea should be diverged (step S1506).

[0081] FIG. 19 is a flowchart showing an example of details of the convergence determination process (step S1506) performed by the idea generation support system according to the first embodiment of the present invention.

[0082] In determining convergence, first, the idea convergence support unit 103 obtains data for which the deletion flag 804 is set to False from the user idea table 800 (step S1901) and sends it to the idea similarity evaluation unit 107. The idea similarity evaluation unit 107 converts the sent data into multivariates (step S1902). Next, the idea similarity evaluation unit 107 calculates the distance between the obtained multivariates (step S1903). This makes it possible to obtain the similarity between idea drafts created by users.

[0083] The idea convergence support unit 103 evaluates whether convergence is sufficient using the multivariates calculated in step S1903. As an example, the unit 103 calculates the variance of the Euclidean distances between the multivariates and determines whether convergence is sufficient based on whether the variance is below a predetermined threshold (step S1904).

[0084] Alternatively, if convergence to a predetermined number of ideas is expected, the predetermined number of clusters can be constructed based on the Euclidean distance between the multivariates, and convergence can be determined by whether the maximum value of the distance within the same cluster is sufficiently small. With this method, convergence can be determined even in a situation where convergence has occurred to two very different ideas.

[0085] The description will continue with returning to Fig. 15. Next, the idea convergence support unit 103 determines whether the entire process is complete (step S1507).

[0086] FIG. 20 is a flowchart showing an example of details of the process of determining the end of the idea convergence process (step S1507) performed by the idea generation support system according to the first embodiment of the present invention.

[0087] First, the idea convergence support unit 103 acquires all user ideas 800 from the user idea candidate database 105 (step S2001). Here, the acquisition is performed regardless of the value of the deletion flag 804, so deleted data is also acquired. Next, the idea convergence support unit 103 sorts the acquired user ideas 800 in order of creation time 802 (step S2002). At this time, if the number of user ideas 800 does not exceed a predetermined number, the idea convergence support unit 103 determines that the fluctuation is small (step S2003). In this example, the threshold value is set to 10.

[0088] Next, the idea similarity evaluation unit 107 converts the acquired user ideas 800 into multivariates (step S2004). Next, the idea convergence support unit 103 calculates the similarities between the user idea 800 with the oldest creation time 802, i.e., the first created user idea 800, and the three user ideas 800 with the most recent creation times 802, and calculates the average (step S2005). If this average does not exceed the threshold, this means that there has not been much change between the initial user idea and the later (i.e., most recent) user ideas, and the idea convergence support unit 103 determines that there is little variation (step S2006).

[0089] On the other hand, if the average calculated in step S2005 exceeds the threshold, this means that the most recent user ideas have changed to a certain extent from the initial user ideas. In this case, the idea convergence support unit 103 next calculates the similarity between the three user ideas 800 with the most recent creation times 802 and the three user ideas with the oldest creation times 802, and calculates the average (step S2007). The idea convergence support unit 103 then determines whether the average exceeds a predetermined threshold (step S2008).

[0090] If the average calculated in step S2007 does not exceed the threshold, this means that the recently generated user ideas have not changed much, and the idea convergence support unit 103 determines that the variation is small (step S2008). On the other hand, if the average calculated in step S2007 exceeds the threshold, this means that there has been some degree of change among the recently generated user ideas. In this case, the idea convergence support unit 103 determines that the variation is large (step S2008).

[0091] In the above steps S2005 and S2007, "three" is just an example, and other numbers may be used.

[0092] 15, if it is determined that the fluctuations in the user ideas are large, it is considered that there is a possibility that the user ideas will fluctuate further (in other words, there is a possibility that new user ideas will be obtained), so the idea divergence process (FIG. 10, etc.) is executed again, and then the idea convergence process (FIG. 15, etc.) is executed again. On the other hand, if it is determined in step S1507 that the fluctuations in the user ideas are small, it is considered that there is little possibility that new user ideas will be obtained even if the idea divergence process and idea convergence process are repeated any further, so next, the idea result summary process (step S304) is executed.

[0093] FIG. 21 is a flowchart showing an example of details of the idea generation result summarizing process (step S304) performed by the idea generation support system according to the first embodiment of the present invention.

[0094] First, the idea result presentation unit 104 extracts keywords from the text of the idea expression data 803 for which the deletion flag 804 in the user idea table 800 is False (step S2101). Next, the related information search unit 109 creates a related article list by submitting the keywords to an external service that searches for Web articles via the Internet 111 and obtaining search results (step S2102). Finally, the idea result presentation unit 104 adds the created related article list to the text of the idea expression data 803 and presents it as a final idea result (step S2103).

[0095] FIG. 22 is an explanatory diagram illustrating an example of a screen presented to the user 300 in the idea generation support system according to the first embodiment of the present invention during the idea generation result summarization process.

[0096] The final idea results are displayed in the upper part of the screen 2201, and a function to download them in a specified format is provided. In addition, the contents of a list of related articles are displayed in the lower part of the screen 2202, and existing cases, etc. can be checked for each idea.

[0097] According to the above-described first embodiment, when a person has an initial idea and tries to develop it, the person can refine the idea into a higher quality idea by repeatedly expanding and summarizing the idea based on the initial idea.

[0098] Next, a second embodiment of the present invention will be described. The configuration and effects of the system according to the second embodiment are the same as those of the system according to the first embodiment shown in Figs. 1 to 22 except for the differences described below, and therefore, the description thereof will be omitted.

[0099] FIG. 23 is a block diagram showing an example of the configuration of an idea support system according to the second embodiment of the present invention.

[0100] The second embodiment differs from the first embodiment in that an end condition for the idea study is accepted in the initial idea estimation process (step S301) and the operation is ended by referring to end condition data 2301.

[0101] The physical implementation of the idea support system 100 of the second embodiment is the same as that of the first embodiment (see FIG. 2). The termination condition data 2301 can be implemented by the processor 201 executing a program that stores data in the storage device 213.

[0102] FIG. 24 is an explanatory diagram showing an example of an initial idea input screen 2400 that is presented to the user 300 in the initial idea estimation process by the idea support system according to the second embodiment of the present invention.

[0103] The initial idea input screen 2400 of the second embodiment has the same functions as the initial idea input screen 901 of the first embodiment, and also has a function for inputting the purpose of the initial idea. The purpose of the idea can be input by selecting an option in the upper part 2401 of the screen. This purpose of the idea is used by checking it against the termination conditions in the termination condition data 2301. The termination condition table 2501 included in this termination condition data 2301 will be described with reference to FIG. 25.

[0104] FIG. 25 is an explanatory diagram illustrating an example of a termination condition table 2501 held by the idea generation support system according to the second embodiment of the present invention.

[0105] The termination condition table 2501 stores idea objectives 2502 to be displayed on the initial idea input screen 2400, and corresponding termination conditions 2503. The termination conditions store mathematical expressions corresponding to the conditions and parameter judgments of the judgment process (steps S2001 to S2008) in the first embodiment shown in Fig. 20, and the termination can be judged by calculating the mathematical expressions. By setting the mathematical expressions in advance according to the objective and preparing multiple patterns, it becomes possible to make judgments according to the objective.

[0106] Here, examples of the purpose of idea generation include, but are not limited to, "improving an existing idea" and "applying an existing idea to another field" in addition to "creating a completely new idea" as shown in FIG. 24 . Furthermore, for example, in the termination condition when the purpose is "improving an existing idea," the threshold for the similarity of ideas may be set lower than when the purpose is "creating a completely new idea." However, this threshold setting is also one example and is not limited to this.

[0107] The termination condition can also be implemented by determining whether the same idea is repeatedly looped based on the degree of similarity, thereby avoiding repetitive fruitless investigations. Alternatively, a condition can be implemented in which the multivariate stored in the idea literature database 110 is compared with the multivariate of the user idea 800, and an unexamined area is detected based on whether any information stored in the idea literature database 110 that is similar to the user idea 800, including deleted information, exists, and the presence or absence of such information is reflected in the termination condition.

[0108] FIG. 26 is a flowchart showing an example of details of the process of determining the end of the idea convergence process (step S1507) performed by the idea generation support system according to the second embodiment of the present invention.

[0109] 20 is that, instead of steps S2005 to S2008, the idea convergence support unit 103 performs a process of calculating an indicator of termination by referring to the termination condition data 2301 (step S2601) and a process of determining whether fluctuations are large based on whether the indicator of termination exceeds a threshold (step S2602). This makes it possible to determine whether to continue or terminate the consideration of ideas depending on the purpose.

[0110] Next, a third embodiment of the present invention will be described. The configuration and effects of the system according to the third embodiment are the same as those of the systems according to the first or second embodiment shown in Figs. 1 to 26 except for the differences described below, and therefore, the description thereof will be omitted.

[0111] FIG. 27 is a block diagram showing an example of the configuration of an idea support system according to the third embodiment of the present invention.

[0112] In the third embodiment, a plurality of functions that operate in the divergence and convergence of ideas are prepared, and they can be switched between for use. In this case, as shown in Fig. 27, the difference from the first embodiment is that there is a function recommendation unit 2701 that recommends appropriate functions to be used according to the contents of the current user idea candidate database 105, and a recommendation record database 2702 that records the results of such functions.

[0113] The physical implementation of the idea support system 100 of the third embodiment is the same as that of the first embodiment (see FIG. 2). The function recommendation unit 2701 can be realized by the processor 201 executing a program recorded in the storage device 203. The recommendation record database 2702 can be implemented by the processor 201 executing a program that stores data in the storage device 213.

[0114] FIG. 28 is an explanatory diagram showing an example of an idea divergence screen 2800 that is presented to the user 300 in the idea divergence process by the idea support system according to the third embodiment of the present invention.

[0115] This screen is presented in the idea divergence process (step S302) of the idea divergence support unit 102. The difference from the first embodiment is that it not only allows the user to select one of the user's ideas, but also has buttons 2801 for calling up two functions, "improvement" and "leap forward." In addition, in case the user is unsure which function to call, a display 2802 indicating that the function is recommended is displayed.

[0116] FIG. 29 is a flowchart showing an example of details of the idea divergence process (step S302) performed by the idea generation support system according to the third embodiment of the present invention.

[0117] This process differs from the idea divergence process of the first embodiment shown in Fig. 10 in that steps S1001, S1004, and S1006 in Fig. 10 are replaced by steps S2901, S2902, and S2903, respectively. In the first process (step S2901), the idea divergence support unit 102 makes a recommendation as to which function to call by inquiring of the function recommendation unit 2701. The result is reflected in step S2902, which generates a user idea proposal.

[0118] In step S2902, a draft is generated by inputting text into the GPT, as in the process described in Example 1, but the content of the generated draft can be adjusted by changing the input depending on the selected function. For example, by providing an example sentence in which the content equivalent to the original text is not reflected much in the draft revision, such as "■ Original text: When using artificial intelligence to support new ideas, a database is referenced to check whether similar ideas exist. ■ Keywords: Distributed database ■ Revision: Using a distributed database speeds up similar searches, making it easier to use artificial intelligence to support new ideas," it becomes possible to generate a draft that is significantly different from the user's idea. This can be used as a "leap" function.

[0119] On the other hand, if the same input as that described in the idea divergence process of Example 1 is performed, a draft sentence will be generated that combines the original text with the newly added keywords, which can be used as a ``revision'' function.

[0120] The record of this recommendation and the result of the recommendation are recorded in the recommendation record database 2702 in the recording of the user's idea proposal (step S2903), which will be described in detail later.

[0121] FIG. 30 is a flowchart showing an example of a recommendation process performed by the function recommendation unit 2701 of the idea support system according to the third embodiment of the present invention.

[0122] First, the function recommendation unit 2701 acquires all data from the user idea candidate database 105 (step S3001). Next, the function recommendation unit 2701 sorts the acquired data in order of creation time 802 (step S3002). Next, the idea similarity evaluation unit 107 converts each piece of data into multivariate data (step S3003). Next, the function recommendation unit 2701 compares this multivariate data string with the recommendation record database 2702 and selects past recommendation cases with similar conditions (step S3004).

[0123] FIG. 31 is an explanatory diagram showing an example of the configuration of a divergent recommendation history table 3101 held by the idea generation support system according to the third embodiment of the present invention.

[0124] The divergence recommendation history table 3101 is a table storing past recommendation cases, and is included in the recommendation record database 2702. The multivariate sequence of the user idea data 3102 stores the multivariate sequence at the time the divergence function was presented. The divergence degree 3103 stores the variance value of the multivariate sequence. The content of the function actually selected under those conditions is stored as the selected content 3104. The change in the variance value of the multivariate sequence as a result of using that function is stored as the divergence degree after function use 3105.

[0125] In searching for similar cases, the multivariate sequence derived from the current user idea is compared with the multivariate sequence 3102 using a known method for comparing sequential data to select similar cases. This comparison method may use a matching algorithm or a method for converting sequential data, such as LSTM, into multivariate data. Once similar cases are selected, the change in divergence between the selection content 3104 of an improvement and that of a breakthrough is calculated, i.e., the average difference between the divergence 3105 and the divergence 3103, and the larger one is recommended (step S3005). User ideas to be recommended are selected based on the larger multivariate sequence 3102, and those with a similar relationship to the user ideas in the selection content 3104 are selected. In this example, the similarity between the selection content 3104 and other ideas is used as the standard, and the average of the minimum values ​​is calculated (step S3006), and ideas with similarly calculated indicators are selected (step S3007). This calculation is merely an example; any method that can recommend conditions similar to those that have successfully diverged in past cases may be used.

[0126] Although the present embodiment has shown an example of a recommendation function in the process of diverging ideas, a recommendation function can also be provided in the process of converging ideas in a similar manner.

[0127] Furthermore, the system according to the embodiment of the present invention may be configured as follows.

[0128] (1) An idea support system, comprising a processor (e.g., processor 201) and a storage device (e.g., storage device 203 and DRAM 202), wherein the storage device holds one or more ideas (e.g., idea data included in a user idea table 800 of a user idea candidate database 105), and the processor executes an idea divergence process (e.g., idea divergence process in step S302) for increasing the number of ideas held in the storage device, and an idea convergence process (e.g., idea convergence process in step S303) for reducing the number of ideas after being increased by the idea divergence process, In this method, the processor calculates the importance of each of the multiple ideas, which becomes lower the more likely it is that the idea can be substituted by other ideas included in the multiple ideas (e.g., steps S1804 to S1806), selects ideas with lower importance as candidates for deletion (e.g., step S1505), thereby reducing the number of ideas, and determines whether the ideas have converged among the multiple ideas stored in the storage device and that have not been selected as candidates for deletion based on their mutual similarities (e.g., step S1506), and if it is determined that the ideas have converged, executes the idea divergence process again.

[0129] This allows you to repeatedly expand and summarize your ideas based on the initial proposal, thereby refining them into higher quality ideas.

[0130] (2) In the idea support system described in (1) above, in the idea convergence process, the processor outputs the multiple ideas to the user, generates one new idea by synthesizing two or more ideas selected by the user (e.g., steps S1501 to S1502), and selects at least one of the two or more ideas selected by the user as the one to be deleted (e.g., step S1505).

[0131] This allows ideas that have once diverged to converge appropriately.

[0132] (3) In the idea generation support system described in (2) above, in the idea convergence process, the processor subtracts the importance of the two or more ideas selected by the user (e.g., step S1806).

[0133] This allows ideas that have once diverged to converge appropriately.

[0134] (4) In the idea support system described in (2) above, in the idea convergence process, the processor generates one or more clusters including the multiple ideas based on the mutual similarities of the multiple ideas (e.g., step S1804), and performs at least one of the following: if the number of ideas included in the cluster is greater than a predetermined standard, subtracting the importance of the idea included in the cluster; if the similarity between the idea and the idea with the highest similarity to the idea is higher than a predetermined standard, subtracting the importance of the idea; and subtracting the importance of the idea whose generation time is older than the predetermined standard (e.g., step S1806).

[0135] This allows ideas that have once diverged to converge appropriately.

[0136] (5) In the idea generation support system described in (2) above, the idea is expressed by a sentence, and in the idea convergence process, the processor generates the one new idea by inputting two or more of the ideas selected by the user into a machine learning model that is trained to generate a continuation of a sentence when an intermediate sentence is input (e.g., step S1502).

[0137] This allows the convergence of ideas to be assisted by processing using machine learning models.

[0138] (6) In the idea support system described in (2) above, in the idea divergence process, the processor outputs to the user one or more keywords corresponding to the idea stored in the storage device (e.g., step S1003), generates a new idea by modifying the idea based on the keyword selected by the user (e.g., step S1004), and stores the idea in the storage device (e.g., step S1006), thereby increasing the number of ideas stored in the storage device.

[0139] This allows you to broaden your ideas.

[0140] (7) In the idea generation support system described in (6) above, in the idea divergence process, the processor outputs one or more keywords extracted from literature similar to the idea stored in the storage device as one or more keywords corresponding to the idea (e.g., steps S1002 to S1003).

[0141] This allows us to provide appropriate information to broaden your ideas.

[0142] (8) The idea support system described in (6) above, wherein the idea is expressed by a sentence, and in the idea divergence process, the processor generates the new idea by inputting the idea stored in the storage device and a keyword selected by the user into a machine learning model trained to generate a continuation of a sentence when a partial sentence is input (e.g., step S1004).

[0143] This allows the divergence of ideas to be supported through processing using machine learning models.

[0144] (9) In the idea generation support system described in (1) above, in the idea generation convergence process, the processor determines whether or not to execute the idea generation divergence process again based on a predetermined condition (for example, step S1507).

[0145] This allows the repetition of diverging and converging ideas to be properly terminated.

[0146] (10) In the idea support system described in (9) above, in the idea convergence process, if the processor determines that the variation in the ideas generated sequentially by repeating the idea divergence process and the idea convergence process multiple times is small (e.g., steps S2001 to S2008), it determines not to execute the idea divergence process again.

[0147] This allows the repetition of diverging and converging ideas to be properly terminated.

[0148] (11) In the idea support system described in (9) above, the storage device holds information (e.g., termination condition data 2301) that associates the purpose of an idea with an end condition, and in the idea convergence process, if the processor determines that the magnitude of variation in the ideas generated sequentially by repeating the idea divergence process and the idea convergence process multiple times satisfies the end condition corresponding to the purpose of the idea input by the user (e.g., steps S2601 to S2602), it determines not to execute the idea divergence process again.

[0149] This allows the repetition of diverging and converging ideas to be ended appropriately according to the purpose of the ideas.

[0150] (12) In the idea support system described in (9) above, if the processor determines in the idea convergence process that the idea divergence process will not be executed again, it executes an idea result summary process (e.g., step S304), and in the idea result summary process, the processor performs an information search based on the idea stored in the storage device (e.g., step S2102), and outputs the information obtained as a result of the information search in association with the idea stored in the storage device (e.g., step S2103).

[0151] This makes it easier to check the relationship between the generated ideas and existing ideas.

[0152] (13) In the idea support system described in (6) above, in at least one of the idea divergence processing and the idea convergence processing, the processor generates the new idea in response to either a modification or a leap of the idea selected by the user (e.g., step S2902), stores the user selection and the generated new idea in the storage device (e.g., step S2903), and, based on the user selection and the generated new idea stored in the storage device, estimates which selection of the modification or the leap will contribute to the divergence of the idea in the idea divergence processing or the convergence of the idea in the idea convergence processing (e.g., steps S3004 to S3006), and outputs information recommending the modification or the leap that is estimated to have a greater contribution (e.g., step S3007).

[0153] This allows new ideas to be generated in an appropriate manner for both divergence and convergence of ideas.

[0154] (14) In the idea support system described in (13) above, the idea is expressed by a sentence, and when the leap is selected, the processor generates the new idea by providing an example sentence to a machine learning model trained to generate a continuation of a sentence when an intermediate sentence is input, such that the original idea is less likely to be reflected in the generated idea (e.g., step S2902).

[0155] This allows new ideas to be generated in an appropriate manner for both divergence and convergence of ideas.

[0156] The present invention is not limited to the above-described embodiments, but includes various modifications. For example, the above-described embodiments have been described in detail to provide a better understanding of the present invention, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0157] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as programs, tables, and files that implement each function can be stored in storage devices such as nonvolatile semiconductor memory, hard disk drives, and solid-state drives (SSDs), or in computer-readable, non-transitory data storage media such as IC cards, SD cards, and DVDs.

[0158] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and not all control lines and information lines in the product are necessarily shown. In reality, it can be considered that almost all components are interconnected.

Claims

1. An idea support system having a processor and a storage device, the storage device holding one or more ideas, the processor executing an idea divergence process to increase the number of ideas held in the storage device, and an idea convergence process to reduce the number of ideas after the increase by the idea divergence process, wherein in the idea convergence process, the processor calculates for each of a plurality of ideas an importance that becomes lower the more likely an idea is to be substituted by other ideas included in the plurality of ideas, and reduces the number of ideas by selecting ideas with a lower importance as candidates for deletion, and determines whether the ideas have converged based on their mutual similarities for ideas among the plurality of ideas held in the storage device that have not been selected as candidates for deletion, and if it is determined that the ideas have converged, executes the idea divergence process again.

2. An idea generation support system as described in claim 1, characterized in that, in the idea convergence process, the processor outputs the multiple ideas to a user, generates a new idea by synthesizing two or more of the ideas selected by the user, and selects at least one of the two or more ideas selected by the user as the one to be deleted.

3. An idea generation support system as claimed in claim 2, characterized in that in the idea convergence process, the processor subtracts the importance of the two or more ideas selected by the user.

4. An idea support system as described in claim 2, wherein in the idea convergence processing, the processor performs at least one of the following: generating one or more clusters including the multiple ideas based on the mutual similarities between the multiple ideas; and, if the number of ideas included in the cluster is greater than a predetermined criterion, subtracting the importance of the idea included in the cluster; if the similarity between the idea and the idea that has the highest similarity to the idea included in the cluster is higher than a predetermined criterion, subtracting the importance of the idea; and subtracting the importance of the idea whose generation time was older than the predetermined criterion.

5. An idea support system as described in claim 2, wherein the idea is expressed by a sentence, and in the idea convergence process, the processor generates the one new idea by inputting two or more of the ideas selected by the user into a machine learning model trained to generate a continuation of a sentence when an intermediate sentence is input.

6. An idea support system as described in claim 2, characterized in that, in the idea divergence process, the processor outputs to the user one or more keywords corresponding to the ideas stored in the storage device, generates new ideas by modifying the ideas based on the keywords selected by the user, and stores the ideas in the storage device, thereby increasing the number of ideas stored in the storage device.

7. An idea generation support system as described in claim 6, characterized in that in the idea divergence process, the processor outputs one or more keywords extracted from literature similar to the idea stored in the storage device as one or more keywords corresponding to the idea.

8. An idea support system as described in claim 6, wherein the idea is expressed by a sentence, and in the idea divergence process, the processor generates the new idea by inputting the idea stored in the storage device and a keyword selected by the user into a machine learning model trained to generate a continuation of a sentence when an intermediate sentence is input.

9. An idea generation support system as claimed in claim 1, characterized in that in the idea convergence process, the processor determines whether or not to execute the idea divergence process again based on a predetermined condition.

10. An idea support system as described in claim 9, characterized in that, in the idea convergence process, when the processor determines that there is little variation in the ideas generated sequentially by repeating the idea divergence process and the idea convergence process multiple times, it decides not to execute the idea divergence process again.

11. An idea support system as described in claim 9, wherein the storage device holds information that corresponds an idea purpose to a termination condition, and in the idea convergence process, the processor determines not to execute the idea divergence process again when it determines that the magnitude of variation in the ideas generated sequentially by repeating the idea divergence process and the idea convergence process multiple times satisfies the termination condition corresponding to the idea purpose input by the user.

12. An idea support system as described in claim 9, wherein, when the processor determines in the idea convergence process that the idea divergence process will not be executed again, it executes an idea result summary process, and in the idea result summary process, the processor performs an information search based on the idea stored in the storage device, and outputs information obtained as a result of the information search in association with the idea stored in the storage device.

13. An idea support system as described in claim 6, wherein in at least one of the idea divergence processing and the idea convergence processing, the processor generates the new idea in response to either a modification or a leap of the idea selected by the user, stores the user's selection and the generated new idea in the storage device, estimates which selection of the modification or the leap will contribute to the divergence of the ideas in the idea divergence processing or the convergence of the ideas in the idea convergence processing based on the user's selection and the generated new idea stored in the storage device, and outputs information recommending the modification or the leap that is estimated to have a greater contribution.

14. An idea support system as described in claim 13, wherein the idea is expressed by a sentence, and when the leap is selected, the processor generates the new idea by providing an example sentence to a machine learning model trained to generate a continuation of a sentence when an intermediate sentence is input, in which the original idea is unlikely to be reflected in the generated idea.

15. A method for supporting ideas using an idea generation support system, the idea generation support system having a processor and a storage device, the storage device holding one or more ideas, the idea generation support method including: an idea divergence procedure in which the processor increases the number of ideas held in the storage device; and an idea convergence procedure in which the processor reduces the number of ideas after being increased by the idea divergence procedure, wherein in the idea convergence procedure, the processor calculates, for each of a plurality of ideas, an importance that becomes lower the more likely it is that the ideas included in the plurality of ideas can be substituted, and reduces the number of ideas by selecting ideas with a lower importance as candidates for deletion, and determines whether the ideas have converged based on their mutual similarities for ideas among the plurality of ideas held in the storage device that have not been selected as candidates for deletion, and if it is determined that the ideas have converged, executes the idea divergence procedure again.

Images