Information processing system, information processing method, and program

Abstract

The present technology relates to an information processing system, an information processing method, and a program that make it possible to improve the quality of idea creation. A persona for making ideas is set, and the persona is changed according to the situation. In addition, an idea generated on the basis of the persona is presented. The present technology can be applied, for example, when creating an idea from brain streaming or the like.

Description

Information Processing System, Information Processing Method, and Program

[0009] ,

[0008] , ,

[0007] , ,

[0001] This technology relates to an information processing system, an information processing method, and a program, and particularly to an information processing system, an information processing method, and a program that can improve the quality of idea creation.

[0002] For the purpose of providing an idea support system with excellent versatility and usefulness, knowledge is acquired by dynamically collecting information related to the phrases extracted from the remarks of the participants in a meeting from the Internet, and support information for assisting the ideas of the participants and / or the progress of the meeting is generated using that knowledge. A technology has been proposed (see, for example, Patent Document 1).

[0003] Japanese Patent Application Laid-Open No. 2021-056829

[0004] For example, in a place for idea creation such as brainstorming, there is a demand to improve the quality of idea creation.

[0005] This technology has been made in view of such a situation and is intended to improve the quality of idea creation.

[0006] The information processing system or program of this technology is an information processing system including a persona setting unit that sets a persona for idea generation, a persona change unit that changes the persona according to the situation, and a presentation control unit that presents ideas generated based on the persona, or a program for causing a computer to function as such an information processing system.

[0007] The information processing method of this technology is an information processing method including setting a persona for idea generation, changing the persona according to the situation, and presenting ideas generated based on the persona.

[0008] In this technology, a persona for idea generation is set. Then, the persona is changed according to the situation, and ideas generated based on the persona are presented.

[0009] An information processing system may be a single, independent device, or it may be an internal block comprising a single independent device. Furthermore, an information processing system can be composed of multiple independent devices.

[0010] The program can be provided by recording it on a recording medium or by transmitting it via a transmission medium.

[0011] This is a block diagram illustrating an example configuration of an embodiment of an information processing system to which this technology is applied. This is a diagram illustrating an overview of idea generation-related processing. This is a diagram illustrating an overview of narrowing-down-related processing. This is a diagram illustrating an example of an initial screen displayed on the UI unit 13 for initial setup of brainstorming. This is a diagram illustrating an example of a prompt given to the LLM at the start of idea generation by a persona (input to the LLM), and the persona ideas generated in the LLM as a result of that prompt (output of the LLM). This is a diagram illustrating an example of a prompt given to the LLM during the repetition of idea generation by a persona (input to the LLM), and the persona ideas generated in the LLM as a result of that prompt (output of the LLM). This is a diagram illustrating an example of an idea generation result screen generated by the idea generation result screen generation unit 16. This is a diagram illustrating an example of generating a summary of the content of the discussion in brainstorming, performed by the summarization unit 19. This is a block diagram illustrating an example configuration of the discussion stagnation determination unit 20. This is a block diagram illustrating an example configuration of the persona generation unit 21. This is a block diagram illustrating an example configuration of the participant persona change unit 22. This is a diagram illustrating an example of a direction narrowing screen generated by the LLM direction narrowing unit 31. This figure shows an example of a prompt for narrowing down directions (input to LLM) and candidate directions generated in LLM in response to that prompt (output of LLM). This figure shows an example of an idea narrowing screen generated by the idea narrowing screen generation unit 32. This figure shows an example of a prompt for narrowing down ideas (input to LLM) and the result of idea narrowing generated in LLM in response to that prompt (output of LLM). This is a flowchart illustrating an example of processing by the information processing system 1. This is a flowchart illustrating an example of processing by the information processing system 1. This is a block diagram showing an example configuration of one embodiment of a computer to which this technology is applied.

[0012] <Information processing system applying this technology>

[0013] Figure 1 is a block diagram showing an example configuration of one embodiment of an information processing system to which this technology is applied.

[0014] Information processing system 1 supports idea generation, for example, through methods such as brainstorming. Support for idea generation may include, if necessary, the provision of a forum for idea generation, such as facilitating brainstorming meetings. Information processing system 1 can be used to support idea generation in meetings conducted by brainstorming or any other method.

[0015] Information processing system 1 generates (sets) various personas in line with the theme (purpose) (including the category of the theme) of idea generation during brainstorming, for example. Information processing system 1 can change the personas depending on the situation, thereby improving the quality and efficiency of idea generation through brainstorming. For example, information processing system 1 can generate personas in response to user operations or using AI (artificial intelligence) models, such as LLM (large language model). Changing personas includes adding new personas and adding new personas and deleting some or all of the existing personas. Persona changes can be made in response to user operations (situations) that request a change in personas, or the status of the discussion (such as whether the discussion is stalled), etc.

[0016] Information processing system 1 can use LLM to allow various personas to participate in brainstorming. Furthermore, information processing system 1 can generate ideas using at least personas and produce a final completed product (final output) that aligns with the theme of the idea generation, just as in actual brainstorming.

[0017] Here, there are many opportunities to generate ideas with multiple people, such as in the development of new products, the launch of projects, and the creation of research proposals and intellectual property. One method of generating ideas is brainstorming with multiple people. Brainstorming, also known as group ideation, is a meeting method in which multiple members freely exchange opinions and generate new ideas (ideas, thoughts, etc.). In brainstorming, multiple people simultaneously generate ideas on a given theme. It is acceptable for ideas to be divergent during the brainstorming phase. After the brainstorming phase, the process involves narrowing down the ideas presented to determine which ones are suitable. In other words, in brainstorming, the quantity of ideas is prioritized first, and then the quality of the ideas is prioritized to narrow them down.

[0018] Brainstorming requires multiple people to discuss at the same time, and even when held online, it involves significant burdens in terms of securing participants and coordinating schedules. Furthermore, if there is a certain bias among the brainstorming participants, it can be difficult to ensure diversity of ideas.

[0019] Information processing system 1, by using LLM, can generate a large number of ideas from multiple perspectives in a short time, even without user (human) participation, or with only one user participating. Because LLM has been trained in advance using a large corpus, it possesses diverse concepts and knowledge. Therefore, in information processing system 1, by using personas generated with LLM to generate ideas and changing the personas as needed, the diversity of ideas can be ensured and the quality of idea generation can be improved. Furthermore, costs such as the number of personnel participating in brainstorming and the time spent on brainstorming can be reduced, improving the efficiency of idea generation.

[0020] Furthermore, Information Processing System 1 can, for example, encourage discussion when it has stalled. It can also narrow down ideas after brainstorming using multiple perspectives. Additionally, Information Processing System 1 can generate and present a summary of the brainstorming discussion to the user, allowing the user to conduct collaborative brainstorming sessions with other participants, including personas, without being constrained by time. For example, by referring to the summary, the user can understand the content of the discussions that took place during their absence and participate smoothly in the brainstorming session.

[0021] The information processing system 1 includes an information processing unit 10. The information processing unit 10 includes an LLM (large language model) storage unit 11, an initial persona setting unit 12, a UI (user interface) unit 13, an LLM input generation unit 14, an LLM idea generation unit 15, an idea generation result screen generation unit 16, an idea exploration unit 17, a history storage unit 18, a summarization unit 19, a discussion stagnation determination unit 20, a persona generation unit 21, and a participant persona modification unit 22. Furthermore, the information processing unit 10 includes an LLM direction narrowing unit 31, an idea narrowing screen generation unit 32, an idea narrowing unit 33, and a final output generation unit 34, etc. Some or all of the blocks from the LLM storage unit 11 to the participant persona modification unit 22, and from the LLM direction narrowing unit 31 to the final output generation unit 34, excluding the UI unit 13, can be provided in one or more separate devices.

[0022] Note that in Figure 1, some of the connection lines representing data exchange between the blocks constituting the information processing unit 10 have been omitted from the diagram to avoid making it too complex. The same applies to the diagrams described later.

[0023] The LLM memory unit 11 stores learned LLMs. The LLMs stored in the LLM can be updated using user input to the information processing system 1.

[0024] The initial persona setting unit 12 sets up a persona that will generate ideas. That is, the initial persona setting unit 12 sets up the persona that will first participate in the brainstorming as the initial persona. The initial persona setting unit 12 sets up one or more initial personas. The initial persona setting unit 12 supplies persona information about the initial personas to the LLM input generation unit 14.

[0025] The UI unit 13 functions as an input receiving unit that receives various user inputs, such as user ideas and information necessary for generating prompts to be given to the LLM, and as a presentation control unit that presents various information, such as ideas generated based on the persona, to the user. Information is presented to the user, for example, by displaying an image on a display or outputting sound from a speaker. The UI unit 13 receives user input and supplies it to the LLM input generation unit 14, the idea generation result screen generation unit 16, and other necessary blocks. The UI unit 13 also receives necessary information from the necessary blocks and presents that information.

[0026] The LLM input generation unit 14 is supplied with persona information relating to the persona that will generate ideas from the initial persona setting unit 12 and the participating persona modification unit 22. In addition, the LLM input generation unit 14 receives necessary information from other necessary blocks. For example, based on the persona information, the LLM input generation unit 14 generates prompts (prompts to be given (inputted) to the LLM in order to generate ideas) for the persona represented by that persona information, and supplies them to the LLM idea generation unit 15 and the LLM direction narrowing unit 31.

[0027] The LLM idea generation unit 15 processes ideas as persona-based idea generation. Specifically, the LLM idea generation unit 15 causes the LLM stored in the LLM storage unit 11 to generate ideas by giving prompts from the LLM input generation unit 14, which is performed as persona-based idea generation (generation of ideas by persona), and supplies the persona ideas generated by this idea generation to the idea generation result screen generation unit 16.

[0028] The Idea Generation Results Screen Generation Unit 16 generates an Idea Generation Results Screen that represents at least the results of idea generation by the persona, based on the persona's ideas from the LLM Idea Generation Unit 15. When a user participates in brainstorming and the user's ideas (input) are supplied to the Idea Generation Results Screen Generation Unit 16 from the UI Unit 13, the Idea Generation Results Screen Generation Unit 16 generates an Idea Generation Results Screen that represents the results of idea generation by the user and the persona, based on the persona's ideas from the LLM Idea Generation Unit 15 and the user's ideas from the UI Unit 13. When ideas begin to be generated from the user or persona, the Idea Generation Results Screen Generation Unit 16 generates an Idea Generation Results Screen and supplies it to the UI Unit 13 for presentation. The Idea Generation Results Screen Generation Unit 16 also supplies the Idea Generation Results Screen (and the information displayed on it) to the History Storage Unit 18 and the Discussion Stagnation Determination Unit 20.

[0029] The idea exploration unit 17 refers to the idea generation result screen generated by the idea generation result screen generation unit 16 and explores the ideas displayed on that screen in more detail. The idea exploration unit 17 supplies the explored ideas to the idea generation result screen generation unit 16, which then reflects them on the idea generation result screen.

[0030] The history storage unit 18 stores various types of information, such as the history of brainstorming. For example, the history storage unit 18 is supplied with the idea generation result screen from the idea generation result screen generation unit 16, the idea refinement screen (and the information displayed therein) from the idea refinement screen generation unit 32 (described later), and the refinement result file (and the information stored therein) from the final output generation unit 34 (described later). The history storage unit 18 acquires (extracts) and stores information that constitutes the history of brainstorming from the information obtained from the idea generation result screen, the idea refinement screen, and the refinement result file, respectively. In addition, the history storage unit 18 can also store information obtained from blocks other than the idea generation result screen generation unit 16, the idea refinement screen generation unit 32, and the final output generation unit 34, such as user input obtained from the UI unit 13, as needed.

[0031] For example, the history storage unit 18 stores the brainstorming theme, user information, persona information, and output format information entered by the user on the initial screen, which will be described later. The history storage unit 18 also stores the user's ideas and persona's ideas displayed on the idea generation results screen, as well as the direction for narrowing down ideas (candidate directions) displayed on the idea narrowing screen. In the history storage unit 18, each piece of information can be stored in association with the time it was entered or generated. Furthermore, in the history storage unit 18, ideas can be stored in association with the user or persona (or information identifying the user) who generated the idea. When the user inputs information by speaking (voice), the information processing system 1 performs speech recognition to convert the utterance into text, and that text can be stored in the history storage unit 18 in association with the time of the utterance and the user (or information identifying the user) who made the utterance.

[0032] The summarization unit 19 uses the brainstorming history stored in the history storage unit 18 to generate a summary of the content of the discussions during the brainstorming sessions conducted by the personas and users.

[0033] The discussion stagnation determination unit 20 determines whether the discussion is stagnant based on information obtained from the idea generation result screen from the idea generation result screen generation unit 16, such as user and persona ideas. The discussion stagnation determination unit 20 supplies the stagnation determination result, which is the result of determining whether the discussion is stagnant, to the persona generation unit 21.

[0034] The persona generation unit 21 generates one or more candidate personas that can be used as new participant personas, which are personas that will participate in the brainstorming session, and supplies them to the participant persona modification unit 22. The persona generation unit 21 generates candidate personas depending on the situation. For example, the persona generation unit 21 generates candidate personas in response to an operation of the UI unit 13 by a user that requests a change in the participant persona. Also, for example, the persona generation unit 21 generates candidate personas in response to the fact that the discussion has stalled, that is, the stallion determination result from the discussion stalling determination unit 20 indicates that the discussion has stalled.

[0035] The participant persona modification unit 22 modifies the participant personas participating in the brainstorming session. For example, the participant persona modification unit 22 selects one or more candidate personas from the candidate personas generated by the persona generation unit 21 to be added as new participant personas to the brainstorming session. The participant persona modification unit 22 also selects, if necessary, people to be removed from the brainstorming session from the list of existing participant personas. The participant persona modification unit 22 supplies the persona information of the new participant personas and the removed personas to the LLM input generation unit 14, causing the new participant personas to be added to the brainstorming session and the removed personas to be removed from the brainstorming session. Adding a persona to a brainstorming session means setting that persona as the persona that generates ideas. Removing a persona from brainstorming means excluding (excluding) that persona from the group of people involved in generating ideas.

[0036] The participant persona modification unit 22 can select some or all of the one or more candidate personas from the persona generation unit 21 as new participant personas.

[0037] The persona generation unit 21 generates candidate personas according to the situation, and the participant persona modification unit 22 modifies the participant personas by selecting new participant personas from the candidate personas. Therefore, it can be said that the participant persona modification unit 22 modifies the participant personas according to the situation.

[0038] The LLM direction refinement unit 31 refines the direction of idea refinement when narrowing down ideas based on persona or user ideas. For example, the LLM direction refinement unit 31 is supplied with prompts from the LLM input generation unit 14 to refine the direction of idea refinement. The LLM direction refinement unit 31 provides the prompts from the LLM input generation unit 14 to the LLM stored in the LLM storage unit 11, thereby narrowing down the direction of idea refinement (generating refined directions). The LLM direction refinement unit 31 supplies the refined directions as candidate directions, which are candidates for the direction of idea refinement, to the idea refinement screen generation unit 32.

[0039] The idea filtering screen generation unit 32 generates an idea filtering screen that displays candidate directions from the LLM direction filtering unit 31. The idea filtering screen generation unit 32 supplies the idea filtering screen to the UI unit 13 for display. The idea filtering screen generation unit 32 also supplies the idea filtering screen (and the information displayed on it) to the history storage unit 18 and the idea filtering unit 33.

[0040] The idea refinement unit 33 determines the direction for narrowing down ideas from the candidate directions displayed on the idea refinement screen generated by the idea refinement screen generation unit 32. The determined direction is also called the determined direction. The idea refinement unit 33 generates prompts for narrowing down ideas (prompts given (inputted) to the LLM in order to narrow down ideas) according to the determined direction.

[0041] The idea narrowing-down unit 33 performs idea narrowing-down by giving a prompt to the LLM stored in the LLM storage unit 11 to cause the LLM to generate the result of idea narrowing-down. The idea narrowing-down unit 33 supplies the result of idea narrowing-down generated by the idea narrowing-down to the final output generation unit 34.

[0042] The final output generation unit 34 generates a narrowing-down result file that displays the result of idea narrowing-down from the idea narrowing-down unit 33 in a predetermined output format as the final output of brainstorming. The final output generation unit 34 supplies the narrowing-down result file to the UI unit 13 for presentation. Further, the final output generation unit 34 supplies the narrowing-down result file to the history storage unit 18.

[0043] The information processing system 1 configured as described above performs brainstorming-related processing related to brainstorming for generating ideas and narrowing-down-related processing related to idea narrowing-down for narrowing down ideas. In the narrowing-down-related processing, the direction for narrowing down ideas is narrowed down, the direction for narrowing down ideas is determined from the narrowed-down direction (candidate direction), and ideas are narrowed down according to the determined direction (determined direction).

[0044] FIG. 2 is a diagram for explaining the outline of the brainstorming-related processing.

[0045] FIG. 2 shows a part related to the brainstorming-related processing in the information processing system 1.

[0046] In the brainstorming-related processing, importance is placed on quantity rather than the quality of ideas, and ideas regarding the theme of brainstorming are generated by at least a plurality of people including personas. In the present embodiment, for simplicity of explanation, an idea is assumed to be expressed in one or a few sentences.

[0047] The user inputs, by operating the UI unit 13 (performing an operation on the UI unit 13), the theme of brainstorming and output format information representing the output format (such as a format) for outputting the result of brainstorming.

[0048] Also, when the user participates in brainstorming, by operating the UI unit 13, the user inputs user information, which is various information about the user, including the user's age, gender, occupation, etc. Furthermore, the user inputs ideas by operating the UI unit 13. The input of the user's ideas can be performed by operating an operation means such as a keyboard, or can be input by voice. Also, in the input of ideas, in addition to ideas expressed in text, ideas partially or entirely expressed in arbitrary media data such as images and sounds can be input. The images and sounds expressing ideas can be generated by an arbitrary method, for example, a method using an image or sound generation model, etc.

[0049] Here, in the information processing system 1, even if a user (person) does not participate, brainstorming can be performed only with personas. However, in the following, it is assumed that the user participates and brainstorming is performed.

[0050] The UI unit 13 supplies user information, the theme of brainstorming, and output format information to the LLM input generation unit 14. Also, the UI unit 13 supplies the user's ideas to the idea output result screen generation unit 16.

[0051] The initial persona setting unit 12 or the participating persona change unit 22 supplies the persona information of the participating persona to the LLM input generation unit 14.

[0052] The LLM input generation unit 14 uses, as necessary, user information, the theme of brainstorming, and output format information from the UI unit 13, and persona information of the participating persona from the initial persona setting unit 12 or the participating persona change unit 22, etc., to generate a prompt to be given to the LLM. For example, based on the persona information and the theme of brainstorming, the LLM input generation unit 14 generates a prompt for causing the persona represented by the persona information to come up with (generate) ideas along the theme of brainstorming, and supplies it to the LLM idea generation unit 15.

[0053] The LLM idea generation unit 15 generates ideas for the LLM stored in the LLM memory unit 11 by providing prompts from the LLM input generation unit 14, which is called persona-based idea generation (generating ideas by persona). The LLM idea generation unit 15 supplies the persona ideas generated by persona-based idea generation to the idea generation result screen generation unit 16.

[0054] The idea generation results screen generation unit 16 generates an idea generation results screen that reflects the persona ideas from the LLM idea generation unit 15 and the user ideas from the UI unit 13, and supplies it to the UI unit 13 for presentation. Ideas displayed on the idea generation results screen are further explored by the idea exploration unit 17 as needed, and the explored content is reflected on the idea generation results screen.

[0055] In the history storage unit 18, information obtained from the idea generation results screen, etc., is stored as a history of the brainstorming session. Then, in the summarization unit 19, a summary of the discussions during the brainstorming session conducted by the persona and user is generated using the history stored in the history storage unit 18. The summary is supplied to the UI unit 13 and presented.

[0056] The discussion stagnation determination unit 20 determines whether the discussion is stagnant based on, for example, user and persona ideas obtained from the idea generation result screen generated by the idea generation result screen generation unit 16. If the discussion stagnation determination unit 20 determines that the discussion is stagnant, the persona generation unit 21 generates candidate personas, and the participant persona modification unit 22 selects a new participant persona to add to the brainstorming session from the candidate personas.

[0057] The participant persona modification unit 22 supplies the persona information of the new participant persona to the LLM input generation unit 14, thereby changing the participant persona by adding the new participant persona to the brainstorming session.

[0058] Then, the same process as described above is repeated as appropriate. Specifically, the LLM input generation unit 14 generates prompts to give to the LLM using the persona information of the modified participating personas as needed, and the LLM idea generation unit 15 gives the LLM the prompts generated by the LLM input generation unit 14, thereby causing the LLM to generate ideas, which is then performed as idea generation by the personas, and this process is repeated as appropriate.

[0059] Figure 3 is a diagram illustrating the overview of the filtering-related processing.

[0060] Figure 3 shows the part of the information processing system 1 that is related to filtering processes.

[0061] In the refinement process, the ideas generated in the idea generation process are examined, and a direction for narrowing down the ideas (direction of idea refinement) is determined in order to produce the final completed product (e.g., a project proposal). The final completed product is then generated according to that direction.

[0062] In the refinement-related processing, the user inputs constraints for narrowing down the direction of idea refinement by operating the UI unit 13. The UI unit 13 supplies the constraints to the LLM input generation unit 14.

[0063] Furthermore, in the filtering-related processing, the history storage unit 18 supplies the LLM input generation unit 14 with a list of ideas (idea list) that were generated in the idea generation-related processing from the information it stores.

[0064] The LLM input generation unit 14 uses constraints from the UI unit 13, and, as necessary, uses user information, brainstorming themes, and output format information supplied from the UI unit 13 during idea generation-related processing, as well as persona information of participating personas supplied from the initial persona setting unit 12 or the participating persona modification unit 22, and idea lists from the history storage unit 18 to generate prompts to be given to the LLM. For example, the LLM input generation unit 14 generates prompts to narrow down the direction of idea refinement according to the constraints from the UI unit 13. The prompts are supplied from the LLM input generation unit 14 to the LLM direction refinement unit 31.

[0065] The LLM direction selection unit 31 provides prompts from the LLM input generation unit 14 to the LLM stored in the LLM storage unit 11, thereby narrowing down the direction of idea selection according to the constraints entered by the user. The LLM direction selection unit 31 supplies the candidate directions obtained by narrowing down the direction of idea selection to the idea selection screen generation unit 32.

[0066] The idea refinement screen generation unit 32 generates an idea refinement screen that displays candidate directions from the LLM direction refinement unit 31 and supplies it to the UI unit 13 for presentation. The idea refinement screen generation unit 32 also supplies the idea refinement screen (and the information displayed on it) to the history storage unit 18 and the idea refinement unit 33.

[0067] The idea refinement unit 33 determines a final direction from the candidate directions displayed on the idea refinement screen generated by the idea refinement screen generation unit 32, and generates prompts for refining ideas according to that final direction.

[0068] In the idea refinement unit 33, the idea refinement is performed by prompting the LLM stored in the LLM storage unit 11, causing the LLM to generate the idea refinement results. The idea refinement unit 33 supplies the idea refinement results generated by the idea refinement to the final output generation unit 34.

[0069] The final output generation unit 34 generates a refinement result file that displays the idea refinement results from the idea refinement unit 33 in the output format indicated by the output format information. The final output generation unit 34 supplies the refinement result file to the UI unit 13 for display. The final output generation unit 34 also supplies the refinement result file to the history storage unit 18.

[0070] In the history storage unit 18, information obtained from the idea filtering screen generated by the idea filtering screen generation unit 32 and the filtering result file generated by the final output generation unit 34 is stored as the brainstorming history. Then, in the summarization unit 19, a summary of the brainstorming discussions conducted by the persona and user is generated using the history stored in the history storage unit 18. The summary is supplied to the UI unit 13 and presented.

[0071] <Initial screen>

[0072] Figure 4 shows an example of the initial screen displayed in the UI unit 13 for initial setup of brainstorming.

[0073] When the user performs an operation to initiate the initial setup for brainstorming, the UI unit 13 generates and displays the initial screen 50. The user then inputs the necessary information on the initial screen 50 by operating the UI unit 13.

[0074] The initial screen 50 includes a topic field 51, a user settings field 52, a persona field 53, an edit button 54, a delete button 55, an add button 56, an output format field 58, and a start button 59, among others.

[0075] In the topic field 51, the user enters the brainstorming theme. In Figure 4, "10-minute animation" (an animation of about 10 minutes) is entered as the brainstorming theme.

[0076] In the user settings field 52, the user information of the real person participating in the brainstorming session is entered by the user. In Figure 4, the user information entered is that the user is a man in his 50s and works for a TV (television) station.

[0077] In the Persona field 53, the user enters the persona information for the initial persona, that is, the hypothetical person who will be the first to participate in the brainstorming session (the persona that the LLM will play). In Figure 4, the persona information for the initial persona is entered as follows: the initial persona is a man in his 30s who works at an advertising agency.

[0078] The edit button 54 and the delete button 55 are located (displayed) to the right of the persona field 53. The edit button 54 is operated (clicked, touched, etc.) to edit the persona information in the persona field 53 to the left of the edit button 54. The delete button 55 is operated to delete the persona information (and the persona it represents) in the persona field 53 to the left of the delete button 55.

[0079] The add button 56 is located at the bottom of the persona section 53 and is operated when adding an initial persona (persona information). When the add button 56 is operated, an input field 57 for entering persona information is displayed. The input field 57 has a column displaying various items (attribute names) such as the gender and occupation of the initial persona, and a field for entering the content (attribute value) of each item. By operating the UI section 13, the user can input the persona information of the initial persona to be added by entering the content of each item in the input field 57. Once the persona information of the initial persona to be added is confirmed, the persona section 53 that displays that persona information, the edit button 54 and delete button 55 corresponding to that persona section 53 are displayed one row below the persona section 53 to the delete button 55, and the add button 56 moves one row below that.

[0080] The content of each item in input field 57 can be entered as text, or by selecting from options displayed in a pull-down menu, etc. Persona information can be entered by entering content in each item in input field 57, or by entering natural language to describe the (initial) persona. For example, natural language such as "A 30-something employee of an advertising agency working in Marunouchi. Spends their days off playing golf" can be entered as persona information.

[0081] On the initial screen 50, instead of entering persona information in the persona field 53, the user can select a persona to be used as the initial persona from the DB (database). For example, the DB can store personas that the user previously entered in the persona field 53, personas that other users have entered, and pre-created default personas (and their persona information). On the initial screen 50, the personas stored in the DB are displayed, and the user can select a persona to be used as the initial persona from among them.

[0082] The output format field 58 is used to input the output format for the idea refinement results. In Figure 4, the output format field 58 is a pull-down menu, allowing the user to select the output format for the idea refinement results from various document formats, such as project proposals, patent applications, and academic papers.

[0083] The start button 59 is operated to begin brainstorming. The start button 59 becomes operable once the necessary information for brainstorming has been entered on the initial screen.

[0084] In the initial persona setting unit 12, the persona represented by the persona information entered in the persona field 53 is set as the initial persona (identified as the initial persona).

[0085] <Idea generation using personas>

[0086] Figure 5 shows an example of the prompt given to the LLM at the start of persona-based idea generation (input to the LLM), and the persona ideas generated by the LLM in response to that prompt (output to the LLM).

[0087] Figure 5A shows an example of an input to the LLM as a prompt generated by the LLM input generation unit 14.

[0088] In Figure 5A, a prompt is generated that reflects the brainstorming theme, user information, and participant persona information, stating, "We are brainstorming based on the following theme. Please present several ideas based on the personas given below..." It is assumed that there are two participant personas (initial personas), Persona A and Persona B.

[0089] Figure 5B shows an example of LLM output as a persona idea, which is generated when the LLM idea generation unit 15 provides the LLM with prompts generated by the LLM input generation unit 14.

[0090] Figure 5B shows the LLM output as ideas for two participant personas, Persona A and Persona B, generated when the prompt in Figure 5A is given in LLM.

[0091] Figure 6 shows an example of prompts given to the LLM (input to the LLM) during the iterative process of generating ideas using personas, and the persona ideas generated by the LLM in response to those prompts (output to the LLM).

[0092] In Figure 6, as in Figure 5, there are two participant personas, Persona A and Persona B.

[0093] Figure 6A shows an example of an input to the LLM as a prompt generated by the LLM input generation unit 14.

[0094] In Figure 6A, the brainstorming theme, user information, participant persona information, and already generated ideas are reflected, and a prompt is generated requesting new ideas: "We are brainstorming based on the following theme. The following ideas have already been generated. Please provide several new ideas based on the personas given below..."

[0095] Figure 6B shows an example of LLM output as a persona idea, which is generated when the LLM idea generation unit 15 provides the LLM with prompts generated by the LLM input generation unit 14.

[0096] Figure 6B shows the LLM output as ideas for two participant personas, Persona A and Persona B, respectively, generated when the prompt in Figure 6A is given in LLM.

[0097] As described above, the LLM input generation unit 14 generates a prompt requesting new ideas that reflect the ideas already generated. The LLM idea generation unit 15 then provides this prompt to the LLM, and, taking into account all the ideas already generated, generates new ideas as persona ideas.

[0098] <Idea generation results screen>

[0099] Figure 7 shows an example of an idea generation result screen generated by the idea generation result screen generation unit 16.

[0100] In Figure 7, the idea generation results screen 70 includes a brainstorming information section 71, as well as a set of idea images 72 and icons 73.

[0101] The brainstorming information section 71 displays the theme section 61, user settings section 62, persona settings section 63, and output format section 64, among others.

[0102] The theme field 61 displays the brainstorming theme entered in the topic field 51 on the initial screen 50 (Figure 4).

[0103] The user settings field 62 displays the user information entered in the user settings field 52 on the initial screen 50, along with an icon representing the user of that user information.

[0104] In the persona setting section 63, the persona information of participating personas (including personas that were participating personas as needed) is displayed along with an icon representing that persona.

[0105] The output format field 64 displays the output format of the idea refinement results entered in the output format field 58 on the initial screen 50.

[0106] Idea image 72 displays ideas submitted by participating personas or users. An icon 73 representing the participating persona or user who submitted the idea shown in idea image 72 is displayed alongside idea image 72. In Figure 7, the icon 73 representing the participating persona or user who submitted the idea shown in idea image 72 is displayed in the upper right corner of the rectangular idea image 72. By viewing this icon 73, it is possible to identify the user or persona who submitted the idea displayed in idea image 72, which is displayed together with the icon 73.

[0107] Idea image 72 can be displayed one per idea. Furthermore, idea image 72 can be displayed in a format that mimics the posting of sticky notes containing ideas, as is done in face-to-face brainstorming sessions.

[0108] On the idea generation results screen 70, for example, idea images 72 displaying ideas from users and each persona can be arranged in any order, for example, in the order in which the ideas were generated. Furthermore, on the idea generation results screen 70, the idea images 72 can be clustered and arranged. For example, the idea images 72 can be clustered and arranged according to the user or persona who generated the idea displayed in that idea image 72. Alternatively, for example, the idea images 72 can be clustered according to the content of the ideas, and similar idea images 72 can be grouped together. Clustering of idea images 72 according to the content of the ideas can be performed, for example, using LLM stored in the LLM storage unit 11.

[0109] Furthermore, on the idea generation results screen 70, the user can view the ideas displayed in each idea image 72, and if they come up with a new idea, they can input the new idea by operating the UI unit 13 and add an idea image 72 displaying that new idea. In addition, the user can specify several (multiple) idea images 72 that they are interested in and request new ideas related to the ideas displayed in the specified idea images 72. In this case, a prompt requesting new ideas related to the ideas displayed in the idea images 72 specified by the user is given to the LLM stored in the LLM storage unit 11, and a new idea is generated. Then, the idea image 72 displaying that new idea is added to the idea generation results screen 70.

[0110] On the idea generation results screen 70, users may wish to explore the ideas displayed in the idea image 72 further, that is, to concretize or expand upon the ideas.

[0111] Therefore, on the idea generation results screen 70, a context menu 74 can be displayed when the idea image 72 is manipulated. The context menu 74 includes a menu for requesting further exploration of the idea. When the user manipulates the context menu 74 to request further exploration of the idea, the idea exploration unit 17 takes the idea displayed on the idea image 72 on which the context menu 74 is displayed as the idea to be explored further and explores that idea further.

[0112] For example, the idea exploration unit 17 generates an idea to be explored, along with a prompt requesting that the idea be explored further, and provides this to the LLM stored in the LLM storage unit 11, thereby generating in-depth content of the idea to be explored further. The idea exploration unit 17 supplies the in-depth content to the idea generation result screen generation unit 16. The idea generation result screen generation unit 16 reflects the in-depth content on the idea generation result screen 70.

[0113] The idea generation results screen 70 in Figure 7 reflects (displays) the in-depth analysis content 75 obtained by further exploring the idea of ​​a certain persona, "Can be viewed during breaks between classes," which was selected as the idea to be explored in more detail.

[0114] Regarding the in-depth exploration of an idea, it can be performed by the persona who generated the idea (by generating a prompt), or it can be performed by a persona other than the one who generated the idea. Furthermore, the user can also perform the in-depth exploration. For example, the user can input the in-depth exploration content by operating the UI unit 13.

[0115] Furthermore, the idea generation results screen 70 can display messages to encourage the progress of the meeting. For example, if the user has not entered any ideas for a certain period of time, a message prompting them to enter ideas can be displayed. Alternatively, for example, a message can be displayed recommending that the user select an idea they are interested in and explore that idea in more detail.

[0116] <Generating a summary>

[0117] Figure 8 illustrates an example of the generation of a summary of the content of a brainstorming discussion performed by the summarization unit 19.

[0118] Figure 8 schematically shows the history of the first brainstorming session, for example, which was conducted on the theme of "10-minute animation." In Figure 8, the brainstorming history is represented by arranging the user and persona ideas in chronological order, along with icons representing the user or persona who came up with the idea.

[0119] Here, brainstorming sessions are often repeated multiple times, as a final conclusion may not be reached in a single session. The first brainstorming session refers to the first brainstorming session in a series of such sessions.

[0120] The summarization unit 19 can generate an overall summary 91, which is a summary of the entire content of the discussion in a single brainstorming session. Furthermore, the summarization unit 19 can generate a partial summary 92, which is a summary of a part of the content of the discussion in a single brainstorming session. For example, in the information processing system 1, if the user is away from their seat, a summary of the discussion content during the period the user was away (including a marginal period) can be generated as a partial summary 92.

[0121] A user may miss a brainstorming session. In this case, the user can understand the content of the discussion in the missed brainstorming session by referring to the overall summary 91. Also, if a brainstorming session is scheduled for the next time, and a to-do list (things to be done before the next brainstorming session) was decided in the brainstorming session the user missed, the user can understand the to-do list by referring to the overall summary 91. The user can then prepare for the next brainstorming session according to the to-do list.

[0122] Users may temporarily leave their seats during a brainstorming session. In this case, because the user is unaware of the content of the discussion while they were away, it may take time for them to grasp (understand) the current discussion after returning to the brainstorming session. Therefore, by referring to the partial summary 92, which summarizes the content of the discussion during the period when the user was away, they can grasp the content of the discussion while they were away and, based on that understanding, quickly grasp the content of the current discussion.

[0123] Furthermore, if there are multiple users participating in the brainstorming session, a partial summary 92 can be generated for each user.

[0124] As described above, users can understand the content of the discussions during periods when they are not present at the brainstorming session by referring to the overall summary 91 and the partial summaries 92. Therefore, users can understand the content of the discussions even if they only participate in the brainstorming session at times that are convenient for them, thus not being bound by the (session) time of the brainstorming.

[0125] <Example configuration of the discussion stagnation determination unit 20, the persona generation unit 21, and the participant persona modification unit 22>

[0126] Figure 9 is a block diagram showing an example configuration of the discussion stagnation determination unit 20.

[0127] In Figure 9, the discussion stagnation determination unit 20 includes a discussion stagnation detection module 101 and a determination module 102.

[0128] The discussion stagnation detection module 101 is supplied with user and persona ideas. Furthermore, the discussion stagnation detection module 101 is supplied with the content of the user's speech (speech content) as a result of speech recognition, as well as user sensing data such as the user's gaze, facial expressions, and biometric information obtained by sensing the user.

[0129] The discussion stagnation detection module 101 detects (calculates) the degree of stagnation in the brainstorming discussion (stagnation level) based on one or more of the user and persona's ideas, the user's utterances, and sensing data, and supplies this to the judgment module 102.

[0130] If no users offer any ideas or speak for several minutes or more, the discussion is likely stalled. Similarly, if users or personas start offering similar ideas repeatedly, reducing the diversity of ideas, or if users' concentration levels decline, the discussion is likely stalled.

[0131] Therefore, the discussion stagnation detection module 101 can calculate a degree of stagnation that is proportional to the duration of the silence, based on the length of time that silence continues, if a state of silence persists for a predetermined period of time or longer.

[0132] Furthermore, the discussion stagnation detection module 101 calculates a diversity index value representing the diversity of user and persona ideas generated within the most recent predetermined time, and based on this diversity index value, it can calculate a degree of stagnation that is inversely proportional to the diversity index value. Examples of diversity index values ​​that can be used include RemoteClique (RC value) and ChamferDistance (CD value).

[0133] In addition, the discussion stagnation detection module 101 can calculate a concentration index value representing the user's level of concentration based on the user's gaze, facial expressions, and biometric information, and based on that concentration index value, it can calculate a degree of stagnation that is inversely proportional to the concentration index value.

[0134] The determination module 102 determines whether the discussion in the brainstorming session is stalled based on the degree of stagnation reported by the discussion stagnation detection module 101, and supplies the stagnation determination result to the persona generation unit 21. For example, the determination module 102 determines that the discussion in the brainstorming session is stalled if the degree of stagnation is above a threshold, or if the degree of stagnation is above the threshold for a predetermined amount of time or number of consecutive times.

[0135] Figure 10 is a block diagram showing an example of the configuration of the persona generation unit 21.

[0136] In Figure 10, the persona generation unit 21 includes a prompt generation module 111 and a persona generation module 112.

[0137] The prompt generation module 111 receives the stagnation determination result from the determination module 102 of the discussion stagnation determination unit 20 (Figure 9). If the stagnation determination result indicates that the discussion is stagnant, the prompt generation module 111 generates a prompt to generate candidate personas that will become new participant personas for brainstorming, and supplies it to the persona generation module 112.

[0138] In Figure 10, the prompt generation module 111 reflects information such as the brainstorming theme, user information, participant persona information, and ideas already generated. Based on this information, a prompt is generated requesting a new persona: "Please suggest candidates for a new participant persona..."

[0139] The persona generation module 112 generates, for example, multiple candidate personas (persona information) by providing prompts from the prompt generation module 111 to the LLM stored in the LLM storage unit 11. The persona generation module 112 supplies the candidate personas (information representing them) along with the persona information to the participant persona modification unit 22.

[0140] Figure 11 is a block diagram showing an example of the configuration of the participant persona modification unit 22.

[0141] In Figure 11, the participant persona modification unit 22 has a selection module 121.

[0142] The selection module 121 is supplied with candidate personas from the persona generation module 112 of the persona generation unit 21 (Figure 10). The selection module 121 selects a new participant persona from the candidate personas from the persona generation module 112.

[0143] The candidate personas that LLM generates (persona information) may include a variety of personas. For example, it may include personas that are directly related to ideas about the brainstorming theme. It may also include personas that are as far removed as possible from existing personas who have already participated in the brainstorming (personas with very low attribute similarity). Furthermore, it may also include personas that are not too far removed from existing personas (personas with not too low attribute similarity).

[0144] From the perspective of ensuring diversity of ideas, it is desirable that new participant personas be as far removed as possible from the user and existing participant personas.

[0145] Therefore, the selection module 121 can select a new participating persona from the candidate personas, one whose attributes are as different as possible from those of the user and existing participating personas.

[0146] For example, the selection module 121 generates attribute vectors representing the attributes of the user, the already participating persona, and the candidate persona. For the user's attribute vector, for example, a (feature) vector obtained by transforming user information using an encoder of an AI model such as SentenceBERT can be used. Similarly, for the attribute vectors of the already participating persona and the candidate persona, vectors obtained by transforming persona information using an encoder of an AI model can be used.

[0147] The selection module 121 calculates a similarity score representing the similarity between the candidate persona and the users and existing participant personas, based on the attribute vectors. As the similarity score between the candidate persona and the users and existing participant personas, for example, a value corresponding to the reciprocal of the distance between the attribute vector of the candidate persona and the attribute vectors of the users and existing participant personas can be adopted.

[0148] The selection module 121 can select candidate personas as new participating personas based on similarity, selecting those with an average low similarity to the user and existing participating personas. For example, the selection module 121 can select the top N (>0) candidate personas with the lowest average similarity to the user and existing participating personas, or candidate personas whose average is below the threshold for persona selection, as new participating personas.

[0149] Furthermore, the method for selecting new participating personas from the candidate personas is not limited to the method described above. For example, new participating personas can be randomly selected from the candidate personas.

[0150] Furthermore, for example, a new participant persona can be selected from candidate personas in response to user input. That is, for example, the UI unit 13 can display persona information for candidate personas, and the user can refer to that persona information and select a candidate persona to be a new participant persona.

[0151] In addition, for example, new participant personas can be selected from candidate personas based on the brainstorming theme (the theme for generating ideas). For example, if the brainstorming theme (category) is related to anime production, candidate personas with anime-related attributes, such as those who are interested in anime or have knowledge of anime, can be selected as new participant personas.

[0152] The selection module 121 supplies persona information of a new participant persona to the LLM input generation unit 14, thereby changing the participant persona by adding the new participant persona to the brainstorming session.

[0153] Changes to participating personas can be made, for example, at the beginning of the next idea generation cycle, where the user and participating personas generate ideas a predetermined number of times (one or more times).

[0154] In the selection module 121, as part of changing the participant personas, new participant personas can be added to the brainstorming session, as well as existing participant personas can be removed (removed from the brainstorming session). In the selection module 121, the selection of existing participant personas to be removed from the brainstorming session can be done by any method.

[0155] For example, the selection module 121 can calculate the similarity of each already participating persona to both the user and other already participating personas. The selection module 121 can then select a predetermined number of already participating personas with high average similarity scores to the user and other already participating personas, or already participating personas whose average score is above a threshold for persona deletion, as personas to be deleted.

[0156] Furthermore, for example, the selection module 121 can calculate the similarity between each existing participant persona and each new participant persona. The selection module 121 can then select a predetermined number of existing participants personas with the highest average similarity scores to each new participant persona, or existing participants personas whose average similarity score is above a threshold for persona deletion, as the personas to be deleted.

[0157] In addition, for example, the selection module 121 can select previously participating personas who have spent more than a predetermined amount of time participating in brainstorming as personas to be deleted, or it can select personas to be deleted in response to user actions.

[0158] In the aforementioned case, if the brainstorming discussion stalls, candidate personas are generated, and the participating personas are changed by selecting a new participating persona from the candidate personas. Changes to participating personas can also be made in response to user actions, such as requests for changes to participating personas.

[0159] Furthermore, changes to participant personas should not be made immediately when brainstorming discussions stall. Instead, the change should be proposed to the user, and only made after the user has accepted the proposal.

[0160] <Direction filtering screen>

[0161] Figure 12 shows an example of a directionality filtering screen generated by the LLM directionality filtering unit 31.

[0162] In the information processing system 1, for example, if a user performs an operation to narrow down ideas targeting personas or user ideas, the system will perform idea narrowing. In narrowing down ideas, first, the LLM direction narrowing unit 31 narrows down the direction of idea narrowing.

[0163] The LLM direction filtering unit 31 generates a direction filtering screen 130 for setting (inputting) constraints for narrowing down the direction of idea filtering. The direction filtering screen 130 is supplied from the LLM direction filtering unit 31 to the UI unit 13 and presented.

[0164] Figure 12 shows an example of the direction selection screen 130.

[0165] In Figure 12, the direction selection screen 130 includes a brainstorming information section 131, an idea list 132, a constraints section 133, and a selection button 134.

[0166] The brainstorming information section 131 is configured in the same way as the brainstorming information section 71 on the idea generation results screen 70 (Figure 7).

[0167] Idea list 132 is a list of ideas generated from users and personas during previous idea generation sessions, and may include the details of any further in-depth analysis.

[0168] The constraint field 133 is where the constraints for narrowing down the direction of idea refinement are entered. These constraints can be entered, for example, by the user operating the UI unit 13. In other words, the constraints can be set according to the user's actions.

[0169] The filter button 134 is operated when narrowing down the direction of user and persona ideas in the idea list 132.

[0170] On the direction refinement screen 130, the user enters constraint conditions in the constraint field 133 and operates the refine button 134. When the refine button 134 is operated, the LLM direction refinement unit 31 refines the directions for narrowing down the user's and persona's ideas according to the constraint conditions entered in the constraint field 133, and generates candidate directions, which are candidates for narrowing down the ideas.

[0171] For example, the LLM direction narrowing unit 31 requests the LLM input generation unit 14 to generate a prompt for narrowing down the direction of an idea (hereinafter also referred to as direction narrowing). The LLM input generation unit 14 generates a prompt for direction narrowing in response to the request from the LLM direction narrowing unit 31 and supplies it to the LLM direction narrowing unit 31. The LLM direction narrowing unit 31 provides the prompt from the LLM input generation unit 14 to the LLM stored in the LLM storage unit 11, thereby generating candidate directions, which are candidates for the direction in which to narrow down an idea, as direction narrowing, and supplies them to the idea narrowing screen generation unit 32.

[0172] Figure 13 shows an example of a prompt for narrowing down directions (input to the LLM) and the candidate directions generated by the LLM in response to that prompt (output of the LLM).

[0173] Figure 13A shows an example of an input to the LLM as a prompt for directional narrowing, which is generated by the LLM input generation unit 14.

[0174] In Figure 13A, in addition to the constraints entered in the constraint field 133, the brainstorming theme, the persona information of the persona that generated the ideas, and the ideas generated by the user and persona are reflected, and a prompt is generated that asks for guidance on how to narrow down the ideas: "...please suggest a way to summarize...". Other information, such as user information and output format information, can be reflected in the prompt.

[0175] Figure 13B shows an example of LLM output as a candidate direction (description) generated when the LLM direction narrowing unit 31 provides the LLM with prompts generated by the LLM input generation unit 14.

[0176] The prompt in Figure 13A reflects the constraint condition "Please combine in two directions." Therefore, in Figure 13B, two candidate directions, direction 1 and direction 2, are generated in accordance with the constraint condition "Please combine in two directions."

[0177] <Idea Selection Screen>

[0178] Figure 14 shows an example of an idea filtering screen generated by the idea filtering screen generation unit 32.

[0179] For example, when candidate directions are supplied from the LLM direction selection unit 31 to the idea selection screen generation unit 32, the idea selection screen generation unit 32 displays the candidate directions and generates an idea selection screen 150 for narrowing down ideas. The idea selection screen 150 is supplied from the idea selection screen generation unit 32 to the UI unit 13 and presented.

[0180] Figure 14 shows an example of the idea refinement screen 150.

[0181] In Figure 14, the idea refinement screen 150 includes a direction refinement information field 151, a direction refinement result information field 152, a back button 153, and a create button 154.

[0182] The direction filtering information section 151 displays information from the direction filtering screen 130 (Figure 12), excluding the filtering button 134.

[0183] The direction selection result information section 152 displays the candidate directions. In Figure 14, the two candidate directions, direction 1 and direction 2, shown in B of Figure 13, are displayed.

[0184] For example, by operating the UI unit 13, the user can select a direction to narrow down their ideas from the candidate directions displayed in the direction narrowing result information field 152. For example, the user can select all the candidate directions displayed in the direction narrowing result information field 152 by checking the checkboxes displayed in the direction narrowing result information field 152. The idea narrowing unit 33 confirms the candidate direction selected by the user from among the candidate directions displayed in the direction narrowing result information field 152 as the confirmed direction, which is the direction to narrow down the ideas. The user can also select candidate directions that they do not want to be confirmed. In this case, the idea narrowing unit 33 confirms the candidate directions that the user did not select from among the candidate directions displayed in the direction narrowing result information field 152 as the confirmed direction. Therefore, if the user does not select any of the candidate directions displayed in the direction narrowing result information field 152, all of the candidate directions displayed in the direction narrowing result information field 152 are confirmed as the confirmed direction.

[0185] As a confirmed direction, one or more candidate directions can be selected, and therefore, some or all of the candidate directions can be selected. If multiple candidate directions are selected as the confirmed direction, and there are conflicting directions among them, the UI unit 13 can warn the user and prompt them to remove one of the conflicting candidate directions from the selection of confirmed directions.

[0186] The back button 153 is operated when generating candidate directions, that is, when the direction narrowing process is restarted. When the back button 153 is operated, the UI unit 13 displays the direction narrowing screen 130 (Figure 12) instead of the idea narrowing screen 150. This allows the user to change the constraints and restart the direction narrowing process as needed.

[0187] The create button 154 is operated when narrowing down ideas according to the confirmed direction. When the create button 154 is operated, the idea narrowing unit 33 narrows down user and persona ideas according to the confirmed direction.

[0188] For example, the idea refinement unit 33 generates prompts for idea refinement according to the determined direction. The idea refinement unit 33 performs idea refinement by giving prompts for idea refinement to the LLM stored in the LLM storage unit 11, thereby causing the LLM to generate the result of idea refinement. The idea refinement unit 33 supplies the result of idea refinement generated by idea refinement to the final output generation unit 34.

[0189] The final output generation unit 34 generates a refinement result file as the final output of the brainstorming session, which displays the results of the idea refinement in the output format indicated by the output format information. In Figure 14, a project proposal is specified as the output format, and a refinement result file in the format of a project proposal is generated. The refinement result file is supplied from the final output generation unit 34 to the UI unit 13 and presented.

[0190] Figure 15 shows an example of a prompt for idea refinement (input to LLM) and the result of idea refinement generated in LLM based on the given prompt (output of LLM).

[0191] Figure 15A shows an example of an input to the LLM as a prompt for idea refinement, generated by the idea refinement unit 33.

[0192] In Figure 15A, in addition to the confirmed direction, the brainstorming theme, ideas generated by the user and persona, output format information, etc. are reflected, and a prompt is generated requesting that the ideas be narrowed down and output be generated according to the output format indicated by the output format information, saying, "...Please generate output according to the output format..."

[0193] Figure 15B shows an example of the output of the LLM as a result of idea refinement, which is generated when the idea refinement unit 33 provides prompts for idea refinement to the LLM.

[0194] For example, the prompt reflects output format information representing the project proposal, and in Figure 15B, the items that should be included in the proposal are generated as a result of the idea refinement process.

[0195] <Processing by Information Processing System 1>

[0196] Figure 16 is a flowchart illustrating an example of the processing of information processing system 1. Figure 17 is a flowchart that follows Figure 16.

[0197] In step S10, the UI unit 13 generates and displays the initial screen 50 (Figure 4), waits for user input on the initial screen 50, and then proceeds to step S11.

[0198] In step S11, the information processing system 1 sets the brainstorming theme. That is, the information processing system 1 sets (stores) the user's input in the topic field 51 on the initial screen 50 as the brainstorming theme. After that, the process proceeds from step S11 to step S12.

[0199] In step S12, the information processing system 1 sets the output format. That is, the information processing system 1 sets the user's input in the output format field 58 of the initial screen 50 as output format information representing the output format. After that, the process proceeds from step S12 to step S13.

[0200] In step S13, the information processing system 1 sets the participating users, who are users participating in the brainstorming session. That is, the information processing system 1 sets the user information of the participating users as the user input in the user setting field 52 on the initial screen 50. After that, the process proceeds from step S13 to step S14.

[0201] In step S14, the initial persona setting unit 12 sets the initial persona, which is the first participant persona to participate in the brainstorming session. That is, the initial persona setting unit 12 sets the user's input in the persona field 53 of the initial screen 50 as the persona information of the participant persona (initial persona).

[0202] The brainstorming theme, output format information, user information of participating users, and persona information of participating personas set in steps S11 to S14 are supplied to the necessary blocks in the information processing system 1.

[0203] Subsequently, when the user operates the start button 59 on the initial screen 50, the process proceeds from step S14 to step S15.

[0204] In step S15, the process of generating ideas is carried out as described in steps S15-1 and S15-2, and the process proceeds to step S16.

[0205] In step S15-1, the UI unit 13 waits for the participating user to input an idea, accepts the input, and supplies the input idea to the idea generation result screen generation unit 16.

[0206] In step S15-2, ideas from the participating personas are generated. Specifically, in step S15-2, first in step S15-21, the LLM input generation unit 14 generates prompts for idea generation by the personas represented by the persona information of the participating personas, based on the persona information of the participating personas. The LLM input generation unit 14 supplies the prompts to the LLM idea generation unit 15, and the process proceeds from step S15-21 to step S15-22. In step S15-22, the LLM idea generation unit 15 provides the LLM with prompts from the LLM input generation unit 14, causing the LLM to generate ideas, which is considered idea generation by the participating personas (idea generation). The LLM idea generation unit 15 supplies the ideas from the participating personas generated by the idea generation by the participating personas to the idea generation result screen generation unit 16.

[0207] In step S16, the idea generation result screen generation unit 16 generates an idea generation result screen 70 (Figure 7) that displays ideas from participating users from the UI unit 13 and persona ideas from the LLM idea generation unit 15. The idea generation result screen 70 is supplied from the idea generation result screen generation unit 16 to the UI unit 13 and presented.

[0208] Subsequently, the process proceeds from step S16 to step S17, where the history storage unit 18 stores various information such as the brainstorming history, and the process proceeds to step S18. The history storage unit 18 stores the brainstorming theme, user information, persona information, and output format information entered on the initial screen 50 (Figure 4). The history storage unit 18 also stores the ideas and in-depth discussion content displayed on the idea generation results screen 70 (Figure 7). The history storage unit 18 stores other necessary information at the necessary timing, but the explanation of the history storage unit 18 storing information will be omitted in the following flowchart explanation.

[0209] In step S18, the discussion stagnation determination unit 20 determines whether the discussion is stagnant. If it determines that the discussion is not stagnant, the process skips steps S19 to S21 and proceeds to step S22.

[0210] Furthermore, if it is determined in step S18 that the discussion has stalled, the process proceeds to step S19, where the persona generation unit 21 determines whether or not to change the participating personas.

[0211] If it is determined in step S19 that the participating persona should not be changed, for example, if the user takes action to avoid changing the participating persona, the process skips steps S20 and S21 and proceeds to step S22.

[0212] Furthermore, if it is determined in step S19 that the participating persona should be changed, for example, if the user performs an operation to change the participating persona, the process proceeds to step S20.

[0213] In step S20, the persona generation unit 21 generates candidate personas and supplies them to the participating persona modification unit 22, and the process proceeds to step S21.

[0214] In step S21, the participant persona modification unit 22 modifies the participant personas, and the process proceeds to step S22. For example, the participant persona modification unit 22 selects one or more candidate personas from the candidate personas from the persona generation unit 21 to be the new participant personas. The participant persona modification unit 22 also selects, if necessary, personas to be removed from the brainstorming session from the existing participant personas who are already participating in the brainstorming session. The participant persona modification unit 22 supplies the persona information of the new participant personas and the personas to be removed to the LLM input generation unit 14, causing the new participant personas to be added to the brainstorming session and the personas to be removed to be removed from the brainstorming session.

[0215] In step S22, the information processing system 1 determines whether to continue brainstorming.

[0216] If it is determined in step S22 that the idea generation will continue, for example, if the user has not taken any action to end the idea generation, the process returns to step S15 and the same process is repeated.

[0217] On the other hand, if it is determined in step S22 that the idea generation will not continue, for example, if the user ends the idea generation and performs an operation to narrow down the ideas, the process proceeds to step S31 in Figure 17.

[0218] Here, the processes in steps S15 to S22 in Figure 16 are idea generation-related processes.

[0219] In step S31 of Figure 17, the LLM direction filtering unit 31 generates the direction filtering screen 130 (Figure 12), and the process proceeds to step S32. The direction filtering screen 130 is supplied from the LLM direction filtering unit 31 to the UI unit 13 and presented.

[0220] In step S32, the information processing system 1 sets constraint conditions. That is, the information processing system 1 waits for user input in the constraint field 133 of the direction narrowing screen 130 and sets (stores) that input as constraint conditions. After that, the process proceeds from step S32 to step S33.

[0221] In step S33, the processes of steps S33-1 and S33-2 are carried out as a process to narrow down the direction of idea refinement, and the process proceeds to step S34.

[0222] In step S33-1, the UI unit 13 waits for input from participating users regarding the direction of idea refinement, accepts that direction input, and supplies the input direction as a candidate direction to the idea refinement screen generation unit 32.

[0223] In step S33-2, the directions are narrowed down according to the constraints set in step S32. Specifically, in step S33-2, first in step S33-21, the LLM input generation unit 14 generates a prompt for narrowing down the directions based on the constraints, etc., and supplies it to the LLM idea generation unit 15. Then, the process proceeds from step S33-21 to step S33-22. In step S33-22, the LLM direction narrowing unit 31 provides the LLM with the prompt from the LLM input generation unit 14, causing it to generate narrowed directions according to the constraints. The LLM direction narrowing unit 31 supplies the narrowed directions as candidate directions to the idea narrowing screen generation unit 32.

[0224] In step S34, the idea refinement screen generation unit 32 generates an idea refinement screen 150 (Figure 14) that displays candidate directions entered by participating users from the UI unit 13, as well as candidate directions from the LLM direction refinement unit 31. The idea refinement screen 150 is supplied from the idea refinement screen generation unit 32 to the UI unit 13 and presented.

[0225] Subsequently, the process proceeds from step S34 to step S35, where the information processing system 1 determines whether to change the constraint conditions.

[0226] In step S35, if it is determined that the constraint conditions should be changed, for example, if the back button 153 on the idea refinement screen 150 is operated, the process returns to step S31. In this case, the direction refinement screen 130 (Figure 12) is displayed again. Therefore, the user can re-enter the constraint conditions in the constraint field 133 of the direction refinement screen 130.

[0227] Furthermore, if it is determined in step S35 that the constraint conditions will not be changed, for example, if the back button 153 on the idea refinement screen 150 has not been operated, the process proceeds to step S36.

[0228] In step S36, the information processing system 1 determines whether to return to the idea generation process.

[0229] In step S36, if it is determined to return to the idea generation phase, for example, if the user performs an action to return to the idea generation phase, the process returns to step S15 in Figure 16.

[0230] Furthermore, if it is determined in step S36 that the process will not return to idea generation, for example, if a final direction is determined from the candidate directions displayed on the idea refinement screen 150 in response to user operation, and then the user operates the create button 154 on the idea refinement screen 150 (Figure 14) to instruct the user to refine the ideas, the process proceeds to step S37.

[0231] In step S37, the processes of steps S37-1 and S37-2 are performed as an idea refinement process, and the process proceeds to step S38.

[0232] In step S37-1, the idea refinement unit 33 generates prompts for refining ideas according to the determined direction, and the process proceeds to step S37-2.

[0233] In step S37-2, the idea refinement unit 33 performs idea refinement by prompting the LLM to generate the idea refinement results. The idea refinement unit 33 supplies the idea refinement results generated by the idea refinement to the final output generation unit 34.

[0234] The final output generation unit 34 generates a refinement result file, which displays the idea refinement results from the idea refinement unit 33 in the output format indicated by the output format information, as the final output in accordance with the output format indicated by the output format information. The refinement result file as the final output is supplied to and presented to the UI unit 13.

[0235] In step S38, the information processing system 1 determines whether to finalize the output.

[0236] In step S38, if it is determined that the final output should not be confirmed, for example, if the user performs an operation that prevents the final output from being confirmed, the process returns to step S31 and the same process is repeated. In this case, as described above, the direction narrowing screen 130 (Figure 12) is displayed again, so the user can re-enter the constraint conditions in the constraint field 133 of the direction narrowing screen 130.

[0237] Therefore, if the user views the filtered result file as the final output and finds it does not meet their expectations, they can choose not to confirm the final output and start over from inputting the constraints.

[0238] Furthermore, in this case, by performing an operation in step S36 that follows, the user returns to the idea generation process in step S15 of Figure 16.

[0239] Therefore, if the user believes that the reason the filtered result file does not meet their expectations is due to a lack of ideas, they can return to the idea generation process in step S15 by performing an operation to return to the idea generation process. This allows them to add or modify ideas.

[0240] As described above, the information processing system 1 can, so to speak, move back and forth between the process of generating ideas and the process of narrowing down those ideas and generating a narrowed-down result file.

[0241] In step S38, if it is determined that the final output should be confirmed, for example, if the user performs an operation to confirm the final output, the process ends.

[0242] The processes in steps S31 to S38 of Figure 17 are related to the idea refinement process, which involves narrowing down the ideas.

[0243] The output format of the filtered result file, which is the final output, is not particularly limited. In this embodiment, the output format of the filtered result file is specified by the user by selecting it from a pull-down menu in the output format field 58 on the initial screen 50 (Figure 4). The output format can also be specified by, for example, specifying a file in a desired document format. Furthermore, the filtered result file can be a file containing any media data such as text, images, or sound. Images and sounds to be included in the filtered result file can be generated by any method, for example, by using an image or sound generation model.

[0244] <Description of a computer using this technology>

[0245] The series of processes described above can be executed by hardware or by software. When the series of processes are executed by software, the programs that make up that software are installed on a computer. Here, a computer includes computers built into dedicated hardware, as well as general-purpose personal computers, for example, that can perform various functions by installing various programs.

[0246] Figure 18 is a block diagram showing an example of the hardware configuration of a computer that executes the series of processes described above using a program.

[0247] In a computer, the processing circuit 901, ROM (Read Only Memory) 902, and RAM (Random Access Memory) 903 are interconnected by a bus 904.

[0248] An input / output interface 905 is further connected to the bus 904. An input unit 906, an output unit 907, a storage unit 908, a communication unit 909, and a drive 910 are connected to the input / output interface 905.

[0249] The input unit 906 may include physical or virtual operating means that the user operates to input information, such as a keyboard, mouse, or touch panel, as well as means that the user inputs information through voice, eye gaze, etc. Furthermore, the input unit 906 may include sensors for inputting various physical quantities to the computer. For example, the input unit 906 may include sensors that acquire physical quantities such as light (including infrared light other than visible light) or sound, such as a camera or microphone. Also, for example, the input unit 906 may include sensors that acquire other physical quantities such as temperature, moisture content, acceleration, distance, etc. The output unit 907 may include means that present information to the user by stimulating the user's perception, such as a display, speaker, or haptic device. The storage unit 908 is composed of a hard disk, non-volatile or volatile memory, etc., and stores various types of information (including programs). The communication unit 909 is a network interface, etc., and performs wired or wireless communication with the outside. The drive 910 drives removable media 911 such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory.

[0250] The processing circuit 901 includes a processor that executes programs such as a CPU (Central Processing Unit) and a DSP (Digital Signal Processor). The processing circuit 901 (its processor) performs the above-described series of processes by loading the program stored in the storage unit 908 into the RAM 903 via the input / output interface 905 and the bus 904 and executing it. The processing circuit 901 can output the processing results of the series of processes from the output unit 907 via the bus 904 and the input / output interface 905 as needed. The processing circuit 901 can also store the processing results in the storage unit 908 or transmit them from the communication unit 909.

[0251] The program executed by the computer (processing circuit 901) can be provided by recording it on a removable medium 911, such as a package medium. The program can also be provided via wired or wireless transmission media, such as a local area network, the internet, or digital satellite broadcasting.

[0252] In a computer, a program can be installed in the storage unit 908 via the input / output interface 905 by inserting a removable media 911 into the drive 910. Alternatively, a program can be received by the communication unit 909 from another device, such as a server, via a wired or wireless transmission medium, and installed in the storage unit 908. Furthermore, programs can be pre-installed in the ROM 902 or the storage unit 908.

[0253] The programs executed by the computer may be programs that are processed chronologically in the order described herein, or they may be programs that are processed in parallel or at necessary times, such as when a call is made.

[0254] The processes that a computer performs according to a program do not necessarily have to follow the order described in the flowchart. In other words, the processes that a computer performs according to a program include processes that are executed in parallel or individually (e.g., parallel processing and object-based processing).

[0255] The program may be processed by a single computer (processor), or it may be processed in a distributed manner by multiple computers. Furthermore, the program may be transferred to a remote computer and executed there.

[0256] When the above-described series of processes are performed by a computer executing a program, the processing circuit 901 (its processor) functions as the LLM storage unit 11 to the participant persona modification unit 22, and the LLM direction narrowing unit 31 to the final output generation unit 34, by executing the program.

[0257] In this specification, a system means one component or a collection of multiple components (devices, modules (parts), etc.). Therefore, one or more components of a computer, for example, only the processor, or a combination of the processor and memory (for example, only the processing circuit 901, or a combination of the processing circuit 901 to the bus 904, etc.), constitute a system. Regarding a collection of multiple components, it is not necessary whether all components reside in the same enclosure. Therefore, multiple devices housed in separate enclosures and connected via a network, or a single device containing multiple modules within a single enclosure, are all systems. Furthermore, for example, the entire computer, or a combination of a computer and other devices such as a server (not shown), also constitute a system.

[0258] Furthermore, the embodiments of this technology are not limited to those described above, and various modifications are possible without departing from the spirit of this technology.

[0259] For example, this technology can be configured as cloud computing, where a single function is shared and processed collaboratively by multiple devices via a network.

[0260] Furthermore, each step described in the flowchart above can be performed by a single device, or it can be divided and performed by multiple devices.

[0261] Furthermore, if a single step includes multiple processes, those processes can be executed by a single device or shared among multiple devices.

[0262] Furthermore, the effects described herein are merely illustrative and not limiting, and other effects may also occur.

[0263] Furthermore, this technology can take the following configuration.

[0264] <1> An information processing system comprising: a persona setting unit for setting a persona for generating ideas; a persona changing unit for changing the persona according to the situation; and a presentation control unit for presenting ideas generated based on the persona. <2> The information processing system according to <1>, wherein the presentation control unit presents the results of idea narrowing, which is performed on the ideas of the persona. <3> The information processing system according to <1> or <2>, wherein the persona changing unit changes the persona according to user operation. <4> The information processing system according to any one of <1> to <3>, wherein the persona changing unit changes the persona according to the status of the discussion. <5> The information processing system according to <4>, wherein the persona changing unit changes the persona when the discussion is stalled. <6> The information processing system according to any one of <1> to <5>, wherein the persona changing unit changes the persona by adding a new persona, or by adding a new persona and deleting an existing persona. <7> The information processing system according to <6>, further comprising a persona generation unit that generates candidate personas that are candidates for the new persona, wherein the persona modification unit selects the new persona from the candidate personas based on the theme of idea generation. <8> The information processing system according to <6>, further comprising a persona generation unit that generates candidate personas that are candidates for the new persona, wherein the persona modification unit selects the new persona from the candidate personas in response to user operation. <9> The information processing system according to <6>, further comprising a persona generation unit that generates candidate personas that are candidates for the new persona, wherein the persona modification unit selects the candidate persona with low similarity to an existing persona as the new persona. <10> The information processing system according to any one of <1> to <9>, wherein the persona idea is generated using an LLM (large language model).<11> An information processing system according to any one of <1> to <10>, further comprising an input receiving unit that receives input of a user's ideas, wherein the presentation control unit presents the results of an idea narrowing process performed on the persona's ideas and the user's ideas. <12> An information processing system according to <11>, further comprising a summarizing unit that generates a summary of the content of a discussion conducted by the persona and the user. <13> An information processing system according to <12>, wherein the summarizing unit generates a summary of the entire content of the discussion. <14> An information processing system according to <12> or <13>, wherein the summarizing unit generates a summary of a part of the content of the discussion. <15> An information processing system according to <14>, wherein the summarizing unit generates a summary of the content of the discussion during the period when the user was away from their desk. <16> An information processing system according to any one of <1> to <15>, wherein the direction for narrowing down ideas is narrowed down according to predetermined constraints, and an idea narrowing process is performed according to the determined direction determined from the narrowed-down direction. <17> An information processing system according to <16> wherein the constraints are set in accordance with user operations. <18> An information processing system according to any one of <1> to <17> further comprising an idea exploration unit for exploring ideas in more depth. <19> An information processing method comprising setting a persona for generating ideas, changing the persona according to the situation, and presenting ideas generated based on the persona. <20> A program for causing a computer to function as a persona setting unit for setting a persona for generating ideas, a persona changing unit for changing the persona according to the situation, and a presentation control unit for presenting ideas generated based on the persona.

[0265] 1 Information Processing System, 10 Information Processing Unit, 11 LLM Memory Unit, 12 Initial Persona Setting Unit, 13 UI Unit, 14 LLM Input Generation Unit, 15 LLM Idea Generation Unit, 16 Idea Generation Result Screen Generation Unit, 17 Idea Deepening Unit, 18 History Memory Unit, 19 Summarization Unit, 20 Discussion Stagnation Judgment Unit, 21 Persona Generation Unit, 22 Participating Persona Change Unit, 31 LLM Direction Narrowing Unit, 32 Idea Narrowing Screen Generation Unit, 33 Idea Narrowing Unit, 34 Final Output Generation Unit, 50 Initial Screen, 51 Topic Field, 52 User Settings Field, 53 Persona Field, 54 Edit Button, 55 Delete Button, 56 Add Button, 57 Input Field, 58 Output Format Field, 59 Start Button, 70 Idea Generation Result Screen 71 Brainstorming Information Section, 72 Idea Image, 73 Icon, 91 Overall Summary, 92 Partial Summary, 101 Discussion Stagnation Detection Module, 102 Judgment Module, 111 Prompt Generation Module, 112 Persona Generation Module, 121 Selection Module, 130 Direction Refinement Screen, 131 Brainstorming Information Section, 132 Idea List, 133 Constraint Section, 134 Refinement Button, 150 Idea Refinement Screen, 151 Direction Refinement Information Section, 152 Direction Refinement Result Information Section, 153 Back Button, 154 Create Button, 901 Processing Circuit, 902 ROM, 903 RAM, 904 Bus, 905 Input / Output Interface, 906 Input Section, 907 Output Section, 908 Memory Section, 909 Communications unit, 910 Drive, 911 Removable media

Claims

1. An information processing system including a persona setting unit for setting up a persona for generating ideas, a persona changing unit for changing the persona according to the situation, and a presentation control unit for presenting ideas generated based on the persona.

2. The information processing system according to claim 1, wherein the presentation control unit presents the results of an idea narrowing process performed on the persona's ideas.

3. The information processing system according to claim 1, wherein the persona changing unit changes the persona in response to user operations.

4. The information processing system according to claim 1, wherein the persona changing unit changes the persona according to the status of the discussion.

5. The information processing system according to claim 4, wherein the persona changing unit changes the persona when the discussion is stalled.

6. The information processing system according to claim 1, wherein the persona modification unit modifies the persona by adding a new persona, or by adding a new persona and deleting an existing persona.

7. The information processing system according to claim 6, further comprising a persona generation unit that generates candidate personas to be candidates for the new persona, wherein the persona modification unit selects the new persona from the candidate personas based on the theme of idea generation.

8. The information processing system according to claim 6, further comprising a persona generation unit that generates candidate personas to be candidates for the new persona, wherein the persona modification unit selects the new persona from the candidate personas in response to user operations.

9. The information processing system according to claim 6, further comprising a persona generation unit that generates candidate personas to be candidates for the new persona, wherein the persona modification unit selects the candidate personas with low similarity to existing personas as the new persona.

10. The information processing system according to claim 1, wherein the persona idea is generated using an LLM (large language model).

11. The information processing system according to claim 1, further comprising an input receiving unit for receiving user ideas, wherein the presentation control unit causes the persona ideas and the user ideas to present the results of an idea narrowing process performed to narrow down the ideas.

12. The information processing system according to claim 11, further comprising a summarization unit that generates a summary of the content of the discussions conducted by the persona and the user.

13. The information processing system according to claim 12, wherein the summarization unit generates a summary of the entire content of the discussion.

14. The information processing system according to claim 12, wherein the summarization unit generates a summary of a part of the content of the discussion.

15. The information processing system according to claim 14, wherein the summarization unit generates a summary of the content of the discussion during the period when the user was away from their seat.

16. An information processing system according to claim 1, wherein the direction for narrowing down ideas is narrowed down according to predetermined constraints, and the idea narrowing is performed according to the determined direction determined from the narrowed-down direction.

17. The information processing system according to claim 16, wherein the constraints are set in response to user operations.

18. The information processing system according to claim 1, further comprising an idea exploration unit for further exploring ideas.

19. An information processing method comprising setting up a persona for generating ideas, changing the persona according to the situation, and presenting ideas generated based on the persona.

20. A program for causing a computer to function as a persona setting unit for setting up a persona for generating ideas, a persona changing unit for changing the persona according to the situation, and a presentation control unit for presenting ideas generated based on the persona.

Images