Idea expansion system, idea expansion method, and idea expansion program

Abstract

To provide new technology that supports the creation of new ideas based on a policy that reflects the user's intentions. [Solution] An idea expansion system comprising an idea receiving unit, a policy receiving unit, and an expansion unit, wherein the idea receiving unit receives idea input indicating a user's idea, the policy receiving unit receives policy input indicating a policy for expanding the idea, and the expansion unit inputs to a generative model based on the idea input and policy input, and obtains an expanded idea that expands the idea based on the output of the generative model.

Description

[Technical Field]

[0001] The present invention relates to an idea expansion system, an idea expansion method, and an idea expansion program. [Background technology]

[0002] A variety of idea generation tools have been proposed so far, including framework-based methods, communication tools for expanding ideas in groups, and idea expansion tools that utilize generative AI.

[0003] Framework methods are ways of generating ideas by following a specific structure or procedure. Examples include SWOT analysis and mind maps. While these methods make it easy to organize ideas systematically, they have the drawback of making ideas tend to become fixed.

[0004] Communication tools are tools for expanding ideas through group brainstorming and discussion. These include online meeting tools and collaboration platforms. While these tools make it easy to incorporate opinions from diverse perspectives, they can also pose challenges due to conflicts of opinion and time constraints.

[0005] Generative AI idea expansion tools have the ability to generate new ideas based on large amounts of data.

[0006] Patent Document 1 also discloses an idea generation support system that accepts a user's idea (user idea) as input, selects related idea candidates (first idea candidates) from a database that stores existing ideas, and then generates second idea candidates by converting the classification of the first and second idea candidates. This system associates the user's idea with the first and second idea candidates and presents them, allowing the user to come up with new ideas without spending much effort or time. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Japanese Patent Application Laid-Open No. 2024-79138 Summary of the Invention [Problem to be solved by the invention]

[0008] However, the technology described in Patent Document 1 has a problem in that, since idea candidates are generated based on existing ideas, the ideas proposed are not necessarily in line with the user's wishes.

[0009] In view of the above circumstances, an object of the present invention is to provide a novel technology for supporting the creation of new ideas in accordance with a policy that reflects the user's intentions. [Means for solving the problem]

[0010] [1] An idea expansion system comprising an idea receiving unit, a policy receiving unit, and an expansion unit, wherein the idea receiving unit receives an idea input indicating a user's idea, the policy receiving unit receives a policy input indicating a policy for expanding the idea, and the expansion unit performs input to a generative model based on the idea input and the policy input, and obtains an expanded idea that expands the idea based on the output of the generative model.

[0011] By adopting such a configuration, it is possible to obtain expansion ideas that reflect the user's input of guidelines.

[0012] [2] The idea expansion system described in [1], wherein the idea receiving unit receives the idea input in natural language, and the expansion unit inputs an instruction statement in natural language to the generative model, which is an instruction to expand the idea, based on the idea input and the policy input.

[0013] [3] The idea expansion system described in [1] or [2], wherein the policy receiving unit receives the policy input by having the user select from a plurality of options, and the expansion unit inputs an instruction statement in natural language to the generative model, the instruction statement including the idea input and a word based on the policy input, to expand the idea.

[0014] This configuration makes it easier for the user to consider how to expand on an idea.

[0015] [4] The idea expansion system according to [3], wherein the options include at least one of a user, a purpose, a problem, a solution method, and a competition.

[0016] [5] An idea expansion system described in any of [1] to [4], further comprising a dialogue unit, which determines the statements of multiple people including a facilitator and an idea person, and conducts a dialogue between three or more people including a user, and the dialogue unit determines a statement that prompts the policy input as the statement of the facilitator based on the idea input, and the policy receiving unit receives the user's statement as the policy input.

[0017] With this configuration, the facilitator's comments make it easier for users to input guidelines.

[0018] [6] The idea expansion system according to [5], wherein the dialogue unit outputs the expanded idea as a statement by the idea person.

[0019] This configuration makes it easy to input opinions and adjustments to expansion ideas while interacting with them.

[0020] [7] The idea expansion system described in [5] or [6], wherein the dialogue unit associates the speech history of the facilitator, the idea person, and the user with the speaker and displays it in chronological order.

[0021] This configuration allows the user to check the process of the investigation, and effectively supports the creation of ideas.

[0022] [8] The extension unit further inputs an instruction to the generative model to output a sentence summarizing the extended idea based on the remarks of the multiple people in response to the user's remarks, and passes the output of the generative model to the dialogue unit, and the dialogue unit outputs the extended idea based on the output of the generative model. An idea extension system described in any of [5] to [7].

[0023] With this configuration, it is possible to expand ideas and exchange opinions through dialogue, and then output an expanded idea that summarizes the results.

[0024] [9] An idea expansion system according to any one of [1] to [8], further comprising an extraction unit, which, based on the idea input, refers to a reference idea storage unit that stores publications in which existing reference ideas are described, and extracts similar publications in which ideas similar to the idea are described.

[0025] With this configuration, it is possible to present ideas similar to the idea input by the user, thereby assisting in the consideration of expanding the idea.

[0026]

[10] The idea expansion system described in [9], wherein the idea receiving unit receives the idea input in natural language, the publication describes the reference idea in sentences, and the extraction unit extracts the similar publications by calculating the similarity between a part of the publication and the idea input.

[0027]

[11] The idea expansion system described in [9] or

[10] , wherein the extraction unit extracts a plurality of the similar publications, the policy reception unit receives a user selection from among the plurality of the similar publications, and the expansion unit inputs to the generative model based on the idea input and the selected similar publications.

[0028] This configuration makes it possible to provide users with extension ideas based on similar publications.

[0029]

[12] An idea expansion system according to any one of [9] to

[11] , further comprising an output unit that displays a proposal screen showing the expansion idea, the proposal screen having a similar idea extraction button for instructing extraction of ideas similar to the expansion idea, and the extraction unit, by operating the similar idea extraction button, extracts from the reference idea storage unit similar publications that describe ideas similar to the expansion idea.

[0030] With this configuration, similar publications can be extracted for the extension idea, and the user can be effectively supported in further deepening their consideration.

[0031]

[13] The idea expansion system according to any one of [9] to

[12] , wherein the publication is a patent document.

[0032]

[14] An idea expansion system according to any one of [1] to

[13] , further comprising an output unit, wherein the output unit displays a proposal screen showing the expansion idea, the proposal screen comprises an input button for inputting the expansion idea, and the idea receiving unit receives the expansion idea as the idea input by operating the input button.

[0033] With this configuration, the outputted extension idea can be further extended, and the user can be assisted in further examining the idea.

[0034]

[15] An idea expansion method executed by a computer having an idea receiving unit, a policy receiving unit, and an expansion unit, wherein the idea receiving unit receives an idea input indicating a user's idea, the policy receiving unit receives a policy input indicating a policy for expanding the idea, the expansion unit inputs to a generative model based on the idea input and the policy input, and obtains an expanded idea that expands the idea based on the output of the generative model.

[0035]

[16] An idea expansion program that causes a computer to execute the idea expansion method described in

[15] . [Effects of the Invention]

[0036] According to the present invention, it is possible to support the creation of new ideas based on a policy that reflects the user's intentions. [Brief explanation of the drawings]

[0037] [Figure 1] FIG. 2 is a schematic diagram showing the main functions of the present embodiment. [Figure 2] FIG. 1 is a configuration diagram of an embodiment of the present invention. [Figure 3] Hardware configuration diagram. [Figure 4] 10 is a flowchart for acquiring expansion ideas. [Figure 5] Example of output from the drawing creation section. [Figure 6] 10 is an example of a screen display related to expansion by policy selection. [Figure 7] 10 is an example of a screen display related to interactive extension. [Figure 8] 1 is a flowchart for extracting similar publications and expanding based on similar ideas. [Figure 9] An example of the similar publications display screen. [Figure 10] FIG. 10 is a schematic diagram relating to calculation of similarity. DETAILED DESCRIPTION OF THE INVENTION

[0038] <Summary> The present invention relates to a technology for supporting further expansion of an idea input by a user (hereinafter referred to as an "input idea"). The present invention will be described in more detail below with reference to the accompanying drawings. While the drawings show preferred embodiments, the present invention may be embodied in many different forms and is not limited to the embodiments set forth herein.

[0039] For example, in this embodiment, the configuration, operation, etc. of the idea expansion system will be described, but similar effects can be achieved by a device having similar functions, a method executed by the device, a computer program that causes a computer device to execute the method, etc. The program may be provided as a non-transitory computer-readable recording medium, or may be provided so as to be downloadable from an external server.

[0040] 1 is a schematic diagram showing an overview of the main functions of the idea expansion system in this embodiment. The main functions of the idea expansion system in this embodiment can be divided into three: an idea input function, a similar idea extraction function, and an idea expansion function.

[0041] The idea input function is a function for accepting input of ideas from the user, and also includes a function for re-inputting extended ideas output by the idea extension function.

[0042] The similar idea extraction function is a function that extracts similar publications that describe ideas similar to the input idea from a reference idea storage unit that stores publications that describe existing reference ideas, based on the input idea (including the extended idea), and then displays information about the similar publications.

[0043] The idea expansion function is a function that proposes an expanded idea for an idea input by a user using the idea input function. In this embodiment, there are three types of expansion methods: expansion by policy selection, expansion by dialogue, and expansion by similar ideas. Expansion by policy selection and expansion by dialogue are methods of expanding an input idea based on a policy instruction by the user.

[0044] In the policy selection extension, the system accepts the user's desired extension policy from multiple options and proposes extension ideas based on the selected policy.In the dialogue extension, the system accepts the user's statements indicating the desired extension policy through a chat format between the user and the system, and proposes extension ideas based on the content of the user's statements.

[0045] The similar idea extension is a method of extending an input idea based on similar publications extracted by a similar idea extraction function. For example, a list of similar publications may be output and the user may select one to use in proposing an extension idea. Note that the extension idea may be proposed based on publications uploaded by the user, not limited to publications extracted by the similar idea extraction function.

[0046] In this embodiment, the idea input function, similar idea extraction function, and idea expansion function work together to support the user's idea creation. This allows the user to check whether there are any similar ideas for the proposed expansion idea, and to repeatedly expand the expansion idea, thereby effectively supporting the user's consideration.

[0047] <System configuration> The present embodiment will be described in detail below. In the following embodiments, the term "unit" may include, for example, hardware resources implemented by a broadly defined circuit and software information processing that can be specifically realized by these hardware resources. In the present embodiment, "information" is represented by, for example, the physical value of a signal value representing voltage or current, the high or low value of a signal value as a binary bit set consisting of 0 or 1, or quantum superposition (so-called quantum bits), and communication and calculation can be performed on a broadly defined circuit.

[0048] A circuit in the broad sense is a circuit realized by appropriately combining a circuit, a processor, a memory, etc. For example, it is a circuit including any of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), etc.

[0049] Fig. 2 is a block diagram showing the configuration of the idea expansion system of this embodiment. As shown in Fig. 2, the idea expansion system includes an idea expansion device 1 and a user terminal 2. The idea expansion device 1 and the user terminal 2 are configured to be able to communicate with each other via a network NW. The idea expansion device 1 is also connected to a generation device 3 via the network NW, and is also configured to be able to communicate with a reference idea storage unit DB. In this embodiment, the network NW is an IP (Internet Protocol) network, but there are no restrictions on the type of communication protocol, the type of network, etc.

[0050] The idea expansion device 1 accepts various inputs, including idea input and policy input, from a user terminal 2, and proposes an expansion idea. When proposing an expansion idea, the device communicates with a generation device 3 and makes a proposal based on the output of a generation model.

[0051] The user terminal 2 accepts various input operations from the user, communicates with the idea expansion device 1, and proposes expansion ideas to the user by displaying on the screen or outputting by voice.

[0052] The generation device 3 is a server that provides a generation model using, for example, large language models (LLMs), etc. Any system or service can be used as the generation device 3, but as an example, it is assumed to use an interactive text generation service that returns an output in natural language in response to an input in natural language.

[0053] 1 is an example, and the idea expansion device 1 may be realized by the cooperation of multiple computers. Also, a part of the configuration of the idea expansion device 1 or the generation device 3 may be realized in another computer such as the user terminal 2.

[0054] <Hardware configuration> Next, the hardware configuration of the idea expansion system in this embodiment will be described. One or more information processing devices 10 (computer devices) such as a general-purpose server or a personal computer can be used as the idea expansion device 1 and the generation device 3. Furthermore, a terminal device 9 (computer device) such as a personal computer, a smartphone, or a tablet terminal can be used as the user terminal 2. In this embodiment, the idea expansion device 1 is an information processing device 10 in which a computer program (idea expansion program) that executes the idea expansion method is installed.

[0055] Fig. 3(a) is a hardware configuration diagram of the information processing device 10. As shown in Fig. 3, the information processing device 10 has a control unit 101, a storage unit 102, and a communication unit 103, which are used to perform the functions of each unit and each process.

[0056] The control unit 101 has a processor such as a CPU that can execute an instruction set, and executes an OS and programs. The storage unit 102 includes a volatile memory such as a RAM capable of storing an instruction set, and a non-volatile recording medium such as an HDD or SSD capable of recording an OS, an idea expansion program, a DBMS, and the like. The communication unit 103 has an interface for physically connecting to a network, and controls communication with the network NW to input and output information.

[0057] Fig. 3(b) is a hardware configuration diagram of the terminal device 9. As shown in Fig. 3, the terminal device 9 has a control unit 901, a storage unit 902, a communication unit 903, an input unit 904, and an output unit 905, which are used to perform the functions of each unit and each process.

[0058] The control unit 901 has a processor such as a CPU that can execute an instruction set, and executes an OS, application programs, and the like. The storage unit 902 includes a volatile memory such as a RAM capable of storing an instruction set, and a non-volatile recording medium such as an HDD or SSD capable of recording an OS, any application program, and the like. The communication unit 903 has an interface for physically connecting to a network, and controls communication with the network NW to input and output information. The input unit 904 includes an operation input device capable of input processing, such as a touch panel or keyboard, and an audio input device capable of audio input, such as a microphone. The output unit 905 includes a display device capable of display processing, such as a display, and an audio output device, such as a speaker.

[0059] <Functional configuration> Next, referring again to FIG. 2, the functional configuration of the idea expansion device 1 will be described. The idea expansion device 1 has an idea receiving unit 11, a policy receiving unit 12, an expansion unit 13, a dialogue unit 14, an extraction unit 15, an output unit 16, and a drawing creation unit 17. This is a specific implementation of software-based information processing using hardware. The idea expansion device 1 is also configured to be able to communicate with a reference idea storage unit DB. Note that these components do not need to be realized by a single computer, but may be realized by multiple computers working together. The reference idea storage unit DB may be provided as part of the idea expansion device 1, or an external service available via a network NW may function as the reference idea storage unit DB.

[0060] The idea receiving unit 11 receives an idea input indicating an idea of ​​a user via the user terminal 2. In this embodiment, the idea input is received as a sentence or a paragraph in a natural language. Alternatively, the idea may be input as information in another format, such as an image, or a combination of an image and a paragraph.

[0061] The policy receiving unit 12 receives a policy input that indicates a policy for expanding an idea. In this embodiment, the policy input is received in two formats: a format in which a policy is selected from a plurality of options, and a format in which an instruction is given in text. More specifically, the policy input is received in the former format in the case of expansion by policy selection, and in the latter format in the case of expansion by dialogue.

[0062] The extension unit 13 performs input to the generative model based on the idea input and the policy input, and acquires an extended idea by expanding the input idea based on the output of the generative model. Specifically, the extension unit 13 generates an instruction statement based on the idea input and the policy input, and transmits the instruction statement to the generation device 3 to perform input to the generative model. Then, the extension unit 13 receives information based on the output of the generative model from the generation device 3, and acquires an extended idea based on the received information.

[0063] In this embodiment, the generation device 3 uses a large-scale language model to output the extended idea as a sentence or paragraph in natural language, which is then received by the extension unit 13 for subsequent processing.

[0064] The dialogue unit 14 determines the statements of multiple characters (characters) including the facilitator and the idea man, and conducts a dialogue between three or more people including the user. In this embodiment, instructions are sent to the generation device 3 for each of the roles of the facilitator and the idea man, causing the generation device 3 to generate each statement based on the user's statements and statements of other characters. Here, the idea man does not need to make a statement every time the user makes a statement, and may be instructed to make a statement with a certain probability, for example.

[0065] For example, instructions for the facilitator might include, "You are the facilitator. Ask questions to help the user flesh out their ideas. In particular, ask questions to dig deeper into their ideas." Similarly, instructions for the idea man might include, "You are the idea man. Please speak about specific ideas to broaden the user's thinking. Please provide ideas about once every two or three times the user speaks."

[0066] The extraction unit 15 refers to the reference idea storage unit based on the idea input, and extracts similar publications that describe ideas similar to the input idea. The details of the process of extracting similar publications will be described later.

[0067] The output unit 16 performs various outputs to the user terminal 2. For example, the output unit 16 of this embodiment performs display processing of a proposal screen showing an extension idea, transmits the processing result to the user terminal 2, and displays the proposal screen on the user terminal 2. The output unit 16 also controls the display of various other input / output screens. Examples of display screens will be described later.

[0068] The diagram creation unit 17 creates diagrams for ideas described in sentences or paragraphs in natural language, including input ideas and extended ideas. Specifically, the diagram creation unit 17 may transmit the idea described in sentences or paragraphs to an available generative model via the generation device 3 or another server, etc., and instruct the generation of a diagram. For example, a flowchart or a block diagram may be generated as the diagram.

[0069] The reference idea storage unit DB stores publications in which existing reference ideas are described. In this embodiment, patent documents are stored as the publications. In particular, in this embodiment, information on patent documents is stored in XML format, and the contents of the documents are stored for each item described in the specification of the application documents. Note that while it is preferable to use documents with a unified format, such as patent documents, as publications in the present invention, the publications are not limited to patent documents, and any document may be used.

[0070] In this embodiment, a dedicated database is constructed and used as the reference idea storage DB, but a database of an external service may also be used as the reference idea storage DB. For example, an existing patent information database may also be used as the reference idea storage DB.

[0071] <Procedure for obtaining expansion ideas> The processing procedure in this embodiment will be described in detail below. First, the extension by policy selection and the extension by dialogue will be described with reference to FIGS.

[0072] FIG. 4 is a flowchart showing the flow of processing up to acquisition of an expansion idea when expansion by policy selection or expansion by dialogue is performed.

[0073] First, in step S11, a user inputs his or her own idea via the user terminal 2. In this embodiment, input is assumed to be in the form of a sentence or paragraph in a natural language, but the input format is not particularly limited. Then, the idea receiving unit 11 receives information based on the idea input from the user terminal 2 via the network NW.

[0074] Next, in step S12, the user inputs a policy that indicates a policy for expanding the input idea via the user terminal 2. The policy receiving unit 12 receives information based on the policy input from the user terminal 2 via the network NW. Then, the policy receiving unit 12 generates an instruction statement regarding the expansion of the idea based on the idea input and the policy input, and transmits it to the generation device 3, thereby inputting it into the generative model.

[0075] In step S13, the extension unit 13 generates an instruction sentence and inputs the instruction sentence to the generative model. Different rules may be applied to the generation of the instruction sentence depending on the type and content of the idea. It is also possible to adaptively generate an instruction sentence by referring to the user's past input history. In this embodiment, the extension unit 13 generates an instruction sentence including an input idea input in natural language and a sentence based on the policy input. For example, when the policy input is performed by selecting from pre-set options, the sentence based on the policy input is a sentence prepared in advance for each option. When the policy input is performed by a sentence, the input sentence may be used as is as the sentence based on the policy input.

[0076] The extension unit 13 transmits the directive to the generation device 3 via the network NW, and the generation device 3 inputs the directive to the generative model, which then outputs the extended idea.

[0077] In step S14, the extension unit 13 acquires the extension idea output by the generative model. In this embodiment, the generative model outputs the extension idea, and the extension unit 13 acquires the output of the generative model as the extension idea as is, but the extension idea may also be acquired by the extension unit 13 performing processing such as selection and modification based on the output of the generative model. The extension idea is provided to the user by the output unit 16 via the user terminal 2.

[0078] Furthermore, the diagram creation unit 17 may further create a diagram for the extended idea. For example, the diagram creation unit 17 may transmit the extended idea described in a sentence or a paragraph to an available generative model via the generation device 3 or another server, etc., and instruct the generation of a diagram. For example, a flowchart or a block diagram may be generated as the diagram.

[0079] As an example, FIG. 5 shows a drawing created by the drawing creation unit 17 of this embodiment based on the following sentence: An idea expansion system including an idea receiving unit, a policy receiving unit, and an expansion unit, the idea receiving unit receives an idea input indicating a user's idea; the policy receiving unit receives a policy input indicating a policy for expanding the idea; The extension unit provides input to a generative model based on the idea input and policy input, and obtains an extended idea by extending the idea based on the output of the generative model.

[0080] The drawing creation unit 17 creates an instruction statement including the above sentence (the idea described in a natural language sentence) and sends it to the generation device 3, thereby inputting it into the generative model, and obtains the drawing shown in Figure 5 based on the output of the generative model. The instruction statement may include instructions on how to create the drawing. For example, it is expected to include content such as "Please convert the idea below into Mermaid notation and create a drawing," which instructs the user to convert the input sentence into a specific notation and then create the drawing. Figure 5 shows an image actually generated by instructing according to the above procedure. In this way, the drawing creation unit 17 inputs an idea expressed in natural language into the generative model, allowing the user to create a drawing representing the idea.

[0081] By following this procedure, an extended idea can be obtained by extending the input idea according to the policy specified by the user. Next, examples of screen displays will be described for both extension by policy selection and extension by dialogue.

[0082] <Example of expansion by policy selection> Fig. 6 is an example of a screen display when performing expansion by policy selection. When inputting an idea, the output unit 16 displays an idea input field and an expansion policy selection field on the user terminal 2, as shown in Fig. 6(a). In this embodiment, five options are set in advance as expansion policies: user detailing, goal detailing, problem detailing, solution method detailing, and competing technology / opponent detailing, and a selection is accepted from among these.

[0083] When an operation to instruct the proposal of an extension idea is performed on the screen shown in Fig. 6(a), the extension idea is acquired in the above-described procedure, and the extension idea is output on the screen shown in Fig. 6(b). In this embodiment, the screen shown in Fig. 6(b) is used as a proposal screen that displays the extension idea. On the proposal screen of this embodiment, the input idea ("input text" in Fig. 6(b)), the selected extension policy ("selected item" in Fig. 6(b)), and the extension idea ("extension result" in Fig. 6(b)) are displayed.

[0084] The proposal screen of this embodiment also includes a similar idea extraction button for instructing extraction of ideas similar to the extended idea ("Extract Similar Ideas" in FIG. 6(b)). When the user selects the similar idea extraction button, a similar publication extraction process (described later) is executed for the displayed extended idea, and similar publications are extracted from the reference idea storage unit DB.

[0085] <Example of interactive expansion> Fig. 7 shows an example of a screen display when expanding through dialogue. When inputting an idea, the output unit 16 first displays an idea input field on the user terminal 2 as shown in Fig. 7(a). When the send button is selected, the dialogue screen shown in Fig. 7(b) is displayed, and dialogue with multiple characters begins.

[0086] In this embodiment, the dialogue unit 14 determines what the facilitator and the idea person will say, and a dialogue is held between three people including the user. There may be multiple idea people, and further, an expansion of the idea may be proposed through a dialogue between four or more people including characters with other roles.

[0087] As shown in Fig. 7(a), the dialogue screen of this embodiment displays the statements of all three dialogue participants, i.e., the user, the facilitator, and the idea person, in chronological order, distinguishing between the speakers. Although omitted in Fig. 7(a), an input field is provided at the bottom of the screen, etc., and the user can input statements by entering characters via the user terminal 2.

[0088] The dialogue unit 14 accepts user utterances, inputs them to the generative model, and passes them on to the extension unit 13. Specifically, the dialogue unit 14 generates instruction sentences including the user utterances and instructs the generative model to generate utterances from the facilitator. As already mentioned, the facilitator is instructed to ask questions to the users to concretize their ideas. Note that in addition to questions, the facilitator may also be configured to organize the discussion. In this embodiment, the dialogue unit 14 causes the generative model to generate the facilitator's utterances every time a user utters, and determines the output of the generative model as the facilitator's utterance. Note that when an idea person utters, the facilitator's utterance may also be determined after the idea person utters.

[0089] The dialogue unit 14 also passes the user's utterance to the extension unit 13. The extension unit 13 accepts the user's utterance as a policy input and acquires an extension idea using the procedure described above. Here, in the case of extension through dialogue, the extension unit 13 inputs to the generative model an instruction statement to output an extension idea as an idea person's utterance based on the input idea and the user's utterance indicating the policy. In response to this, the generative model outputs the idea person's utterance proposing an extension idea, and the extension unit 13 acquires this utterance as an extension idea. The extension unit 13 then passes this utterance to the dialogue unit 14, which then decides on the idea person's utterance.

[0090] The comments determined in this way are displayed on the dialogue screen as shown in Fig. 7(b). Therefore, the dialogue screen also includes comments from idea people showing extended ideas, and the dialogue screen is an example of a proposal screen in the present invention.

[0091] The above-described procedure and the cooperation of each unit are merely examples, and any changes can be made as long as the series of procedures of inputting to the generative model based on the user's utterances and determining the idea person's utterances based on the output of the generative model can be executed. In other words, the processes of the dialogue unit 14 and the extension unit 13 do not need to be separate, and they may function as an integrated processing unit to determine the idea person's utterances.

[0092] In addition, the extension unit 13 of this embodiment summarizes the exchanges made in a series of dialogues to extend ideas during dialogue-based extension. For example, when a user says "summarize," the extension unit 13 generates an instruction sentence including a series of statements (statement history) made by the facilitator, idea person, and user (all participants in the dialogue), and inputs the instruction sentence into the generative model to generate a sentence summarizing the extended ideas that came up in the dialogue. The extension unit 13 then obtains the output of the generative model, and the output unit 16 outputs a summary of the extended ideas (summary of the dialogue), as shown in FIG. 7(c). Here, the summary of the dialogue is also included in the extended ideas of the present invention.

[0093] The interactive screen of this embodiment also has a similar idea extraction button for instructing the extraction of ideas similar to the extended idea ("Extract similar ideas" in FIG. 7(c)). The interactive screen may also have a similar idea extraction button for each statement made by the idea person. When the user selects the similar idea extraction button, a similar publication extraction process, which will be described later, is executed for the displayed extended idea, and similar publications are extracted from the reference idea storage unit DB. The idea extraction button is an example of an input button of the present invention. In addition, an input button for further extension of the extended idea may also be provided.

[0094] <Similar publication extraction process - expansion based on similar ideas> Next, a flow from the similar publication extraction process by the extraction unit 15 to the expansion using similar ideas will be described with reference to Figures 8 and 9. Figure 8 is a flowchart of the process from when the extraction unit 15 extracts similar publications from the reference idea storage unit DB to when the expansion idea based on the similar idea is acquired.

[0095] First, in step S21, the idea receiving unit 11 receives an idea input by a user. The idea input may be an operation of the similar idea extraction button described above. In this case, the extended idea corresponding to the operated similar idea extraction button is treated as the input idea.

[0096] In step S22, the extraction unit 15 calculates the similarity between the input idea and each publication stored in the reference idea storage unit DB. The calculation of the similarity will be described later.

[0097] Then, in step S23, the extraction unit 15 extracts similar publications based on the similarity calculated in step S22. Specifically, publications with a high similarity to the input idea are extracted as similar publications. For example, a predetermined number of similar publications may be extracted in descending order of similarity, or publications with a similarity to the input idea that is equal to or greater than a predetermined reference value may be selected as similar publications.

[0098] In step S24, the output unit 16 outputs the extracted similar publications and displays them on the user terminal 2. Fig. 9 shows an example of the display of a similar publication display screen in this embodiment. The similar publication display screen has an idea input field at the top of the screen, and the idea input in step S21 described above is accepted when the user inputs an idea in text (selecting the "Go" button in Fig. 9 with the text entered in the input field). As a result, a similar publication display screen including the input idea and a list of similar publications is displayed.

[0099] In the list of similar publications, basic information for each similar publication is displayed, including the name of the invention, ID, the content distance indicating the degree of similarity, and the publication number (in this embodiment, a patent publication). When "Details" is selected for each publication, detailed information is displayed, as in the case of the third publication in Figure 9.

[0100] In this embodiment, patent publications are stored as publications in the reference idea storage unit DB. Therefore, information about each publication may further include descriptions in the abstract and descriptions in the "Means for Solving the Problems" section of the specification. Furthermore, in this embodiment, a summary of the "Mode for Carrying Out the Invention" ("AI Summary" in FIG. 9) is displayed, which is a summary of the text using a large-scale language model. The extraction unit 15 may generate a summary each time by inputting a directive to generate a summary of the publication's contents into the generative model. Alternatively, summaries generated by the generative model may be stored in advance in the reference idea storage unit DB in association with each publication. Furthermore, the items to be summarized may be changed; for example, the entire specification may be summarized and displayed.

[0101] Furthermore, it is preferable to provide links in the list of similar publications so that the full text of the publication can be viewed. For example, it is expected that the list will display links to information about the publication stored in the reference idea storage unit DB, or links to the pages of each publication in external databases such as J-PlatPat.

[0102] In the list of similar publications, a selection column (check box) for each publication is displayed so that the user can select it. Then, when one or more publications are selected and a button (hereinafter referred to as the "expand button") is operated, the expansion unit 13 accepts the selection of similar publications in step S25.

[0103] Then, in step S26, the extension unit 13 generates a directive statement including the input idea and the selected similar publication (hereinafter referred to as the selected publication) and inputs it to the generative model. As a result, in step S27, the extension unit 13 obtains an extended idea based on the input idea and the selected publication as the output of the generative model. Here, as in the cases of extension by policy selection and extension by dialogue, the output of the generative model may be used as the extended idea as is, or information processed based on the output of the generative model may be used as the extended idea.

[0104] The display of the extended ideas can be performed in the same manner as the proposal screen shown in Fig. 6. The proposal screen has a button for extracting similar ideas, allowing repeated extraction of similar publications. As explained in Fig. 6, an input button for further expanding the extended ideas may also be provided.

[0105] This makes it easy for users to repeatedly execute the steps of inputting an idea, checking similar ideas, and expanding the idea. Therefore, while referring to similar ideas, users can receive proposals for expanding their idea from various angles according to their intended direction, and can also check similar ideas for the expanded idea. In this way, the repeated review process can be smoothly carried out, effectively supporting users in generating ideas.

[0106] <Similarity calculation> Next, the calculation of the similarity will be described with reference to Fig. 10. Fig. 10 is a schematic diagram relating to the calculation of the similarity between the input idea and patent documents stored in the reference idea storage unit DB in this embodiment.

[0107] 10, in this embodiment, the reference idea storage unit DB stores information on patent documents in XML format. The information stored in the reference idea storage unit DB can be, for example, information obtainable from the patent information platform "J-PlatPat" provided by the National Institute of Industrial Property Information and Research (INPIT).

[0108] The information on patent documents stored in the reference idea storage unit DB is organized by documents such as claims and specifications, and items in the specifications (problems to be solved by the invention, effects of the invention, modes for carrying out the invention, etc.). In this embodiment, the sentences that make up the patent documents are converted into vectors for each item, and similarity calculations are performed between the vectors of the input idea to calculate the similarity between each publication and the input idea.

[0109] Text vectorization can be performed by any method. For example, methods such as doc2vec, ELMo, and Sentence Bert are known. The reference idea storage unit DB stores the vectorized information in advance, in association with each publication and item. Note that vectorization may be performed each time a similarity is calculated.

[0110] In this embodiment, all of the items constituting a patent document are vectorized. In particular, it is preferable to perform vectorization and similarity calculation on the claims and the problems that the invention aims to solve. Note that instead of or in addition to vectorizing each item, the entire text of the publication may be further vectorized. Also, vectorization and similarity calculation may be performed on only some of the items. The input ideas are also vectorized by the extraction unit 15. The input ideas are vectorized using the same method as that used for vectorizing publications stored in the reference idea storage unit DB. Then, the vector for each item of the publication is compared with the vector of the input idea to calculate the similarity. Specifically, the extraction unit 15 of this embodiment calculates the cosine similarity between both vectors.

[0111] The extraction unit 15 then integrates the similarity for each item to calculate the similarity between each publication and the input idea. The similarity may be integrated, for example, using a function weighted for each item, a simple sum, or an average value. Once the similarity for each publication is calculated, the extraction unit 15 extracts similar publications based on the similarity.

[0112] As described above, according to this embodiment, it is possible to propose an extension idea that reflects the instruction of the policy for an input idea, and to show similar ideas for the input idea or the extension idea. It is also possible to extend the input idea based on similar ideas. [Explanation of symbols]

[0113] 1: Idea expansion device 11: Idea Reception Department 12: Policy Reception Department 13: Extension 14: Dialogue section 15:Extraction part 16: Output section 17: Drawing Creation Department 2: User terminal 3:Generation device 10: Information processing device 101: Control unit 102: Storage section 103: Communications Department 9: Terminal device 901: Control unit 902: Storage section 903: Communications Department 904: Input section 905: Output section NW: Network

Claims

1. An idea expansion system including an idea receiving unit, an extraction unit, an output unit, a policy receiving unit, and an expansion unit, the idea receiving unit receives an idea input indicating a user's idea; the extraction unit refers to a reference idea storage unit that stores publications in which existing reference ideas are described, based on the idea input, and extracts similar publications in which ideas similar to the idea are described; the output unit presents the list of similar publications to a user; the list includes at least one of the names or titles of the inventions of the similar publications, abstracts of the similar publications, and at least a portion of the claims of the similar publications, and a selection field for inputting a selection of the similar publications; the policy receiving unit receives a designation of a selected publication for expanding the idea from the similar publications from the user via the selection field; the extension unit performs an input to a generative model including an instruction to output an extended idea obtained by extending the idea based on the idea input and the selected publication; the generative model outputs the expansion ideas based on the content of the selected publications; The extension unit obtains the extended idea by extending the idea based on the output of the generative model.

2. The policy receiving unit receives a policy input indicating a policy for expanding an idea, the extension unit performs input to the generative model based on the idea input and the policy input, and acquires a first extended idea by extending the idea based on an output of the generative model; the output unit presents a proposal screen showing the first extension idea to a user; the proposal screen includes a similar idea extraction button for instructing extraction of ideas similar to the first extended idea, the idea receiving unit further receives the first extended idea as a new idea input in response to an input via the similar idea extraction button, the output unit presents to the user a list of the similar publications extracted based on the new idea input; the extension unit performs an input to the generative model including an instruction to output a second extension idea obtained by extending the first extension idea based on the first extension idea and the selected publication designated via the selection field in the list of similar publications based on the first extension idea; The idea expansion system according to claim 1 , wherein the expansion unit obtains the second expansion idea based on an output of the generative model.

3. The expansion unit accepts designation of the plurality of selected publications via the selection field, The idea expansion system according to claim 1 , wherein the expansion unit provides input to the generative model including an instruction to output the expanded idea obtained by expanding the idea based on the idea input and the plurality of selected publications.

4. the idea receiving unit receives the idea input in natural language; 2. The idea expansion system according to claim 1, wherein the expansion unit inputs an instruction statement, which is expressed in natural language as an instruction to expand the idea, to the generative model based on the idea input and the policy input.

5. the policy receiving unit receives the policy input by having a user select from a plurality of options; 3. The idea expansion system according to claim 2, wherein the expansion unit inputs an instruction statement, which is expressed in natural language as an instruction to expand the idea and includes the idea input and a word based on the policy input, into the generative model.

6. The idea expansion system according to claim 5 , wherein the options include at least one of a user, a purpose, a problem, a solution method, and a competition.

7. Further comprising a dialogue unit, the dialogue unit determines statements to be made by a plurality of people including a facilitator and an idea person, and conducts a dialogue among three or more people including the user; the dialogue unit determines, based on the idea input, a statement that prompts the user to input a policy as a statement from the facilitator; The idea expansion system according to claim 2 , wherein the policy receiving unit receives a statement from the user as the policy input.

8. The idea expansion system according to claim 7 , wherein the dialogue unit outputs the expansion idea as a statement by the idea person.

9. 8. The idea expansion system according to claim 7, wherein the dialogue unit displays speech histories of the facilitator, the idea person, and the user in chronological order in association with the speaker.

10. the extension unit further inputs, in response to a user's utterance, an instruction to output a sentence summarizing the extension idea based on the utterance histories of the plurality of people to the generative model, and transfers the output of the generative model to the dialogue unit; The idea expansion system according to claim 7 , wherein the dialogue unit outputs the expansion idea based on an output of the generative model.

11. the idea receiving unit receives the idea input in natural language; The similar publication describes the referenced idea in writing, The idea expansion system according to claim 1 , wherein the extraction unit extracts the similar publications by calculating a similarity between a part of the similar publication and the idea input.

12. The idea expansion system of claim 1 , wherein the similar publication is a patent document.

13. An idea expansion method executed by a computer including an idea receiving unit, an extraction unit, an output unit, a policy receiving unit, and an expansion unit, the idea receiving unit receives an idea input indicating a user's idea; the extraction unit refers to a reference idea storage unit that stores publications in which existing reference ideas are described, based on the idea input, and extracts similar publications in which ideas similar to the idea are described; the output unit presents the list of similar publications to a user; the list includes at least one of the names or titles of the inventions of the similar publications, abstracts of the similar publications, and at least a portion of the claims of the similar publications, and a selection field for inputting a selection of the similar publications; the policy receiving unit receives a designation of a selected publication for expanding the idea from the similar publications from the user via the selection field; the extension unit performs an input to a generative model including an instruction to output an extended idea obtained by extending the idea based on the idea input and the selected publication; the generative model outputs the expansion ideas based on the content of the selected publications; An idea extension method in which the extension unit obtains the extended idea by extending the idea based on the output of the generative model.

14. An idea expansion program that causes a computer to execute the idea expansion method according to claim 13.

Images