System

Abstract

To provide a system.SOLUTION: A system includes: means for allowing a user to input regenerative AI models for discussing among rolls given stances of departments as roles, and a prompt sentence for designating the regenerative AI models to give opinions from respective viewpoints; means for allowing a server to receive the prompt sentence and select the appropriate regenerative AI models; means for sending the prompt sentence to each of the regenerative AI models selected by the server; means for allowing each of the regenerative AI models to start discussion and give an opinion from each viewpoint; means for allowing the server to collect the opinions of the regenerative AI models and integrating them as a discussion result; and means for returning the integrated discussion result to the user.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including a description and related instruction sentence regarding the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2022-180282 A Summary of the Invention [Problem to be solved by the invention]

[0004] In the mobile network-related department, numerous meetings are held daily between teams and departments to improve network quality and achieve the Ministry of Internal Affairs and Communications' objectives. However, these meetings involve human intervention, which increases man-hours and slows down discussions. There is also a possibility that some points may be overlooked. Furthermore, there are problems with the dependency of knowledge on individuals and handover. [Means for solving the problem]

[0005] The present invention provides a generation AI that discusses among roles (maintenance AI, construction AI, head office AI, etc.) that are given the stance of each department as a role. This generation AI identifies concerns and initiates discussions without human intervention, reducing labor hours, accelerating discussions, and preventing oversights. In addition, this system does not require the personalization or handover of knowledge. [Brief description of the drawings]

[0006] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Diagram 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. FIG. [Diagram 3] FIG. 11 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Diagram 5] FIG. 13 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6] FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 13 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 13 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 2 is a sequence diagram showing a process flow of the data processing system in the first embodiment of the first form example. [Figure 12] FIG. 11 is a sequence diagram showing a process flow of the data processing system in the application example 1 of the first embodiment. [Figure 13]FIG. 11 is a sequence diagram showing a process flow of the data processing system in the second embodiment of the second embodiment. [Figure 14] FIG. 13 is a sequence diagram showing a process flow of the data processing system in application example 2 of embodiment 2. [Figure 15] FIG. 13 is a sequence diagram showing a process flow of a data processing system in a third embodiment of the present invention. [Figure 16] FIG. 13 is a sequence diagram showing a process flow of the data processing system in the third application example of the third embodiment. [Figure 17] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in the first embodiment of the first form example when an emotion engine is combined. [Figure 18] FIG. 13 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1 of Form Example 1 when combined with an emotion engine. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0007] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0008] First, the terms used in the following description will be explained.

[0009] In the following embodiments, a signed processor (hereinafter simply referred to as a "processor") may be one arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be one type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), or a TPU (TENSOR PROCESSING UNIT (registered trademark)).

[0010] In the following embodiments, a signed RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by the processor.

[0011] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0012] In the following embodiments, a communication I / F (Interface) with a code is an interface including a communication processor and an antenna. The communication I / F controls communication between multiple computers. An example of a communication standard applied to the communication I / F is a wireless communication standard including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0013] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. In addition, in this specification, the same idea as "A and / or B" is also applied when three or more things are expressed by connecting them with "and / or."

[0014] [First embodiment]

[0015] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0016] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0017] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0018] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0019] The reception device 38 includes a touch panel 38A and a microphone 38B, and receives user input. The touch panel 38A detects contact with a pointer (e.g., a pen or a finger) to receive user input by the touch of the pointer. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0020] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (e.g., voice and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs voice according to instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a Complementary Metal-Oxide-Semiconductor (CMOS) image sensor or a Charge Coupled Device (CCD) image sensor.

[0021] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54.

[0022] FIG. 2 shows an example of main functions of the data processing device 12 and the smart device 14.

[0023] As shown in Fig. 2, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32. The specific process program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific process program 56 from the storage 32, and executes the read specific process program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific process program 56 executed on the RAM 30.

[0024] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0025] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores a reception output program 60. The reception output program 60 is used together with the specific processing program 56 by the data processing system 10. The processor 46 reads out the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0026] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0027] "Example 1"

[0028] The present invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0029] "Example 2"

[0030] As a concrete example, the maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies. These AIs hold discussions based on their respective stances and propose optimal solutions. This discussion is conducted without human intervention, and the results of the discussion are saved as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0031] The process flow of each embodiment will be described below.

[0032] "Example 1"

[0033] Step 1: The system generates AIs with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.).

[0034] Step 2: The generated AIs begin to discuss with each other. The content of the discussions will be related to improving network quality and achieving the Ministry of Internal Affairs and Communications' objectives.

[0035] Step 3: The results of the discussion are saved as meeting minutes and can be referenced later.

[0036] Step 1: The maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies.

[0037] Step 2: These AIs debate based on their respective stances and propose the optimal solution.

[0038] Step 3: The discussion is conducted without human intervention, and the results are stored as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0039] Example 1

[0040] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal".

[0041] In the conventional discussion system, it was difficult to hold discussions that reflected the stance of each department, and there was a problem that it took a lot of time to identify concerns and start discussions. In addition, the human intervention hindered the acceleration of discussions and increased the risk of overlooking something. Furthermore, the necessity of personalizing knowledge and handover made it difficult to hold efficient discussions.

[0042] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0043] In this invention, the server includes a generation AI means for discussion between roles that are given the stance of each department, a means for a user to input a prompt text, a means for the server to receive the prompt text and select an appropriate generation AI model, a means for the server to send the prompt text to the generation AI model selected by the server, a means for the generation AI model to start a discussion and for each AI to give an opinion from its own perspective, a means for the server to collect and integrate the opinions of each AI, a means for the server to return the integrated discussion result to the user, a means for identifying concerns and introducing discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This enables efficient and comprehensive discussions.

[0044] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0045] A "prompt sentence" refers to an instruction sentence regarding the topic or subject of discussion that is entered by the user.

[0046] "Server" refers to a computer system for receiving prompts, selecting an appropriate generative AI model, and managing the discussion.

[0047] A "generative AI model" refers to an artificial intelligence model with a specific role or perspective, such as maintenance AI, construction AI, or headquarters AI.

[0048] "Conservation AI" refers to a generative AI model that has a maintenance department stance.

[0049] "Construction AI" refers to a generative AI model that has a construction sector stance.

[0050] "Headquarters AI" refers to a generative AI model that has the stance of the headquarters department.

[0051] "Discussion results" refers to the conclusions and proposals reached through discussions between generative AI models.

[0052] "Reducing labor hours" refers to reducing the time and effort required for work.

[0053] "Accelerating the discussion" refers to speeding up the discussion.

[0054] "Preventing gaps in consideration" refers to making sure that important matters are not overlooked in the discussion.

[0055] "Personalization of knowledge" refers to the concentration of knowledge and information in the hands of specific individuals.

[0056] "Handover" refers to the transfer of work or knowledge to the next person in charge.

[0057] This invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0058] Hardware and Software Used

[0059] The server uses a high-performance server (e.g., a server equipped with an NVIDIA GPU). The software uses a generative AI model (e.g., GPT-4 (registered trademark)) and a discussion management system (e.g., a dedicated discussion framework).

[0060] Program Processing

[0061] The server receives a prompt from the user and selects an appropriate generative AI model. It then sends the prompt to the selected generative AI model, starting a discussion. The generative AI model starts a discussion with each AI, such as the maintenance AI, construction AI, and headquarters AI, and gives opinions from their own perspectives. The server collects and integrates the opinions of each AI. Finally, it returns the integrated discussion results to the user.

[0062] Examples

[0063] Using the terminal, the user inputs a prompt sentence related to the topic or agenda of the discussion. For example, the user inputs "Please start a discussion on improving network quality." The server receives this prompt sentence and analyzes it using a natural language processing algorithm. Based on the analysis results, the server selects a maintenance AI, a construction AI, or a headquarters AI. The server sends the prompt sentence to the selected generative AI model via API. The generative AI model receives the prompt sentence and begins preparing for the discussion.

[0064] The generative AI model initiates a discussion between the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to prioritize due to budget constraints." The server collects the opinions sent by each AI and integrates them using a natural language processing algorithm. This results in a comprehensive discussion result.

[0065] The server returns the integrated discussion results to the user, who can check the results on the screen of his / her device. For example, the conclusion "Strengthen network maintenance and inspection, and gradually introduce new infrastructure" may be displayed.

[0066] In this way, the server enables efficient and comprehensive discussions, reducing labor costs, accelerating discussions, and preventing items from being overlooked.

[0067] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0068] Step 1:

[0069] The user inputs a prompt text. The user inputs "Please start a discussion on improving the quality of the network" in the input field of the terminal and clicks the send button. The input prompt text is sent to the server.

[0070] Step 2:

[0071] The server receives the prompt sentence and selects an appropriate generative AI model. The server receives the prompt sentence and analyzes the prompt sentence using a natural language processing algorithm. Based on the analysis result, the maintenance AI, construction AI, or head office AI is selected. The input is the prompt sentence, and the output is the selected generative AI model.

[0072] Step 3:

[0073] The server sends a prompt to the selected generative AI model. The server sends the prompt to the selected generative AI model via API. The generative AI model receives the prompt and starts preparing for discussion. The input is the selected generative AI model and the prompt, and the output is a state ready for discussion.

[0074] Step 4:

[0075] The generative AI model starts a discussion, and each AI gives an opinion from its own perspective. The generative AI model starts a discussion with the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to set priorities due to budget constraints." The input is the prompt text and the role of each AI, and the output is the opinion of each AI.

[0076] Step 5:

[0077] The server collects and integrates the opinions of each AI. The server collects the opinions sent by each AI and integrates the opinions using a natural language processing algorithm. This results in a comprehensive discussion result. The input is the opinion of each AI, and the output is the integrated discussion result.

[0078] Step 6:

[0079] The server returns the integrated discussion results to the user. The server returns the integrated discussion results to the user. The user can check the results of the discussion on their terminal screen. For example, the conclusion "strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed. The input is the integrated discussion results, and the output is the discussion results that are displayed to the user.

[0080] (Application example 1)

[0081] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0082] Conventional discussion systems had problems in that it was difficult to gather the opinions of each department and proceed with discussions efficiently, which required a lot of man-hours and made it easy for considerations to be overlooked. Also, in discussions on how to improve production efficiency within factories, there was a lack of a way to generate AI with the stance of each department and have those AIs discuss with each other. Furthermore, the implementation as an application to be installed on smartphones was also insufficient.

[0083] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a generating AI means for discussion between roles that are given the stance of each department as a role, a means for identifying concerns and initiating discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This makes it possible to generate AI with the stance of each department in a factory, and for these AIs to discuss ways to improve production efficiency. In addition, by implementing it as an application to be installed on a smartphone, users can easily use it.

[0084] A "generative AI means" is an artificial intelligence means that is generated to hold discussions between roles that are assigned the stance of each department.

[0085] "Means for identifying concerns and initiating discussions without human intervention" refers to a means for automatically extracting concerns or issues that arise during the discussion process and starting a discussion.

[0086] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" are measures to improve the efficiency of discussions and cover all necessary items without missing anything.

[0087] "A means of generating AI that has the perspective of each department within a factory and having these AIs discuss with each other how to improve production efficiency" refers to a means of generating artificial intelligence that has the perspective of each department within a factory (production, quality control, maintenance, etc.) and having these AIs discuss with each other how to improve production efficiency.

[0088] "Means including an application installed on a smartphone" refers to means implemented as an application installed on a smartphone.

[0089] "Roles of maintenance AI, construction AI, headquarters AI, production AI, quality control AI, maintenance AI, etc." are the roles of artificial intelligence that have the responsibilities and perspectives of each department.

[0090] As an embodiment of the present invention, we will create an AI that has the stance of each department in a factory, and build a system in which these AIs discuss ways to improve production efficiency. The specific embodiment will be described below.

[0091] System Configuration

[0092] The system is composed of a server, a smartphone, a generative AI model, and a prompt. The server provides the generative AI means for discussions between roles that are assigned the stances of each department. The smartphone is a device on which users can install applications that allow them to easily use the system.

[0093] Hardware and Software Used

[0094] Hardware: Servers, smartphones

[0095] Software: Python, OpenAI® API

[0096] Data processing and calculation

[0097] The server processes and calculates data in the following procedure.

[0098] 1. Using generative AI models:

[0099] The server uses the OpenAI API to generate an AI with the stance of each department. Specifically, it inputs a prompt sentence and generates an AI with the stance of each department.

[0100] 2. Conducting the discussion:

[0101] The server holds discussions between the generated AIs of each department, creates discussion prompts, and uses the OpenAI API to obtain the results of the discussion.

[0102] 3. Display the results:

[0103] The results of the discussion are displayed to the user through an application installed on their smartphone.

[0104] Examples

[0105] Below are some examples of prompt sentences.

[0106] Generate an AI with a production sector stance.

[0107] Generate an AI with a quality control department stance.

[0108] Generate an AI with a maintenance department stance.

[0109] Please discuss with the AIs below how to improve production efficiency.

[0110] By using this prompt, it is possible to generate AI with the stance of each department and hold discussions. For example, the production department's AI can propose "what kind of data should be collected to identify bottlenecks in the production line," while the quality control department's AI can discuss "what kind of inspection method should be introduced as a measure to improve quality control."

[0111] In this way, a system is realized that enables efficient discussions on how to improve production efficiency within a factory.

[0112] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0113] Step 1:

[0114] The server creates a prompt sentence to generate an AI with the stance of each department. Specifically, it generates a prompt sentence like the following based on the department specified by the user (e.g. production, quality control, maintenance): "Please generate an AI with the stance of the production department." This prompt sentence is the input, and the generated AI is the output.

[0115] Step 2:

[0116] The server sends the generated prompt sentence to OpenAI's API to generate an AI with the stance of each department. Specifically, it calls OpenAI's API and provides the prompt sentence as input. The API uses the generative AI model to generate an AI with the stance of each department and returns the result to the server. This result becomes the output.

[0117] Step 3:

[0118] The server stores the generated AI for each department in a list. Specifically, the AI ​​returned from the API is saved in list format and used in the subsequent discussion process. This list is the input, and the stored AI list is the output.

[0119] Step 4:

[0120] The server creates a prompt for the AIs in each department to discuss with each other. Specifically, it generates the following prompt based on the AIs stored in the list: "Please hold a discussion between the AIs below to improve production efficiency." This prompt is the input, and the results of the discussion are the output.

[0121] Step 5:

[0122] The server sends the discussion prompt to the OpenAI API and obtains the discussion result. Specifically, it calls the API and provides the discussion prompt as input. The API uses a generative AI model to generate the discussion result and returns the result to the server. This result becomes the output.

[0123] Step 6:

[0124] The server transmits the acquired results of the discussion to an application installed on the smartphone. Specifically, it converts the results of the discussion into a format that the user can view and transmits them to the smartphone. The results of this discussion are the input, and data in a format that the user can view is the output.

[0125] Step 7:

[0126] The user checks the results of the discussion displayed on the smartphone and makes additional discussions or adjustments as necessary. Specifically, the user views the results of the discussion through the smartphone application and decides on the next action. The results of this discussion are the input, and the user's actions are the output.

[0127] Example 2

[0128] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0129] In conventional discussion systems, it was necessary for people with specialized knowledge from each department to directly participate in the discussions, which led to problems with personalizing knowledge and handover. In addition, it took a lot of time to progress the discussions and record the results, making it difficult to accelerate the discussions and prevent oversights. This made it difficult to hold efficient discussions and propose optimal solutions.

[0130] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0131] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for a user to input the topic of discussion, a means for the server to start a discussion and collect opinions from each AI model, and a means for automatically generating the results of the discussion as minutes and saving them in a database. This makes it possible to have efficient discussions and propose optimal solutions without the need for personalizing or handing over knowledge.

[0132] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0133] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically extracts discussion topics and concerns and starts discussions.

[0134] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[0135] "Means for users to input a topic for discussion" refers to an interface that allows users to input a topic for discussion into the system.

[0136] "Means for the server to initiate discussions and aggregate opinions from each AI model" refers to technology that enables the server to initiate discussions and collect and integrate opinions from each AI model.

[0137] "Means for automatically generating the results of discussions as minutes and storing them in a database" refers to technology for automatically recording the content and results of discussions in text format and storing them in a database.

[0138] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[0139] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[0140] "Headquarters AI" refers to artificial intelligence that understands the company's overall policies and strategies.

[0141] "Database" refers to an information management system for storing the results of discussions and minutes.

[0142] This invention is a system that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions. Specifically, three AI models, a maintenance AI, a construction AI, and a headquarters AI, hold discussions using their respective expertise, and the results are automatically saved as minutes.

[0143] Hardware and Software Used

[0144] Hardware: Servers, terminals

[0145] Software: Generative AI models (e.g. GPT-4), database management systems (e.g. MySQL®)

[0146] Generating a Program

[0147] The server generates programs including the maintenance AI, construction AI, and headquarters AI models. The programs are implemented using programming languages ​​such as Python and Java. Each AI model is based on a generative AI model (e.g., GPT-4) and is trained to have its own specialized knowledge.

[0148] Program Processing

[0149] The user inputs the topic of discussion

[0150] The user inputs the topic of discussion into the generative AI model using the device. For example, if the user wants to input the topic "optimal strategy for designing new network infrastructure," the user writes the topic in the input form on the device and clicks the send button.

[0151] The server starts the discussion.

[0152] The server receives the topic entered by the user and has the maintenance AI, construction AI, and headquarters AI start a discussion. Specifically, the server generates a prompt for each AI model and sends an API request to hold a discussion. For example, the server generates a prompt like the one below.

[0153] Example prompt:

[0154] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[0155] The server aggregates the results of the discussion.

[0156] The server aggregates the opinions returned by each AI model. Specifically, it analyzes the responses from each AI model and applies an algorithm to propose the optimal solution. For example, if the maintenance AI suggests that "a backup system should be installed to ensure network redundancy" and the construction AI suggests that "fiber optic cables should be used for cost-effectiveness," the server integrates these opinions to derive the optimal solution of "using fiber optic cables and installing a backup system."

[0157] The server generates the minutes.

[0158] The server automatically generates the results of the discussion as minutes. Specifically, it summarizes the contents of the discussion and the final solution in text format and saves them in a database management system (e.g., MySQL). Users can later refer to these minutes. For example, when a user clicks the "Show minutes" button on a terminal, the server retrieves the minutes from the database and displays them on the terminal.

[0159] Examples

[0160] For example, if a user inputs the theme "optimal strategy for designing new network infrastructure," the following prompt sentence is input to the generative AI model.

[0161] Example prompt:

[0162] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[0163] Based on this prompt, the server will have each AI start a discussion and propose the optimal solution. The results of the discussion are automatically saved as minutes, so the user can check the contents later.

[0164] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0165] Step 1:

[0166] Generating a Program

[0167] The server generates a program including each AI model: maintenance AI, construction AI, and headquarters AI. Specifically, using programming languages ​​such as Python and Java, each AI model is trained based on a generative AI model (e.g., GPT-4). Training data for each AI model is required as input, and each AI model is generated as output.

[0168] Step 2:

[0169] Enter the discussion topic

[0170] The user uses a terminal to input the topic of discussion into the generative AI model. Specifically, the user enters the topic in the input form on the terminal and clicks the send button. The input is the topic of discussion entered by the user, and the output is the topic sent to the server.

[0171] Step 3:

[0172] Starting the discussion

[0173] The server receives the topic input by the user and starts a discussion with the maintenance AI, construction AI, and headquarters AI. Specifically, the server generates a prompt for each AI model and sends an API request to hold the discussion. The input is the discussion topic received from the user, and the output is the generated prompt to be sent to each AI model.

[0174] Step 4:

[0175] Gathering opinions

[0176] The server collects the opinions returned from each AI model. Specifically, it analyzes the response from each AI model and obtains the opinions in text format. The input is the response from each AI model, and the output is the collected opinions of each AI model in text format.

[0177] Step 5:

[0178] Summarizing the results of discussions

[0179] The server proposes the optimal solution based on the collected opinions. Specifically, it integrates the opinions from each AI model and applies an algorithm to derive the optimal solution. The inputs are the opinions of each AI model, and the output is the proposed optimal solution.

[0180] Step 6:

[0181] Generate minutes

[0182] The server automatically generates the results of the discussion as minutes and saves them in a database. Specifically, the contents of the discussion and the final solution are summarized in text format and saved in a database management system (e.g., MySQL). The input is the results of the discussion, and the output is the minutes that are generated and saved in the database.

[0183] Step 7:

[0184] View minutes

[0185] A user refers to the minutes from a terminal. Specifically, when the user clicks the "Show minutes" button, the server retrieves the minutes from the database and displays them on the terminal. The input is a user request, and the output is the minutes displayed on the terminal.

[0186] (Application example 2)

[0187] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0188] In the conventional system, it took a lot of time and effort for specialists from each department to hold discussions and come up with optimal solutions. It was also time-consuming to accurately record the results of discussions and create meeting minutes for future reference. Furthermore, knowledge was easily tampered with and problems with handover occurred, making efficient operation difficult.

[0189] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0190] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for automatically saving the results of the discussion as minutes, and a means for AIs with multiple specialized knowledge to cooperate and propose optimal solutions. This makes it possible to efficiently and quickly derive optimal solutions, automatically create minutes, and prevent knowledge from being personalized.

[0191] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[0192] "Issues of concern" refer to issues, challenges, risks, etc. that are subject to discussion.

[0193] "Introducing the discussion" is the process of preparing and setting up the discussion.

[0194] "Reducing labor hours" refers to reducing the time and effort required for work.

[0195] "Accelerating discussions" refers to moving discussions forward more quickly.

[0196] "Preventing oversights" refers to making sure that important matters are not overlooked during the discussion or deliberation process.

[0197] A "minutes" is a document that records the content and results of a discussion.

[0198] "Specialist knowledge" refers to advanced knowledge or skills in a particular field.

[0199] "Personalization of knowledge" refers to a state in which knowledge or skills become dependent on specific individuals.

[0200] "Handover" is the process of transferring work and knowledge to the next person in charge.

[0201] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means for automatically saving the results of the discussion as minutes, and a means for multiple AIs with specialized knowledge to work together to propose optimal solutions.

[0202] Hardware and Software Used

[0203] Hardware: Smartphone

[0204] Software: Python, JSON

[0205] Data processing and calculation

[0206] The server uses a generation AI that is assigned a role based on the stance of each department to hold discussions. Specifically, AIs with specialized knowledge, such as maintenance AI, construction AI, and head office AI, work together to identify concerns and introduce discussions. The results of the discussion are automatically saved as minutes and can be viewed on smartphones.

[0207] Examples

[0208] For example, if a factory manager is considering installing a new production line, he or she could begin the discussion with the following prompt:

[0209] Examples of prompt statements

[0210] issue = "Introduction of new production line"

[0211] solutions = manager.conduct_meeting(issue)

[0212] print(solutions)

[0213] By executing this prompt, the maintenance AI will propose "solutions based on network maintenance," and the construction AI will propose "solutions based on network construction." This allows factory managers to check optimal production plans and maintenance schedules on their smartphones.

[0214] This system makes it possible to efficiently and quickly arrive at optimal solutions, automatically create meeting minutes, and prevent knowledge from becoming personalized.

[0215] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0216] Step 1:

[0217] The server initializes the generation AI, which is given the stance of each department as a role. Specifically, it creates instances of maintenance AI, construction AI, headquarters AI, etc., and sets their respective expertise. The name and expertise of each AI are required as input, and the initialized AI instance is obtained as output.

[0218] Step 2:

[0219] The user inputs the concern to be discussed, for example, setting the agenda "Installation of a new production line." As input, the concern to be discussed is required, and as output, the agenda is obtained.

[0220] Step 3:

[0221] The server provides each AI with an agenda and asks them to propose a solution. Each AI generates a solution based on its own expertise. The inputs are the agenda and the expertise of each AI, and the output is the solution proposed by each AI.

[0222] Step 4:

[0223] The server collects the solutions obtained from each AI and selects the optimal solution. As input, it needs the solution of each AI, and as output, it gets the optimal solution.

[0224] Step 5:

[0225] The server automatically saves the results of the discussion as minutes. Specifically, it saves the agenda, each AI's solution, and the optimal solution in JSON format. The inputs are the agenda, each AI's solution, and the optimal solution, and the output is the saved minutes.

[0226] Step 6:

[0227] The user checks the saved minutes using a smartphone. The saved minutes are required as input, and the contents of the minutes are displayed on the smartphone as output.

[0228] Furthermore, an emotion engine that estimates the emotion of the user may be combined. That is, the identification processing unit 290 may estimate the emotion of the user using the emotion identification model 59, and perform identification processing using the emotion of the user.

[0229] "Example 1"

[0230] In one embodiment of the present invention, the generating AI means includes an emotion engine that recognizes the user's emotions. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generating AI. The generating AI adjusts the progress of the discussion based on this emotion information. For example, if the user shows anger or dissatisfaction, the generating AI pauses the progress of the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generating AI smoothly progresses the discussion and actively incorporates the user's opinion.

[0231] "Example 2"

[0232] In another embodiment of the present invention, the emotion engine adjusts the results of the discussion according to the user's emotions. For example, if the user is dissatisfied, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user is dissatisfied with the quality of the network, the generation AI prioritizes discussions on improving the quality and reports the results to the user. This enables discussions that take the user's emotions into account.

[0233] The process flow of each embodiment will be described below.

[0234] "Example 1"

[0235] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[0236] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[0237] Step 3: The generation AI adjusts the progress of the discussion based on the emotion information. Specifically, if the user shows anger or dissatisfaction, the generation AI pauses the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generation AI smoothly advances the discussion and actively incorporates the user's opinion.

[0238] "Example 2"

[0239] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[0240] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[0241] Step 3: The generation AI adjusts the results of the discussion based on the emotion information. Specifically, if the user expresses dissatisfaction, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user expresses dissatisfaction with the quality of the network, the generation AI prioritizes discussions on improving quality and reports the results to the user.

[0242] Example 1

[0243] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal".

[0244] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and move the discussion forward. In addition, the user's feelings cannot be properly reflected during the discussion, which can lead to a decline in the quality of the discussion. Furthermore, it is difficult to quickly collect the results of the discussion and provide them as a report. In order to solve these problems, an efficient discussion system using generative AI that has the stance of each department is needed.

[0245] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0246] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for adjusting the progress of discussions using an emotion engine that recognizes the emotions of users, and a means for aggregating the results of discussions and generating reports. This makes it possible to efficiently aggregate the opinions of each department, progress discussions while appropriately reflecting the emotions of users, and quickly provide the results of discussions as reports.

[0247] "Generative AI means" refers to artificial intelligence generated to hold discussions between roles that are assigned the stance of each department.

[0248] "Means of identifying concerns and initiating discussions without human intervention" refers to a process whereby the system automatically identifies concerns and initiates discussions.

[0249] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to methods for increasing the efficiency of discussions and covering all necessary items without missing anything.

[0250] An "emotion engine" refers to technology that analyzes emotions from a user's voice, facial expressions, text input, etc., and provides the results to generative AI.

[0251] The "means for adjusting the progress of the discussion" refers to a method for appropriately adjusting the progress of the discussion based on the user's emotion information obtained from the emotion engine.

[0252] "Means of aggregating the results of discussions and generating a report" refers to the process for summarizing the results of each AI's discussions and creating a final report.

[0253] "Conservation AI" refers to artificial intelligence that takes the stance of the maintenance department and engages in discussions about maintenance.

[0254] "Construction AI" refers to artificial intelligence that takes a stance on the construction sector and engages in discussions about construction.

[0255] "Headquarters AI" refers to artificial intelligence that takes the stance of the headquarters department and engages in discussions regarding overall management and strategy.

[0256] "User" refers to a person who accesses the system to advance a discussion.

[0257] "Server" refers to a computer system for executing various means such as generative AI means and emotion engines.

[0258] This invention is a system that holds discussions using generation AI that is given a role based on the stance of each department. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs hold discussions with each other. The content of the discussions is related to improving network quality and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0259] Hardware and Software Used

[0260] The server is a computer system for executing the generative AI model. The generative AI model is built using machine learning libraries such as TENSORFLOW (registered trademark) and PyTorch. The server also includes an emotion engine for analyzing the user's emotions. The emotion engine uses voice recognition software and image analysis software to analyze the user's voice, facial expressions, text input, etc.

[0261] A terminal is a device through which a user accesses the system, and includes a PC, a smartphone, etc. A user accesses the system using a web browser and inputs a topic for discussion.

[0262] Program Processing

[0263] The server detects user access and launches a generative AI model. The generative AI model includes pre-trained AI algorithms. The server obtains the necessary information from each department's database to generate an AI with the stance of each department. For example, a maintenance AI is generated based on past maintenance records.

[0264] The server sets the topic of discussion provided by the user. For example, it sets the topic, "What measures are necessary to improve the quality of the network?" The server sends a prompt to each AI to start the discussion. Each AI expresses its opinion based on its own stance.

[0265] The server uses an emotion engine to analyze the user's emotions. Specifically, it analyzes the user's voice, facial expressions, text input, etc. to grasp the user's emotional state. The server adjusts the progress of the discussion based on the user's emotional information obtained from the emotion engine. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[0266] The server aggregates the results of each AI's discussion and generates a report. The report includes each AI's opinion and the conclusion of the discussion. The user can use the terminal to check the generated report and provide feedback if necessary.

[0267] Examples of concrete examples and prompts

[0268] As a concrete example, consider the case of discussing how to improve the quality of a network. The server provides the following topics to the maintenance AI, construction AI, and head office AI:

[0269] "What measures are necessary to improve the quality of the network?"

[0270] Each AI debates as follows:

[0271] Maintenance AI: "The aging of current network equipment is an issue, and regular maintenance is necessary."

[0272] Construction AI: "We need to introduce new network equipment, especially the latest communication technology."

[0273] Headquarters AI: "We need to consider the cost. We need to take the best measures within our budget."

[0274] The emotion engine analyzes the user's emotions, and if the user indicates dissatisfaction, the generative AI pauses the discussion and prioritizes the user's opinion.

[0275] Example prompt:

[0276] "What measures are necessary to improve the quality of the network? Please give us your opinions on maintenance AI, construction AI, and headquarters AI."

[0277] In this way, the server, terminals, and users work together to execute system processing.

[0278] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0279] Step 1:

[0280] A user accesses the system

[0281] A user accesses the system using a terminal (such as a PC or smartphone). Specifically, the user opens a web browser, enters the system's URL, and logs in. The user's authentication information (user name, password) is required as input, and the system's dashboard is displayed as output.

[0282] Step 2:

[0283] The server runs the generative AI model

[0284] The server detects user access and launches the generative AI model. The user's login information is required as input, and the generative AI model is launched as output. The generative AI model is built using machine learning libraries such as TensorFlow and PyTorch.

[0285] Step 3:

[0286] The server generates AI with a stance for each division.

[0287] The server generates AI with the stance of the maintenance department, construction department, head office department, etc. As input, information obtained from the database of each department is required, and as output, a maintenance AI, a construction AI, a head office AI, etc. For example, a maintenance AI is generated based on past maintenance records.

[0288] Step 4:

[0289] The server sets the topic of discussion

[0290] The server sets the topic of discussion provided by the user. As input, it requires the topic entered by the user (e.g., "What measures are necessary to improve the quality of the network?"), and as output, a prompt sentence is generated for each AI.

[0291] Step 5:

[0292] The server starts the discussion for each AI.

[0293] The server has each AI start a discussion based on the set theme. A prompt is required as input, and the opinion of each AI is generated as output. For example, the maintenance AI may say, "The aging of the current network equipment is a problem, and regular maintenance is necessary."

[0294] Step 6:

[0295] The server uses an emotion engine to analyze the user's emotions.

[0296] The server uses an emotion engine to analyze the user's emotions. The inputs required are the user's voice, facial expressions, text input, etc., and the output is the analyzed user's emotional state. Specifically, the analysis is performed using voice recognition software and image analysis software.

[0297] Step 7:

[0298] The server coordinates the progress of the discussion.

[0299] The server adjusts the progress of the discussion based on the user's emotion information obtained from the emotion engine. The user's emotion information is required as input, and the progress of the discussion is adjusted as output. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[0300] Step 8:

[0301] The server aggregates the results of the discussions and generates a report.

[0302] The server aggregates the results of each AI's discussion and generates a report. The input is the results of each AI's discussion, and the output is a PDF report. The report contains detailed descriptions of each AI's opinions and the conclusions of the discussion.

[0303] Step 9:

[0304] User checks the report

[0305] The user uses a terminal to review the generated report. As input, the report downloaded through the system interface is required, and as output, the report content is displayed. The user can also provide feedback on the report content.

[0306] (Application example 1)

[0307] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0308] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and advance the discussion. In addition, because the discussion proceeds without considering the user's feelings, the user's dissatisfaction or anger may hinder the progress of the discussion. Furthermore, there is a problem that the efficiency of the discussion decreases due to a lack of means to check the progress of the discussion in real time and intervene if necessary. In order to solve these issues, a system is needed that uses a generative AI that has the stance of each department to hold the discussion, recognizes the user's feelings, and adjusts the progress of the discussion.

[0309] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0310] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion through a smartphone or a head-mounted display and intervening as necessary. This makes it possible to efficiently collect the opinions of each department and progress the discussion while taking into account the user's emotions, thereby improving the efficiency and effectiveness of the discussion.

[0311] "Generative AI means" is an artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0312] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically analyzes the content of discussions, extracts important concerns, and starts discussions.

[0313] "Means of reducing labor costs, accelerating discussions, and preventing items from being overlooked" are techniques for increasing the efficiency of discussions and ensuring that all necessary items are not overlooked.

[0314] An "emotion engine" is a technology that analyzes emotions from a user's voice, facial expressions, text input, etc.

[0315] The "means for adjusting the progress of the discussion based on emotional information" is a technique for appropriately controlling the progress of the discussion depending on the emotional state of the users.

[0316] "Means of checking the progress of the discussion through a smartphone or head-mounted display and intervening if necessary" refers to technology that uses mobile devices or wearable devices to monitor the progress of the discussion in real time and allows users to intervene if necessary.

[0317] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion via a smartphone or head-mounted display and intervening if necessary.

[0318] System Programming

[0319] The server implements a generative AI model using the OpenAI API to generate an AI with the stance of each department. This allows discussions to be held from the perspectives of the maintenance department, manufacturing department, quality control department, etc. To recognize the user's emotions, the EmotionRecognizer library is used to capture the user's facial expressions through a webcam. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generative AI. The generative AI adjusts the progress of the discussion based on this emotional information.

[0320] Hardware and Software

[0321] Hardware: Smartphone, head-mounted display, webcam

[0322] Software: OpenAI API, OpenCV, EmotionRecognizer library

[0323] Data processing and calculation

[0324] The server uses a generative AI model to generate discussions from each department's perspective. To recognize the user's emotions, a webcam is used to capture facial expressions in real time and the emotions are analyzed with the EmotionRecognizer library. The analyzed emotion information is provided to the generative AI and used to adjust the progress of the discussion. Through a smartphone or head-mounted display, the user can check the progress of the discussion in real time and intervene if necessary.

[0325] Examples

[0326] For example, if you are discussing the introduction of a new manufacturing line, you might use the following prompt:

[0327] "Discuss from the maintenance department's perspective: Discuss concerns regarding the implementation of a new production line."

[0328] "Discuss from the manufacturing perspective: Discuss the concerns regarding the implementation of a new manufacturing line."

[0329] "Discuss from the quality control department's perspective: Discuss concerns regarding the implementation of a new production line."

[0330] This makes it possible to efficiently collect opinions from each department and conduct discussions that take into account the feelings of users, thereby improving the efficiency and effectiveness of discussions.

[0331] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0332] Step 1:

[0333] The server initializes the generative AI model using the OpenAI API. As input, it receives a prompt sentence with the stance of each department and passes it to the generative AI model. The generative AI model generates a discussion from the perspective of each department and returns the discussion content in text format as output.

[0334] Step 2:

[0335] The server captures the user's facial expressions in real time using a webcam. It passes the captured image data as input to the EmotionRecognizer library, which analyzes the image data and outputs the user's emotional information.

[0336] Step 3:

[0337] The server provides the generative AI model with emotion information obtained from the EmotionRecognizer library. As input, the generative AI model passes the emotion information and the content of the discussion to adjust the progress of the discussion. The generative AI model appropriately controls the progress of the discussion based on the emotion information and returns the adjusted content of the discussion as output.

[0338] Step 4:

[0339] The terminal (smartphone or head-mounted display) receives the coordinated discussion content from the server. As input, it receives the discussion content from the server and displays it to the user. The user can check the progress of the discussion in real time and intervene if necessary.

[0340] Step 5:

[0341] Users intervene in the discussion through a smartphone or a head-mounted display. As input, users' opinions and feedback are entered into the device and sent to the server. The server receives the users' opinions and passes them back to the generative AI model to update the discussion content. As output, the updated discussion content is displayed again on the device.

[0342] Step 6:

[0343] The server saves the progress of the discussion as a log. As input, it receives the discussion content and emotion information generated at each step and saves them in a database. This allows the discussion history to be checked later.

[0344] Example 2

[0345] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0346] In conventional discussion systems, human intervention was required to hold discussions that reflected the stance of each department, which not only took time but also often slowed down the progress of discussions. In addition, manual work was required to save the results of discussions, which led to problems with personalizing knowledge and handover. Furthermore, since discussions did not take into account the feelings of users, user satisfaction could decline.

[0347] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0348] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing man-hours, accelerating discussions, and preventing oversights, an emotion engine means for adjusting the results of discussions according to the emotions of users, and a means for saving the results of discussions as meeting minutes. This makes it possible to streamline discussions, prevent knowledge from becoming personal, and hold discussions that take into account the emotions of users.

[0349] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0350] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically initiates discussions and identifies key issues.

[0351] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[0352] "Emotion engine means" refers to technology for adjusting the outcome of a discussion according to the user's emotions.

[0353] "Means of saving the results of the discussion as meeting minutes" refers to technology for automatically recording and saving the content and results of the discussion.

[0354] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[0355] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[0356] "Headquarters AI" refers to artificial intelligence that understands the overall policies and strategies of the company and holds discussions based on them.

[0357] This invention is a system that includes a generation AI means for holding discussions between roles that are assigned roles that represent the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, an emotion engine means for adjusting the results of the discussion according to the user's emotions, and a means for saving the results of the discussion as meeting minutes.

[0358] Generating a Program

[0359] The server generates the system's programs, which include the maintenance AI, construction AI, headquarters AI, and emotion engine. These programs allow each AI to hold discussions based on their respective expertise and propose optimal solutions.

[0360] Program processing explanation

[0361] The server processes the data in the following steps:

[0362] 1. Data collection:

[0363] The server collects data on network maintenance status, construction plans, and company policies and strategies. This data serves as the basis for each AI's discussion. Specifically, the server obtains data from network monitoring tools (e.g., Nagios) and project management software (e.g., JIRA).

[0364] 2. Start the discussion:

[0365] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson (registered trademark)).

[0366] 3. Discussion process:

[0367] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on maintaining the network, while the construction AI proposes a plan to build a new network. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center."

[0368] 4. Save the results of the discussion:

[0369] The server saves the results of the discussion as minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL).

[0370] 5. Emotion engine intervention:

[0371] When a user expresses dissatisfaction, the server activates an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, if a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality.

[0372] 6. Reporting results:

[0373] The server reports the final discussion result to the user. Specifically, the result is displayed on the user's dashboard (e.g., a web application). For example, the report may be in the form of "We will implement the following measures to improve the quality of the network."

[0374] Examples of concrete examples and prompts

[0375] Examples:

[0376] If a user is dissatisfied with the quality of the network, the following prompt sentence is fed into the generative AI model:

[0377] Example prompt:

[0378] "Users are complaining about the quality of their network. Please prioritize discussions on how to improve the quality and report the results."

[0379] Based on this prompt, the generation AI holds a discussion about improving quality, with the maintenance AI proposing, "We should introduce a new monitoring tool to strengthen network monitoring," and the construction AI proposing, "We should improve network throughput by introducing a new router." The server displays the results of the discussion on the user's dashboard, reporting, "We will introduce a new monitoring tool and router to improve network quality."

[0380] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0381] Step 1:

[0382] The server collects data on network maintenance status, construction plans, and company policies and strategies. Specifically, it obtains data from network monitoring tools (e.g. Nagios) and project management software (e.g. JIRA). As input, it receives data from these tools, and as output, it generates the basic information for each AI to discuss.

[0383] Step 2:

[0384] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson). As input, it provides each AI with the collected data, and as output, it performs initial settings for each AI to start a discussion.

[0385] Step 3:

[0386] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on network maintenance, while the construction AI proposes a new network construction plan. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center." As input, it receives suggestions from each AI, and as output, it generates an optimal solution.

[0387] Step 4:

[0388] The server saves the results of the discussion as the minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL). It receives the results of the discussion as input, and generates minutes as output and saves them in the database.

[0389] Step 5:

[0390] When a user expresses dissatisfaction, the server launches an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, when a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality. It receives the user's emotion data as input and generates instructions to adjust the outcome of the discussion as output.

[0391] Step 6:

[0392] The server reports the final discussion result to the user. Specifically, it displays the result on the user's dashboard (e.g., a web application). For example, it reports in the form of "We will implement the following measures to improve the quality of the network." It receives the final discussion result as input and generates information to report to the user as output.

[0393] (Application example 2)

[0394] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0395] In conventional network security systems, it is difficult to integrate the expertise of each department to derive optimal solutions, and there is a lack of consideration for user feelings, making it difficult to improve user satisfaction. Furthermore, efficient discussion and coordination is required in detecting security risks, proposing countermeasures, and designing and implementing new security infrastructure.

[0396] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0397] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, an emotion engine means for analyzing user emotions and adjusting the results of discussions, a security preservation AI means for detecting security risks and proposing measures, a security construction AI means for proposing the design and implementation of new security infrastructure, and a headquarters AI means for understanding the security policy and strategy of the entire company and adjusting the overall security strategy. This makes it possible to integrate the expertise of each department and propose optimal security measures that take into account the emotions of users.

[0398] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[0399] An "emotion engine" is an engine that has the function of analyzing the user's emotions and adjusting the results of the discussion according to those emotions.

[0400] "Security preservation AI" is an artificial intelligence that has knowledge and experience in network preservation, detects security risks, and proposes countermeasures.

[0401] "Security Construction AI" is an artificial intelligence that has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods.

[0402] "Headquarters AI" is an artificial intelligence that understands the security policies and strategies of the entire company and adjusts the overall security strategy.

[0403] A "role" is a virtual role with expertise and experience that is assigned to each department's stance.

[0404] "Introducing a discussion" is the process of identifying concerns and starting a discussion without human intervention.

[0405] "Reducing labor hours" refers to measures aimed at reducing the time and effort required for work.

[0406] "Accelerating discussions" is a measure aimed at speeding up the progress of discussions.

[0407] "Preventing omissions in consideration" is a measure to ensure that important matters are not overlooked during discussion or consideration.

[0408] A system for implementing this invention is configured as follows: The server includes a generation AI that holds discussions between roles that are assigned the stance of each department, an emotion engine that analyzes user emotions and adjusts the results of the discussion, a security preservation AI that detects security risks and proposes countermeasures, a security construction AI that proposes the design and implementation of new security infrastructure, and a headquarters AI that understands the security policy and strategy of the entire company and adjusts the overall security strategy.

[0409] Hardware and Software Configuration

[0410] The server uses the following hardware and software:

[0411] Hardware: High-performance server (e.g. Dell PowerEdge series)

[0412] Software: Python, sentiment analysis libraries (e.g. TextBlob), generative AI models (e.g. GPT-3 (registered trademark))

[0413] Data processing and calculation

[0414] The server performs the following data processing and calculations.

[0415] 1. Generative AI: Each department is assigned a role based on their stance, and the roles are discussed to propose optimal solutions. A generative AI model (e.g. GPT-3) is used to generate the content of the discussion.

[0416] 2. Sentiment Engine: Analyze user sentiment and adjust the discussion outcome depending on the sentiment. Use a sentiment analysis library (e.g. TextBlob) to parse user feedback.

[0417] 3. Security Preservation AI: With knowledge and experience in network preservation, it detects security risks and proposes countermeasures. It uses generative AI models to detect risks and propose countermeasures.

[0418] 4. Security Construction AI: Has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods. Uses generative AI models to make design and implementation suggestions.

[0419] 5. HQ AI: Understand the security policies and strategies of the entire enterprise and adjust the overall security strategy. Use generative AI models to make adjustments to the strategy.

[0420] Examples

[0421] For example, if a user provides feedback like "I'm dissatisfied with the current security measures," the emotion engine will analyze this feedback and recognize that the user is expressing dissatisfaction. Based on this, the security preservation AI will suggest updating firewall rules.

[0422] Examples of prompt statements

[0423] User feedback: "I'm not satisfied with the current security measures."

[0424] AI discussion result: "Security Preservation AI proposes updating firewall rules."

[0425] In this way, it becomes possible to integrate the expertise of each department and propose optimal security measures that take into account the user's feelings.

[0426] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0427] Step 1:

[0428] The server receives feedback from the user. The input is the user's feedback text, and the output is the storage of the feedback text. For example, if the user enters "I am not satisfied with the current security measures," the server stores this text in a database.

[0429] Step 2:

[0430] The server parses the user feedback using an emotion engine. The input is the stored feedback text and the output is the user's emotional state. Specifically, it uses a sentiment analysis library (e.g., TextBlob) to extract the emotion "unhappy" from the feedback text.

[0431] Step 3:

[0432] The server sends a prompt to the generation AI based on the analysis results of the emotion engine. The input is the user's emotional state and feedback text, and the output is a prompt sentence for the generation AI. Specifically, it generates the prompt sentence "User feedback: 'I am dissatisfied with the current security measures.'"

[0433] Step 4:

[0434] The server uses a generative AI to hold a discussion and generate an optimal solution. The input is the prompt sentence, and the output is the result of the discussion. Specifically, a generative AI model (e.g., GPT-3) is used to generate the discussion result "Security preservation AI proposes updating firewall rules."

[0435] Step 5:

[0436] The server reports the generated discussion results to the user. The input is the discussion result, and the output is a report message to the user. Specifically, the server sends the message "Security preservation AI proposes updating firewall rules" to the user.

[0437] Step 6:

[0438] The server stores the results of the discussion in a database. The input is the discussion result and the output is the stored discussion result. Specifically, the generated discussion result is stored in the database for future reference and analysis.

[0439] In this way, the server performs a series of processes: receiving user feedback, analyzing emotions, using generative AI to propose optimal solutions, and reporting the results to the user.

[0440] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the voice data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0441] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0442] Another example of generative AI is Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0443] In the above embodiment, an example was given in which the specific process was performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0444] [Second embodiment]

[0445] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0446] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0447] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0448] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0449] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[0450] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[0451] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[0452] Fig. 4 shows an example of main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0453] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0454] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0455] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0456] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0457] "Example 1"

[0458] The present invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0459] "Example 2"

[0460] As a concrete example, the maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies. These AIs hold discussions based on their respective stances and propose optimal solutions. This discussion is conducted without human intervention, and the results of the discussion are saved as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0461] The process flow of each embodiment will be described below.

[0462] "Example 1"

[0463] Step 1: The system generates AIs with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.).

[0464] Step 2: The generated AIs begin to discuss with each other. The content of the discussions will be related to improving network quality and achieving the Ministry of Internal Affairs and Communications' objectives.

[0465] Step 3: The results of the discussion are saved as meeting minutes and can be referenced later.

[0466] Step 1: The maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies.

[0467] Step 2: These AIs debate based on their respective stances and propose the optimal solution.

[0468] Step 3: The discussion is conducted without human intervention, and the results are stored as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0469] Example 1

[0470] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0471] In the conventional discussion system, it was difficult to hold discussions that reflected the stance of each department, and there was a problem that it took a lot of time to identify concerns and start discussions. In addition, the human intervention hindered the acceleration of discussions and increased the risk of overlooking something. Furthermore, the necessity of personalizing knowledge and handover made it difficult to hold efficient discussions.

[0472] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0473] In this invention, the server includes a generation AI means for discussion between roles that are given the stance of each department, a means for a user to input a prompt text, a means for the server to receive the prompt text and select an appropriate generation AI model, a means for the server to send the prompt text to the generation AI model selected by the server, a means for the generation AI model to start a discussion and for each AI to give an opinion from its own perspective, a means for the server to collect and integrate the opinions of each AI, a means for the server to return the integrated discussion result to the user, a means for identifying concerns and introducing discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This enables efficient and comprehensive discussions.

[0474] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0475] A "prompt sentence" refers to an instruction sentence regarding the topic or subject of discussion that is entered by the user.

[0476] "Server" refers to a computer system for receiving prompts, selecting an appropriate generative AI model, and managing the discussion.

[0477] A "generative AI model" refers to an artificial intelligence model with a specific role or perspective, such as maintenance AI, construction AI, or headquarters AI.

[0478] "Conservation AI" refers to a generative AI model that has a maintenance department stance.

[0479] "Construction AI" refers to a generative AI model that has a construction sector stance.

[0480] "Headquarters AI" refers to a generative AI model that has the stance of the headquarters department.

[0481] "Discussion results" refers to the conclusions and proposals reached through discussions between generative AI models.

[0482] "Reducing labor hours" refers to reducing the time and effort required for work.

[0483] "Accelerating the discussion" refers to speeding up the discussion.

[0484] "Preventing gaps in consideration" refers to making sure that important matters are not overlooked in the discussion.

[0485] "Personalization of knowledge" refers to the concentration of knowledge and information in the hands of specific individuals.

[0486] "Handover" refers to the transfer of work or knowledge to the next person in charge.

[0487] This invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0488] Hardware and Software Used

[0489] The server uses a high-performance server (e.g., a server equipped with an NVIDIA GPU). The software uses a generative AI model (e.g., GPT-4) and a discussion management system (e.g., a dedicated discussion framework).

[0490] Program Processing

[0491] The server receives a prompt from the user and selects an appropriate generative AI model. It then sends the prompt to the selected generative AI model, starting a discussion. The generative AI model starts a discussion with each AI, such as the maintenance AI, construction AI, and headquarters AI, and gives opinions from their own perspectives. The server collects and integrates the opinions of each AI. Finally, it returns the integrated discussion results to the user.

[0492] Examples

[0493] Using the terminal, the user inputs a prompt sentence related to the topic or agenda of the discussion. For example, the user inputs "Please start a discussion on improving network quality." The server receives this prompt sentence and analyzes it using a natural language processing algorithm. Based on the analysis results, the server selects a maintenance AI, a construction AI, or a headquarters AI. The server sends the prompt sentence to the selected generative AI model via API. The generative AI model receives the prompt sentence and begins preparing for the discussion.

[0494] The generative AI model initiates a discussion between the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to prioritize due to budget constraints." The server collects the opinions sent by each AI and integrates them using a natural language processing algorithm. This results in a comprehensive discussion result.

[0495] The server returns the integrated discussion results to the user, who can check the results on the screen of his / her device. For example, the conclusion "Strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed.

[0496] In this way, the server enables efficient and comprehensive discussions, reducing labor costs, accelerating discussions, and preventing items from being overlooked.

[0497] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0498] Step 1:

[0499] The user inputs a prompt text. The user inputs "Please start a discussion on improving the quality of the network" in the input field of the terminal and clicks the send button. The input prompt text is sent to the server.

[0500] Step 2:

[0501] The server receives the prompt sentence and selects an appropriate generative AI model. The server receives the prompt sentence and analyzes the prompt sentence using a natural language processing algorithm. Based on the analysis result, the maintenance AI, construction AI, or head office AI is selected. The input is the prompt sentence, and the output is the selected generative AI model.

[0502] Step 3:

[0503] The server sends a prompt to the selected generative AI model. The server sends the prompt to the selected generative AI model via API. The generative AI model receives the prompt and starts preparing for discussion. The input is the selected generative AI model and the prompt, and the output is a state ready for discussion.

[0504] Step 4:

[0505] The generative AI model starts a discussion, and each AI gives an opinion from its own perspective. The generative AI model starts a discussion with the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to set priorities due to budget constraints." The input is the prompt text and the role of each AI, and the output is the opinion of each AI.

[0506] Step 5:

[0507] The server collects and integrates the opinions of each AI. The server collects the opinions sent by each AI and integrates the opinions using a natural language processing algorithm. This results in a comprehensive discussion result. The input is the opinion of each AI, and the output is the integrated discussion result.

[0508] Step 6:

[0509] The server returns the integrated discussion results to the user. The server returns the integrated discussion results to the user. The user can check the results of the discussion on their terminal screen. For example, the conclusion "strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed. The input is the integrated discussion results, and the output is the discussion results that are displayed to the user.

[0510] (Application example 1)

[0511] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0512] Conventional discussion systems had problems in that it was difficult to gather the opinions of each department and proceed with discussions efficiently, which required a lot of man-hours and made it easy for considerations to be overlooked. Also, in discussions on how to improve production efficiency within factories, there was a lack of a way to generate AI with the stance of each department and have those AIs discuss with each other. Furthermore, the implementation as an application to be installed on smartphones was also insufficient.

[0513] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a generating AI means for discussion between roles that are given the stance of each department as a role, a means for identifying concerns and initiating discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This makes it possible to generate AI with the stance of each department in a factory, and for these AIs to discuss ways to improve production efficiency. In addition, by implementing it as an application to be installed on a smartphone, users can easily use it.

[0514] A "generative AI means" is an artificial intelligence means that is generated to hold discussions between roles that are assigned the stance of each department.

[0515] "Means for identifying concerns and initiating discussions without human intervention" refers to a means for automatically extracting concerns or issues that arise during the discussion process and starting a discussion.

[0516] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" are measures to improve the efficiency of discussions and cover all necessary items without missing anything.

[0517] "A means of generating AI that has the perspective of each department within a factory and having these AIs discuss with each other how to improve production efficiency" refers to a means of generating artificial intelligence that has the perspective of each department within a factory (production, quality control, maintenance, etc.) and having these AIs discuss with each other how to improve production efficiency.

[0518] "Means including an application installed on a smartphone" refers to means implemented as an application installed on a smartphone.

[0519] "Roles of maintenance AI, construction AI, headquarters AI, production AI, quality control AI, maintenance AI, etc." are the roles of artificial intelligence that have the responsibilities and perspectives of each department.

[0520] As an embodiment of the present invention, we will create an AI that has the stance of each department in a factory, and build a system in which these AIs discuss ways to improve production efficiency. The specific embodiment will be described below.

[0521] System Configuration

[0522] The system is composed of a server, a smartphone, a generative AI model, and a prompt. The server provides the generative AI means for discussions between roles that are assigned the stances of each department. The smartphone is a device on which users can install applications that allow them to easily use the system.

[0523] Hardware and Software Used

[0524] Hardware: Servers, smartphones

[0525] Software: Python, OpenAI API

[0526] Data processing and calculation

[0527] The server processes and calculates data in the following procedure.

[0528] 1. Using generative AI models:

[0529] The server uses the OpenAI API to generate an AI with the stance of each department. Specifically, it inputs a prompt sentence and generates an AI with the stance of each department.

[0530] 2. Conducting the discussion:

[0531] The server holds discussions between the generated AIs of each department, creates discussion prompts, and uses the OpenAI API to obtain the results of the discussion.

[0532] 3. Display the results:

[0533] The results of the discussion are displayed to the user through an application installed on their smartphone.

[0534] Examples

[0535] Below are some examples of prompt sentences.

[0536] Generate an AI with a production sector stance.

[0537] Generate an AI with a quality control department stance.

[0538] Generate an AI with a maintenance department stance.

[0539] Please discuss with the AIs below how to improve production efficiency.

[0540] By using this prompt, it is possible to generate AI with the stance of each department and hold discussions. For example, the production department's AI can propose "what kind of data should be collected to identify bottlenecks in the production line," while the quality control department's AI can discuss "what kind of inspection method should be introduced as a measure to improve quality control."

[0541] In this way, a system is realized that enables efficient discussions on how to improve production efficiency within a factory.

[0542] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0543] Step 1:

[0544] The server creates a prompt sentence to generate an AI with the stance of each department. Specifically, it generates a prompt sentence like the following based on the department specified by the user (e.g. production, quality control, maintenance): "Please generate an AI with the stance of the production department." This prompt sentence is the input, and the generated AI is the output.

[0545] Step 2:

[0546] The server sends the generated prompt sentence to OpenAI's API to generate an AI with the stance of each department. Specifically, it calls OpenAI's API and provides the prompt sentence as input. The API uses the generative AI model to generate an AI with the stance of each department and returns the result to the server. This result becomes the output.

[0547] Step 3:

[0548] The server stores the generated AI for each department in a list. Specifically, the AI ​​returned from the API is saved in list format and used in the subsequent discussion process. This list is the input, and the stored AI list is the output.

[0549] Step 4:

[0550] The server creates a prompt for the AIs in each department to discuss with each other. Specifically, it generates the following prompt based on the AIs stored in the list: "Please hold a discussion between the AIs below to improve production efficiency." This prompt is the input, and the results of the discussion are the output.

[0551] Step 5:

[0552] The server sends the discussion prompt to the OpenAI API and obtains the discussion result. Specifically, it calls the API and provides the discussion prompt as input. The API uses a generative AI model to generate the discussion result and returns the result to the server. This result becomes the output.

[0553] Step 6:

[0554] The server transmits the acquired results of the discussion to an application installed on the smartphone. Specifically, it converts the results of the discussion into a format that the user can view and transmits them to the smartphone. The results of this discussion are the input, and data in a format that the user can view is the output.

[0555] Step 7:

[0556] The user checks the results of the discussion displayed on the smartphone and makes additional discussions or adjustments as necessary. Specifically, the user views the results of the discussion through the smartphone application and decides on the next action. The results of this discussion are the input, and the user's actions are the output.

[0557] Example 2

[0558] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0559] In conventional discussion systems, it was necessary for people with specialized knowledge from each department to directly participate in the discussions, which led to problems with personalizing knowledge and handover. In addition, it took a lot of time to progress the discussions and record the results, making it difficult to accelerate the discussions and prevent oversights. This made it difficult to hold efficient discussions and propose optimal solutions.

[0560] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0561] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for a user to input the topic of discussion, a means for the server to start a discussion and collect opinions from each AI model, and a means for automatically generating the results of the discussion as minutes and saving them in a database. This makes it possible to have efficient discussions and propose optimal solutions without the need for personalizing or handing over knowledge.

[0562] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0563] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically extracts discussion topics and concerns and starts discussions.

[0564] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[0565] "Means for users to input a topic for discussion" refers to an interface that allows users to input a topic for discussion into the system.

[0566] "Means for the server to initiate discussions and aggregate opinions from each AI model" refers to technology that enables the server to initiate discussions and collect and integrate opinions from each AI model.

[0567] "Means for automatically generating the results of discussions as minutes and storing them in a database" refers to technology for automatically recording the content and results of discussions in text format and storing them in a database.

[0568] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[0569] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[0570] "Headquarters AI" refers to artificial intelligence that understands the company's overall policies and strategies.

[0571] "Database" refers to an information management system for storing the results of discussions and minutes.

[0572] This invention is a system that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions. Specifically, three AI models, a maintenance AI, a construction AI, and a headquarters AI, hold discussions using their respective expertise, and the results are automatically saved as minutes.

[0573] Hardware and Software Used

[0574] Hardware: Servers, terminals

[0575] Software: Generative AI models (e.g. GPT-4), database management systems (e.g. MySQL)

[0576] Generating a Program

[0577] The server generates programs including the maintenance AI, construction AI, and headquarters AI models. The programs are implemented using programming languages ​​such as Python and Java. Each AI model is based on a generative AI model (e.g., GPT-4) and is trained to have its own specialized knowledge.

[0578] Program Processing

[0579] The user inputs the topic of discussion

[0580] The user inputs the topic of discussion into the generative AI model using the device. For example, if the user wants to input the topic "optimal strategy for designing new network infrastructure," the user writes the topic in the input form on the device and clicks the send button.

[0581] The server starts the discussion.

[0582] The server receives the topic entered by the user and has the maintenance AI, construction AI, and headquarters AI start a discussion. Specifically, the server generates a prompt for each AI model and sends an API request to hold a discussion. For example, the server generates a prompt like the one below.

[0583] Example prompt:

[0584] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[0585] The server aggregates the results of the discussion.

[0586] The server aggregates the opinions returned by each AI model. Specifically, it analyzes the responses from each AI model and applies an algorithm to propose the optimal solution. For example, if the maintenance AI suggests that "a backup system should be installed to ensure network redundancy" and the construction AI suggests that "fiber optic cables should be used for cost-effectiveness," the server integrates these opinions to derive the optimal solution of "using fiber optic cables and installing a backup system."

[0587] The server generates the minutes.

[0588] The server automatically generates the results of the discussion as minutes. Specifically, it summarizes the contents of the discussion and the final solution in text format and saves them in a database management system (e.g., MySQL). Users can later refer to these minutes. For example, when a user clicks the "Show minutes" button on a terminal, the server retrieves the minutes from the database and displays them on the terminal.

[0589] Examples

[0590] For example, if a user inputs the theme "optimal strategy for designing new network infrastructure," the following prompt sentence is input to the generative AI model.

[0591] Example prompt:

[0592] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[0593] Based on this prompt, the server will have each AI start a discussion and propose the optimal solution. The results of the discussion are automatically saved as minutes, so the user can check the contents later.

[0594] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0595] Step 1:

[0596] Generating a Program

[0597] The server generates a program including each AI model: maintenance AI, construction AI, and headquarters AI. Specifically, using programming languages ​​such as Python and Java, each AI model is trained based on a generative AI model (e.g., GPT-4). Training data for each AI model is required as input, and each AI model is generated as output.

[0598] Step 2:

[0599] Enter the discussion topic

[0600] The user uses a terminal to input the topic of discussion into the generative AI model. Specifically, the user enters the topic in the input form on the terminal and clicks the send button. The input is the topic of discussion entered by the user, and the output is the topic sent to the server.

[0601] Step 3:

[0602] Starting the discussion

[0603] The server receives the topic input by the user and starts a discussion with the maintenance AI, construction AI, and headquarters AI. Specifically, the server generates a prompt for each AI model and sends an API request to hold the discussion. The input is the discussion topic received from the user, and the output is the generated prompt to be sent to each AI model.

[0604] Step 4:

[0605] Gathering opinions

[0606] The server collects the opinions returned from each AI model. Specifically, it analyzes the response from each AI model and obtains the opinions in text format. The input is the response from each AI model, and the output is the collected opinions of each AI model in text format.

[0607] Step 5:

[0608] Summarizing the results of discussions

[0609] The server proposes the optimal solution based on the collected opinions. Specifically, it integrates the opinions from each AI model and applies an algorithm to derive the optimal solution. The inputs are the opinions of each AI model, and the output is the proposed optimal solution.

[0610] Step 6:

[0611] Generate minutes

[0612] The server automatically generates the results of the discussion as minutes and saves them in a database. Specifically, the contents of the discussion and the final solution are summarized in text format and saved in a database management system (e.g., MySQL). The input is the results of the discussion, and the output is the minutes that are generated and saved in the database.

[0613] Step 7:

[0614] View minutes

[0615] A user refers to the minutes from a terminal. Specifically, when the user clicks the "Show minutes" button, the server retrieves the minutes from the database and displays them on the terminal. The input is a user request, and the output is the minutes displayed on the terminal.

[0616] (Application example 2)

[0617] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0618] In the conventional system, it took a lot of time and effort for specialists from each department to hold discussions and come up with optimal solutions. It was also time-consuming to accurately record the results of discussions and create meeting minutes for future reference. Furthermore, knowledge was easily tampered with and problems with handover occurred, making efficient operation difficult.

[0619] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0620] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for automatically saving the results of the discussion as minutes, and a means for AIs with multiple specialized knowledge to cooperate and propose optimal solutions. This makes it possible to efficiently and quickly derive optimal solutions, automatically create minutes, and prevent knowledge from being personalized.

[0621] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[0622] "Issues of concern" refer to issues, challenges, risks, etc. that are subject to discussion.

[0623] "Introducing the discussion" is the process of preparing and setting up the discussion.

[0624] "Reducing labor hours" refers to reducing the time and effort required for work.

[0625] "Accelerating discussions" refers to moving discussions forward more quickly.

[0626] "Preventing oversights" refers to making sure that important matters are not overlooked during the discussion or deliberation process.

[0627] A "minutes" is a document that records the content and results of a discussion.

[0628] "Specialist knowledge" refers to advanced knowledge or skills in a particular field.

[0629] "Personalization of knowledge" refers to a state in which knowledge or skills become dependent on specific individuals.

[0630] "Handover" is the process of transferring work and knowledge to the next person in charge.

[0631] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means for automatically saving the results of the discussion as minutes, and a means for multiple AIs with specialized knowledge to work together to propose optimal solutions.

[0632] Hardware and Software Used

[0633] Hardware: Smartphone

[0634] Software: Python, JSON

[0635] Data processing and calculation

[0636] The server uses a generation AI that is assigned a role based on the stance of each department to hold discussions. Specifically, AIs with specialized knowledge, such as maintenance AI, construction AI, and head office AI, work together to identify concerns and introduce discussions. The results of the discussion are automatically saved as minutes and can be viewed on smartphones.

[0637] Examples

[0638] For example, if a factory manager is considering installing a new production line, he or she could begin the discussion with the following prompt:

[0639] Examples of prompt statements

[0640] issue = "Introduction of new production line"

[0641] solutions = manager.conduct_meeting(issue)

[0642] print(solutions)

[0643] By executing this prompt, the maintenance AI will propose "solutions based on network maintenance," and the construction AI will propose "solutions based on network construction." This allows factory managers to check optimal production plans and maintenance schedules on their smartphones.

[0644] This system makes it possible to efficiently and quickly arrive at optimal solutions, automatically create meeting minutes, and prevent knowledge from becoming personalized.

[0645] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0646] Step 1:

[0647] The server initializes the generation AI, which is given the stance of each department as a role. Specifically, it creates instances of maintenance AI, construction AI, headquarters AI, etc., and sets their respective expertise. The name and expertise of each AI are required as input, and the initialized AI instance is obtained as output.

[0648] Step 2:

[0649] The user inputs the concern to be discussed, for example, setting the agenda "Installation of a new production line." As input, the concern to be discussed is required, and as output, the agenda is obtained.

[0650] Step 3:

[0651] The server provides each AI with an agenda and asks them to propose a solution. Each AI generates a solution based on its own expertise. The inputs are the agenda and the expertise of each AI, and the output is the solution proposed by each AI.

[0652] Step 4:

[0653] The server collects the solutions obtained from each AI and selects the optimal solution. As input, it needs the solution of each AI, and as output, it gets the optimal solution.

[0654] Step 5:

[0655] The server automatically saves the results of the discussion as minutes. Specifically, it saves the agenda, each AI's solution, and the optimal solution in JSON format. The inputs are the agenda, each AI's solution, and the optimal solution, and the output is the saved minutes.

[0656] Step 6:

[0657] The user checks the saved minutes using a smartphone. The saved minutes are required as input, and the contents of the minutes are displayed on the smartphone as output.

[0658] In addition, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0659] "Example 1"

[0660] In one embodiment of the present invention, the generating AI means includes an emotion engine that recognizes the user's emotions. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generating AI. The generating AI adjusts the progress of the discussion based on this emotion information. For example, if the user shows anger or dissatisfaction, the generating AI pauses the progress of the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generating AI smoothly progresses the discussion and actively incorporates the user's opinion.

[0661] "Example 2"

[0662] In another embodiment of the present invention, the emotion engine adjusts the results of the discussion according to the user's emotions. For example, if the user is dissatisfied, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user is dissatisfied with the quality of the network, the generation AI prioritizes discussions on improving the quality and reports the results to the user. This enables discussions that take the user's emotions into account.

[0663] The process flow of each embodiment will be described below.

[0664] "Example 1"

[0665] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[0666] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[0667] Step 3: The generation AI adjusts the progress of the discussion based on the emotion information. Specifically, if the user shows anger or dissatisfaction, the generation AI pauses the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generation AI smoothly advances the discussion and actively incorporates the user's opinion.

[0668] "Example 2"

[0669] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[0670] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[0671] Step 3: The generation AI adjusts the results of the discussion based on the emotion information. Specifically, if the user expresses dissatisfaction, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user expresses dissatisfaction with the quality of the network, the generation AI prioritizes discussions on improving quality and reports the results to the user.

[0672] Example 1

[0673] Next, a description will be given of Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0674] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and move the discussion forward. In addition, the user's feelings cannot be properly reflected during the discussion, which can lead to a decline in the quality of the discussion. Furthermore, it is difficult to quickly collect the results of the discussion and provide them as a report. In order to solve these problems, an efficient discussion system using generative AI that has the stance of each department is needed.

[0675] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0676] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for adjusting the progress of discussions using an emotion engine that recognizes the emotions of users, and a means for aggregating the results of discussions and generating reports. This makes it possible to efficiently aggregate the opinions of each department, progress discussions while appropriately reflecting the emotions of users, and quickly provide the results of discussions as reports.

[0677] "Generative AI means" refers to artificial intelligence generated to hold discussions between roles that are assigned the stance of each department.

[0678] "Means of identifying concerns and initiating discussions without human intervention" refers to a process whereby the system automatically identifies concerns and initiates discussions.

[0679] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to methods for increasing the efficiency of discussions and covering all necessary items without missing anything.

[0680] An "emotion engine" refers to technology that analyzes emotions from a user's voice, facial expressions, text input, etc., and provides the results to generative AI.

[0681] The "means for adjusting the progress of the discussion" refers to a method for appropriately adjusting the progress of the discussion based on the user's emotion information obtained from the emotion engine.

[0682] "Means of aggregating the results of discussions and generating a report" refers to the process for summarizing the results of each AI's discussions and creating a final report.

[0683] "Conservation AI" refers to artificial intelligence that takes the stance of the maintenance department and engages in discussions about maintenance.

[0684] "Construction AI" refers to artificial intelligence that takes a stance on the construction sector and engages in discussions about construction.

[0685] "Headquarters AI" refers to artificial intelligence that takes the stance of the headquarters department and engages in discussions regarding overall management and strategy.

[0686] "User" refers to a person who accesses the system to advance a discussion.

[0687] "Server" refers to a computer system for executing various means such as generative AI means and emotion engines.

[0688] This invention is a system that holds discussions using generation AI that is given a role based on the stance of each department. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs hold discussions with each other. The content of the discussions is related to improving network quality and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0689] Hardware and Software Used

[0690] The server is a computer system for running the generative AI model. The generative AI model is built using machine learning libraries such as TensorFlow and PyTorch. The server also includes an emotion engine for analyzing the user's emotions. The emotion engine uses voice recognition software and image analysis software to analyze the user's voice, facial expressions, text input, etc.

[0691] A terminal is a device through which a user accesses the system, and includes a PC, a smartphone, etc. A user accesses the system using a web browser and inputs a topic for discussion.

[0692] Program Processing

[0693] The server detects user access and launches a generative AI model. The generative AI model includes pre-trained AI algorithms. The server obtains the necessary information from each department's database to generate an AI with the stance of each department. For example, a maintenance AI is generated based on past maintenance records.

[0694] The server sets the topic of discussion provided by the user. For example, it sets the topic, "What measures are necessary to improve the quality of the network?" The server sends a prompt to each AI to start the discussion. Each AI expresses its opinion based on its own stance.

[0695] The server uses an emotion engine to analyze the user's emotions. Specifically, it analyzes the user's voice, facial expressions, text input, etc. to grasp the user's emotional state. The server adjusts the progress of the discussion based on the user's emotional information obtained from the emotion engine. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[0696] The server aggregates the results of each AI's discussion and generates a report. The report includes each AI's opinion and the conclusion of the discussion. The user can use the terminal to check the generated report and provide feedback if necessary.

[0697] Examples of concrete examples and prompts

[0698] As a concrete example, consider the case of discussing how to improve the quality of a network. The server provides the following topics to the maintenance AI, construction AI, and head office AI:

[0699] "What measures are necessary to improve the quality of the network?"

[0700] Each AI debates as follows:

[0701] Maintenance AI: "The aging of current network equipment is an issue, and regular maintenance is necessary."

[0702] Construction AI: "We need to introduce new network equipment, especially the latest communication technology."

[0703] Headquarters AI: "We need to consider the cost. We need to take the best measures within our budget."

[0704] The emotion engine analyzes the user's emotions, and if the user indicates dissatisfaction, the generative AI pauses the discussion and prioritizes the user's opinion.

[0705] Example prompt:

[0706] "What measures are necessary to improve the quality of the network? Please give us your opinions on maintenance AI, construction AI, and headquarters AI."

[0707] In this way, the server, terminals, and users work together to execute system processing.

[0708] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0709] Step 1:

[0710] A user accesses the system

[0711] A user accesses the system using a terminal (such as a PC or smartphone). Specifically, the user opens a web browser, enters the system's URL, and logs in. The user's authentication information (user name, password) is required as input, and the system's dashboard is displayed as output.

[0712] Step 2:

[0713] The server runs the generative AI model

[0714] The server detects user access and launches the generative AI model. The user's login information is required as input, and the generative AI model is launched as output. The generative AI model is built using machine learning libraries such as TensorFlow and PyTorch.

[0715] Step 3:

[0716] The server generates AI with a stance for each division.

[0717] The server generates AI with the stance of the maintenance department, construction department, head office department, etc. As input, information obtained from the database of each department is required, and as output, a maintenance AI, a construction AI, a head office AI, etc. For example, a maintenance AI is generated based on past maintenance records.

[0718] Step 4:

[0719] The server sets the topic of discussion

[0720] The server sets the topic of discussion provided by the user. As input, it requires the topic entered by the user (e.g., "What measures are necessary to improve the quality of the network?"), and as output, a prompt sentence is generated for each AI.

[0721] Step 5:

[0722] The server starts the discussion for each AI.

[0723] The server has each AI start a discussion based on the set theme. A prompt is required as input, and the opinion of each AI is generated as output. For example, the maintenance AI may say, "The aging of the current network equipment is a problem, and regular maintenance is necessary."

[0724] Step 6:

[0725] The server uses an emotion engine to analyze the user's emotions.

[0726] The server uses an emotion engine to analyze the user's emotions. The inputs required are the user's voice, facial expressions, text input, etc., and the output is the analyzed user's emotional state. Specifically, the analysis is performed using voice recognition software and image analysis software.

[0727] Step 7:

[0728] The server coordinates the progress of the discussion.

[0729] The server adjusts the progress of the discussion based on the user's emotion information obtained from the emotion engine. The user's emotion information is required as input, and the progress of the discussion is adjusted as output. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[0730] Step 8:

[0731] The server aggregates the results of the discussions and generates a report.

[0732] The server aggregates the results of each AI's discussion and generates a report. The input is the results of each AI's discussion, and the output is a PDF report. The report contains detailed descriptions of each AI's opinions and the conclusions of the discussion.

[0733] Step 9:

[0734] User checks the report

[0735] The user uses a terminal to review the generated report. As input, the report downloaded through the system interface is required, and as output, the report content is displayed. The user can also provide feedback on the report content.

[0736] (Application example 1)

[0737] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal".

[0738] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and advance the discussion. In addition, because the discussion proceeds without considering the user's feelings, the user's dissatisfaction or anger may hinder the progress of the discussion. Furthermore, there is a problem that the efficiency of the discussion decreases due to a lack of means to check the progress of the discussion in real time and intervene if necessary. In order to solve these issues, a system is needed that uses a generative AI that has the stance of each department to hold the discussion, recognizes the user's feelings, and adjusts the progress of the discussion.

[0739] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0740] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion through a smartphone or a head-mounted display and intervening as necessary. This makes it possible to efficiently collect the opinions of each department and progress the discussion while taking into account the user's emotions, thereby improving the efficiency and effectiveness of the discussion.

[0741] "Generative AI means" is an artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0742] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically analyzes the content of discussions, extracts important concerns, and starts discussions.

[0743] "Means of reducing labor costs, accelerating discussions, and preventing items from being overlooked" are techniques for increasing the efficiency of discussions and ensuring that all necessary items are not overlooked.

[0744] An "emotion engine" is a technology that analyzes emotions from a user's voice, facial expressions, text input, etc.

[0745] The "means for adjusting the progress of the discussion based on emotional information" is a technique for appropriately controlling the progress of the discussion depending on the emotional state of the users.

[0746] "Means of checking the progress of the discussion through a smartphone or head-mounted display and intervening if necessary" refers to technology that uses mobile devices or wearable devices to monitor the progress of the discussion in real time and allows users to intervene if necessary.

[0747] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion via a smartphone or head-mounted display and intervening if necessary.

[0748] System Programming

[0749] The server implements a generative AI model using the OpenAI API to generate an AI with the stance of each department. This allows discussions to be held from the perspectives of the maintenance department, manufacturing department, quality control department, etc. To recognize the user's emotions, the EmotionRecognizer library is used to capture the user's facial expressions through a webcam. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generative AI. The generative AI adjusts the progress of the discussion based on this emotional information.

[0750] Hardware and Software

[0751] Hardware: Smartphone, head-mounted display, webcam

[0752] Software: OpenAI API, OpenCV, EmotionRecognizer library

[0753] Data processing and calculation

[0754] The server uses a generative AI model to generate discussions from each department's perspective. To recognize the user's emotions, a webcam is used to capture facial expressions in real time and the emotions are analyzed with the EmotionRecognizer library. The analyzed emotion information is provided to the generative AI and used to adjust the progress of the discussion. Through a smartphone or head-mounted display, the user can check the progress of the discussion in real time and intervene if necessary.

[0755] Examples

[0756] For example, if you are discussing the introduction of a new manufacturing line, you might use the following prompt:

[0757] "Discuss from the maintenance department's perspective: Discuss concerns regarding the implementation of a new production line."

[0758] "Discuss from the manufacturing perspective: Discuss the concerns regarding the implementation of a new manufacturing line."

[0759] "Discuss from the quality control department's perspective: Discuss concerns regarding the implementation of a new production line."

[0760] This makes it possible to efficiently collect opinions from each department and conduct discussions that take into account the feelings of users, thereby improving the efficiency and effectiveness of discussions.

[0761] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0762] Step 1:

[0763] The server initializes the generative AI model using the OpenAI API. As input, it receives a prompt sentence with the stance of each department and passes it to the generative AI model. The generative AI model generates a discussion from the perspective of each department and returns the discussion content in text format as output.

[0764] Step 2:

[0765] The server captures the user's facial expressions in real time using a webcam. It passes the captured image data as input to the EmotionRecognizer library, which analyzes the image data and outputs the user's emotional information.

[0766] Step 3:

[0767] The server provides the generative AI model with emotion information obtained from the EmotionRecognizer library. As input, the generative AI model passes the emotion information and the content of the discussion to adjust the progress of the discussion. The generative AI model appropriately controls the progress of the discussion based on the emotion information and returns the adjusted content of the discussion as output.

[0768] Step 4:

[0769] The terminal (smartphone or head-mounted display) receives the coordinated discussion content from the server. As input, it receives the discussion content from the server and displays it to the user. The user can check the progress of the discussion in real time and intervene if necessary.

[0770] Step 5:

[0771] Users intervene in the discussion through a smartphone or a head-mounted display. As input, users' opinions and feedback are entered into the device and sent to the server. The server receives the users' opinions and passes them back to the generative AI model to update the discussion content. As output, the updated discussion content is displayed again on the device.

[0772] Step 6:

[0773] The server saves the progress of the discussion as a log. As input, it receives the discussion content and emotion information generated at each step and saves them in a database. This allows the discussion history to be checked later.

[0774] Example 2

[0775] Next, a description will be given of Example 2 of the embodiment 2. In the following description, the data processing device 12 is referred to as a "server" and the smart glasses 214 are referred to as a "terminal".

[0776] In conventional discussion systems, human intervention was required to hold discussions that reflected the stance of each department, which not only took time but also often slowed down the progress of discussions. In addition, manual work was required to save the results of discussions, which led to problems with personalizing knowledge and handover. Furthermore, since discussions did not take into account the feelings of users, user satisfaction could decline.

[0777] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0778] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing man-hours, accelerating discussions, and preventing oversights, an emotion engine means for adjusting the results of discussions according to the emotions of users, and a means for saving the results of discussions as meeting minutes. This makes it possible to streamline discussions, prevent knowledge from becoming personal, and hold discussions that take into account the emotions of users.

[0779] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0780] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically initiates discussions and identifies key issues.

[0781] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[0782] "Emotion engine means" refers to technology for adjusting the outcome of a discussion according to the user's emotions.

[0783] "Means of saving the results of the discussion as meeting minutes" refers to technology for automatically recording and saving the content and results of the discussion.

[0784] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[0785] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[0786] "Headquarters AI" refers to artificial intelligence that understands the overall policies and strategies of the company and holds discussions based on them.

[0787] This invention is a system that includes a generation AI means for holding discussions between roles that are assigned roles that represent the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, an emotion engine means for adjusting the results of the discussion according to the user's emotions, and a means for saving the results of the discussion as meeting minutes.

[0788] Generating a Program

[0789] The server generates the system's programs, which include the maintenance AI, construction AI, headquarters AI, and emotion engine. These programs allow each AI to hold discussions based on their respective expertise and propose optimal solutions.

[0790] Program processing explanation

[0791] The server processes the data in the following steps:

[0792] 1. Data collection:

[0793] The server collects data on network maintenance status, construction plans, and company policies and strategies. This data serves as the basis for each AI's discussion. Specifically, the server obtains data from network monitoring tools (e.g., Nagios) and project management software (e.g., JIRA).

[0794] 2. Start the discussion:

[0795] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson).

[0796] 3. Discussion process:

[0797] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on maintaining the network, while the construction AI proposes a plan to build a new network. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center."

[0798] 4. Save the results of the discussion:

[0799] The server saves the results of the discussion as minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL).

[0800] 5. Emotion engine intervention:

[0801] When a user expresses dissatisfaction, the server activates an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, if a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality.

[0802] 6. Reporting results:

[0803] The server reports the final discussion result to the user. Specifically, the result is displayed on the user's dashboard (e.g., a web application). For example, the report may be in the form of "We will implement the following measures to improve the quality of the network."

[0804] Examples of concrete examples and prompts

[0805] Examples:

[0806] If a user is dissatisfied with the quality of the network, the following prompt sentence is fed into the generative AI model:

[0807] Example prompt:

[0808] "Users are complaining about the quality of their network. Please prioritize discussions on how to improve the quality and report the results."

[0809] Based on this prompt, the generation AI holds a discussion about improving quality, with the maintenance AI proposing, "We should introduce a new monitoring tool to strengthen network monitoring," and the construction AI proposing, "We should improve network throughput by introducing a new router." The server displays the results of the discussion on the user's dashboard, reporting, "We will introduce a new monitoring tool and router to improve network quality."

[0810] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0811] Step 1:

[0812] The server collects data on network maintenance status, construction plans, and company policies and strategies. Specifically, it obtains data from network monitoring tools (e.g. Nagios) and project management software (e.g. JIRA). As input, it receives data from these tools, and as output, it generates the basic information for each AI to discuss.

[0813] Step 2:

[0814] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson). As input, it provides each AI with the collected data, and as output, it performs initial settings for each AI to start a discussion.

[0815] Step 3:

[0816] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on network maintenance, while the construction AI proposes a new network construction plan. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center." As input, it receives suggestions from each AI, and as output, it generates an optimal solution.

[0817] Step 4:

[0818] The server saves the results of the discussion as the minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL). It receives the results of the discussion as input, and generates minutes as output and saves them in the database.

[0819] Step 5:

[0820] When a user expresses dissatisfaction, the server launches an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, when a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality. It receives the user's emotion data as input and generates instructions to adjust the outcome of the discussion as output.

[0821] Step 6:

[0822] The server reports the final discussion result to the user. Specifically, it displays the result on the user's dashboard (e.g., a web application). For example, it reports in the form of "We will implement the following measures to improve the quality of the network." It receives the final discussion result as input and generates information to report to the user as output.

[0823] (Application example 2)

[0824] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the smart glasses 214 will be referred to as a "terminal".

[0825] In conventional network security systems, it is difficult to integrate the expertise of each department to derive optimal solutions, and there is a lack of consideration for user feelings, making it difficult to improve user satisfaction. Furthermore, efficient discussion and coordination is required in detecting security risks, proposing countermeasures, and designing and implementing new security infrastructure.

[0826] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0827] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, an emotion engine means for analyzing user emotions and adjusting the results of discussions, a security preservation AI means for detecting security risks and proposing measures, a security construction AI means for proposing the design and implementation of new security infrastructure, and a headquarters AI means for understanding the security policy and strategy of the entire company and adjusting the overall security strategy. This makes it possible to integrate the expertise of each department and propose optimal security measures that take into account the emotions of users.

[0828] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[0829] An "emotion engine" is an engine that has the function of analyzing the user's emotions and adjusting the results of the discussion according to those emotions.

[0830] "Security preservation AI" is an artificial intelligence that has knowledge and experience in network preservation, detects security risks, and proposes countermeasures.

[0831] "Security Construction AI" is an artificial intelligence that has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods.

[0832] "Headquarters AI" is an artificial intelligence that understands the security policies and strategies of the entire company and adjusts the overall security strategy.

[0833] A "role" is a virtual role with expertise and experience that is assigned to each department's stance.

[0834] "Introducing a discussion" is the process of identifying concerns and starting a discussion without human intervention.

[0835] "Reducing labor hours" refers to measures aimed at reducing the time and effort required for work.

[0836] "Accelerating discussions" is a measure aimed at speeding up the progress of discussions.

[0837] "Preventing omissions in consideration" is a measure to ensure that important matters are not overlooked during discussion or consideration.

[0838] A system for implementing this invention is configured as follows: The server includes a generation AI that holds discussions between roles that are assigned the stance of each department, an emotion engine that analyzes user emotions and adjusts the results of the discussion, a security preservation AI that detects security risks and proposes countermeasures, a security construction AI that proposes the design and implementation of new security infrastructure, and a headquarters AI that understands the security policy and strategy of the entire company and adjusts the overall security strategy.

[0839] Hardware and Software Configuration

[0840] The server uses the following hardware and software:

[0841] Hardware: High-performance server (e.g. Dell PowerEdge series)

[0842] Software: Python, sentiment analysis libraries (e.g. TextBlob), generative AI models (e.g. GPT-3)

[0843] Data processing and calculation

[0844] The server performs the following data processing and calculations.

[0845] 1. Generative AI: Each department is assigned a role based on their stance, and the roles are discussed to propose optimal solutions. A generative AI model (e.g. GPT-3) is used to generate the content of the discussion.

[0846] 2. Sentiment Engine: Analyze user sentiment and adjust the discussion outcome depending on the sentiment. Use a sentiment analysis library (e.g. TextBlob) to analyze user feedback.

[0847] 3. Security Preservation AI: With knowledge and experience in network preservation, it detects security risks and proposes countermeasures. It uses generative AI models to detect risks and propose countermeasures.

[0848] 4. Security Construction AI: Has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods. Uses generative AI models to make design and implementation suggestions.

[0849] 5. HQ AI: Understand the security policies and strategies of the entire enterprise and adjust the overall security strategy. Use generative AI models to make adjustments to the strategy.

[0850] Examples

[0851] For example, if a user provides feedback like "I'm dissatisfied with the current security measures," the emotion engine will analyze this feedback and recognize that the user is expressing dissatisfaction. Based on this, the security preservation AI will suggest updating firewall rules.

[0852] Examples of prompt statements

[0853] User feedback: "I'm not satisfied with the current security measures."

[0854] AI discussion result: "Security Preservation AI proposes updating firewall rules."

[0855] In this way, it becomes possible to integrate the expertise of each department and propose optimal security measures that take into account the user's feelings.

[0856] The flow of the specific process in the application example 2 will be described with reference to FIG.

[0857] Step 1:

[0858] The server receives feedback from the user. The input is the user's feedback text, and the output is the storage of the feedback text. For example, if the user enters "I am not satisfied with the current security measures," the server stores this text in a database.

[0859] Step 2:

[0860] The server parses the user feedback using an emotion engine. The input is the stored feedback text and the output is the user's emotional state. Specifically, it uses a sentiment analysis library (e.g., TextBlob) to extract the emotion "unhappy" from the feedback text.

[0861] Step 3:

[0862] The server sends a prompt to the generation AI based on the analysis results of the emotion engine. The input is the user's emotional state and feedback text, and the output is a prompt sentence for the generation AI. Specifically, it generates the prompt sentence "User feedback: 'I am dissatisfied with the current security measures.'"

[0863] Step 4:

[0864] The server uses a generative AI to hold a discussion and generate an optimal solution. The input is the prompt sentence, and the output is the result of the discussion. Specifically, a generative AI model (e.g., GPT-3) is used to generate the discussion result "Security preservation AI proposes updating firewall rules."

[0865] Step 5:

[0866] The server reports the generated discussion results to the user. The input is the discussion result, and the output is a report message to the user. Specifically, the server sends the message "Security preservation AI proposes updating firewall rules" to the user.

[0867] Step 6:

[0868] The server stores the results of the discussion in a database. The input is the discussion result and the output is the stored discussion result. Specifically, the generated discussion result is stored in the database for future reference and analysis.

[0869] In this way, the server performs a series of processes: receiving user feedback, analyzing emotions, using generative AI to propose optimal solutions, and reporting the results to the user.

[0870] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires a voice indicating a user input for the result of the specific processing. The control unit 46A transmits the voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[0871] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0872] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[0873] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0874] [Third embodiment]

[0875] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0876] 5, the data processing system 310 includes the data processing device 12 and a headset type terminal 314. An example of the data processing device 12 is a server.

[0877] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[0878] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0879] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[0880] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[0881] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[0882] Fig. 6 shows an example of main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0883] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0884] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0885] In the headset type terminal 314, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0886] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[0887] "Example 1"

[0888] The present invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0889] "Example 2"

[0890] As a concrete example, the maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies. These AIs hold discussions based on their respective stances and propose optimal solutions. This discussion is conducted without human intervention, and the results of the discussion are saved as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0891] The process flow of each embodiment will be described below.

[0892] "Example 1"

[0893] Step 1: The system generates AIs with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.).

[0894] Step 2: The generated AIs begin to discuss with each other. The content of the discussions will be related to improving network quality and achieving the Ministry of Internal Affairs and Communications' objectives.

[0895] Step 3: The results of the discussion are saved as meeting minutes and can be referenced later.

[0896] Step 1: The maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies.

[0897] Step 2: These AIs debate based on their respective stances and propose the optimal solution.

[0898] Step 3: The discussion is conducted without human intervention, and the results are stored as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[0899] Example 1

[0900] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[0901] In the conventional discussion system, it was difficult to hold discussions that reflected the stance of each department, and there was a problem that it took a lot of time to identify concerns and start discussions. In addition, the human intervention hindered the acceleration of discussions and increased the risk of overlooking something. Furthermore, the necessity of personalizing knowledge and handover made it difficult to hold efficient discussions.

[0902] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0903] In this invention, the server includes a generation AI means for discussion between roles that are given the stance of each department, a means for a user to input a prompt text, a means for the server to receive the prompt text and select an appropriate generation AI model, a means for the server to send the prompt text to the generation AI model selected by the server, a means for the generation AI model to start a discussion and for each AI to give an opinion from its own perspective, a means for the server to collect and integrate the opinions of each AI, a means for the server to return the integrated discussion result to the user, a means for identifying concerns and introducing discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This enables efficient and comprehensive discussions.

[0904] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0905] A "prompt sentence" refers to an instruction sentence regarding the topic or subject of discussion that is entered by the user.

[0906] "Server" refers to a computer system for receiving prompts, selecting an appropriate generative AI model, and managing the discussion.

[0907] A "generative AI model" refers to an artificial intelligence model with a specific role or perspective, such as maintenance AI, construction AI, or headquarters AI.

[0908] "Conservation AI" refers to a generative AI model that has a maintenance department stance.

[0909] "Construction AI" refers to a generative AI model that has a construction sector stance.

[0910] "Headquarters AI" refers to a generative AI model that has the stance of the headquarters department.

[0911] "Discussion results" refers to the conclusions and proposals reached through discussions between generative AI models.

[0912] "Reducing labor hours" refers to reducing the time and effort required for work.

[0913] "Accelerating the discussion" refers to speeding up the discussion.

[0914] "Preventing gaps in consideration" refers to making sure that important matters are not overlooked in the discussion.

[0915] "Personalization of knowledge" refers to the concentration of knowledge and information in the hands of specific individuals.

[0916] "Handover" refers to the transfer of work or knowledge to the next person in charge.

[0917] This invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[0918] Hardware and Software Used

[0919] The server uses a high-performance server (e.g., a server equipped with an NVIDIA GPU). The software uses a generative AI model (e.g., GPT-4) and a discussion management system (e.g., a dedicated discussion framework).

[0920] Program Processing

[0921] The server receives a prompt from the user and selects an appropriate generative AI model. It then sends the prompt to the selected generative AI model, starting a discussion. The generative AI model starts a discussion with each AI, such as the maintenance AI, construction AI, and headquarters AI, and gives opinions from their own perspectives. The server collects and integrates the opinions of each AI. Finally, it returns the integrated discussion results to the user.

[0922] Examples

[0923] Using the terminal, the user inputs a prompt sentence related to the topic or agenda of the discussion. For example, the user inputs "Please start a discussion on improving network quality." The server receives this prompt sentence and analyzes it using a natural language processing algorithm. Based on the analysis results, the server selects a maintenance AI, a construction AI, or a headquarters AI. The server sends the prompt sentence to the selected generative AI model via API. The generative AI model receives the prompt sentence and begins preparing for the discussion.

[0924] The generative AI model initiates a discussion between the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to prioritize due to budget constraints." The server collects the opinions sent by each AI and integrates them using a natural language processing algorithm. This results in a comprehensive discussion result.

[0925] The server returns the integrated discussion results to the user, who can check the results on the screen of his / her device. For example, the conclusion "Strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed.

[0926] In this way, the server enables efficient and comprehensive discussions, reducing labor costs, accelerating discussions, and preventing items from being overlooked.

[0927] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0928] Step 1:

[0929] The user inputs a prompt text. The user inputs "Please start a discussion on improving the quality of the network" in the input field of the terminal and clicks the send button. The input prompt text is sent to the server.

[0930] Step 2:

[0931] The server receives the prompt sentence and selects an appropriate generative AI model. The server receives the prompt sentence and analyzes the prompt sentence using a natural language processing algorithm. Based on the analysis result, the maintenance AI, construction AI, or head office AI is selected. The input is the prompt sentence, and the output is the selected generative AI model.

[0932] Step 3:

[0933] The server sends a prompt to the selected generative AI model. The server sends the prompt to the selected generative AI model via API. The generative AI model receives the prompt and starts preparing for discussion. The input is the selected generative AI model and the prompt, and the output is a state ready for discussion.

[0934] Step 4:

[0935] The generative AI model starts a discussion, and each AI gives an opinion from its own perspective. The generative AI model starts a discussion with the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to set priorities due to budget constraints." The input is the prompt text and the role of each AI, and the output is the opinion of each AI.

[0936] Step 5:

[0937] The server collects and integrates the opinions of each AI. The server collects the opinions sent by each AI and integrates the opinions using a natural language processing algorithm. This results in a comprehensive discussion result. The input is the opinion of each AI, and the output is the integrated discussion result.

[0938] Step 6:

[0939] The server returns the integrated discussion results to the user. The server returns the integrated discussion results to the user. The user can check the results of the discussion on their terminal screen. For example, the conclusion "strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed. The input is the integrated discussion results, and the output is the discussion results that are displayed to the user.

[0940] (Application example 1)

[0941] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[0942] Conventional discussion systems had problems in that it was difficult to gather the opinions of each department and proceed with discussions efficiently, which required a lot of man-hours and made it easy for considerations to be overlooked. Also, in discussions on how to improve production efficiency within factories, there was a lack of a way to generate AI with the stance of each department and have those AIs discuss with each other. Furthermore, the implementation as an application to be installed on smartphones was also insufficient.

[0943] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a generating AI means for discussion between roles that are given the stance of each department as a role, a means for identifying concerns and initiating discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This makes it possible to generate AI with the stance of each department in a factory, and for these AIs to discuss ways to improve production efficiency. In addition, by implementing it as an application to be installed on a smartphone, users can easily use it.

[0944] A "generative AI means" is an artificial intelligence means that is generated to hold discussions between roles that are assigned the stance of each department.

[0945] "Means for identifying concerns and initiating discussions without human intervention" refers to a means for automatically extracting concerns or issues that arise during the discussion process and starting a discussion.

[0946] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" are measures to improve the efficiency of discussions and cover all necessary items without missing anything.

[0947] "A means of generating AI that has the perspective of each department within a factory and having these AIs discuss with each other how to improve production efficiency" refers to a means of generating artificial intelligence that has the perspective of each department within a factory (production, quality control, maintenance, etc.) and having these AIs discuss with each other how to improve production efficiency.

[0948] "Means including an application installed on a smartphone" refers to means implemented as an application installed on a smartphone.

[0949] "Roles of maintenance AI, construction AI, headquarters AI, production AI, quality control AI, maintenance AI, etc." are the roles of artificial intelligence that have the responsibilities and perspectives of each department.

[0950] As an embodiment of the present invention, we will create an AI that has the stance of each department in a factory, and build a system in which these AIs discuss ways to improve production efficiency. The specific embodiment will be described below.

[0951] System Configuration

[0952] The system is composed of a server, a smartphone, a generative AI model, and a prompt. The server provides the generative AI means for discussions between roles that are assigned the stances of each department. The smartphone is a device on which users can install applications that allow them to easily use the system.

[0953] Hardware and Software Used

[0954] Hardware: Servers, smartphones

[0955] Software: Python, OpenAI API

[0956] Data processing and calculation

[0957] The server processes and calculates data in the following procedure.

[0958] 1. Using generative AI models:

[0959] The server uses the OpenAI API to generate an AI with the stance of each department. Specifically, it inputs a prompt sentence and generates an AI with the stance of each department.

[0960] 2. Conducting the discussion:

[0961] The server holds discussions between the generated AIs of each department, creates discussion prompts, and uses the OpenAI API to obtain the results of the discussion.

[0962] 3. Display the results:

[0963] The results of the discussion are displayed to the user through an application installed on their smartphone.

[0964] Examples

[0965] Below are some examples of prompt sentences.

[0966] Generate an AI with a production sector stance.

[0967] Generate an AI with a quality control department stance.

[0968] Generate an AI with a maintenance department stance.

[0969] Please discuss with the AIs below how to improve production efficiency.

[0970] By using this prompt, it is possible to generate AI with the stance of each department and hold discussions. For example, the production department's AI can propose "what kind of data should be collected to identify bottlenecks in the production line," while the quality control department's AI can discuss "what kind of inspection method should be introduced as a measure to improve quality control."

[0971] In this way, a system is realized that enables efficient discussions on how to improve production efficiency within a factory.

[0972] The flow of the specific process in the application example 1 will be described with reference to FIG.

[0973] Step 1:

[0974] The server creates a prompt sentence to generate an AI with the stance of each department. Specifically, it generates a prompt sentence like the following based on the department specified by the user (e.g. production, quality control, maintenance): "Please generate an AI with the stance of the production department." This prompt sentence is the input, and the generated AI is the output.

[0975] Step 2:

[0976] The server sends the generated prompt sentence to OpenAI's API to generate an AI with the stance of each department. Specifically, it calls OpenAI's API and provides the prompt sentence as input. The API uses the generative AI model to generate an AI with the stance of each department and returns the result to the server. This result becomes the output.

[0977] Step 3:

[0978] The server stores the generated AI for each department in a list. Specifically, the AI ​​returned from the API is saved in list format and used in the subsequent discussion process. This list is the input, and the stored AI list is the output.

[0979] Step 4:

[0980] The server creates a prompt for the AIs in each department to discuss with each other. Specifically, it generates the following prompt based on the AIs stored in the list: "Please hold a discussion between the AIs below to improve production efficiency." This prompt is the input, and the results of the discussion are the output.

[0981] Step 5:

[0982] The server sends the discussion prompt to the OpenAI API and obtains the discussion result. Specifically, it calls the API and provides the discussion prompt as input. The API uses a generative AI model to generate the discussion result and returns the result to the server. This result becomes the output.

[0983] Step 6:

[0984] The server transmits the acquired results of the discussion to an application installed on the smartphone. Specifically, it converts the results of the discussion into a format that the user can view and transmits them to the smartphone. The results of this discussion are the input, and data in a format that the user can view is the output.

[0985] Step 7:

[0986] The user checks the results of the discussion displayed on the smartphone and makes additional discussions or adjustments as necessary. Specifically, the user views the results of the discussion through the smartphone application and decides on the next action. The results of this discussion are the input, and the user's actions are the output.

[0987] Example 2

[0988] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[0989] In conventional discussion systems, it was necessary for people with specialized knowledge from each department to directly participate in the discussions, which led to problems with personalizing knowledge and handover. In addition, it took a lot of time to progress the discussions and record the results, making it difficult to accelerate the discussions and prevent oversights. This made it difficult to hold efficient discussions and propose optimal solutions.

[0990] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0991] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for a user to input the topic of discussion, a means for the server to start a discussion and collect opinions from each AI model, and a means for automatically generating the results of the discussion as minutes and saving them in a database. This makes it possible to have efficient discussions and propose optimal solutions without the need for personalizing or handing over knowledge.

[0992] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[0993] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically extracts discussion topics and concerns and starts discussions.

[0994] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[0995] "Means for users to input a topic for discussion" refers to an interface that allows users to input a topic for discussion into the system.

[0996] "Means for the server to initiate discussions and aggregate opinions from each AI model" refers to technology that enables the server to initiate discussions and collect and integrate opinions from each AI model.

[0997] "Means for automatically generating the results of discussions as minutes and storing them in a database" refers to technology for automatically recording the content and results of discussions in text format and storing them in a database.

[0998] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[0999] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[1000] "Headquarters AI" refers to artificial intelligence that understands the company's overall policies and strategies.

[1001] "Database" refers to an information management system for storing the results of discussions and minutes.

[1002] This invention is a system that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions. Specifically, three AI models, a maintenance AI, a construction AI, and a headquarters AI, hold discussions using their respective expertise, and the results are automatically saved as minutes.

[1003] Hardware and Software Used

[1004] Hardware: Servers, terminals

[1005] Software: Generative AI models (e.g. GPT-4), database management systems (e.g. MySQL)

[1006] Generating a Program

[1007] The server generates programs including the maintenance AI, construction AI, and headquarters AI models. The programs are implemented using programming languages ​​such as Python and Java. Each AI model is based on a generative AI model (e.g., GPT-4) and is trained to have its own specialized knowledge.

[1008] Program Processing

[1009] The user inputs the topic of discussion

[1010] The user inputs the topic of discussion into the generative AI model using the device. For example, if the user wants to input the topic "optimal strategy for designing new network infrastructure," the user writes the topic in the input form on the device and clicks the send button.

[1011] The server starts the discussion.

[1012] The server receives the topic entered by the user and has the maintenance AI, construction AI, and headquarters AI start a discussion. Specifically, the server generates a prompt for each AI model and sends an API request to hold a discussion. For example, the server generates a prompt like the one below.

[1013] Example prompt:

[1014] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[1015] The server aggregates the results of the discussion.

[1016] The server aggregates the opinions returned by each AI model. Specifically, it analyzes the responses from each AI model and applies an algorithm to propose the optimal solution. For example, if the maintenance AI suggests that "a backup system should be installed to ensure network redundancy" and the construction AI suggests that "fiber optic cables should be used for cost-effectiveness," the server integrates these opinions to derive the optimal solution of "using fiber optic cables and installing a backup system."

[1017] The server generates the minutes.

[1018] The server automatically generates the results of the discussion as minutes. Specifically, it summarizes the contents of the discussion and the final solution in text format and saves them in a database management system (e.g., MySQL). Users can later refer to these minutes. For example, when a user clicks the "Show minutes" button on a terminal, the server retrieves the minutes from the database and displays them on the terminal.

[1019] Examples

[1020] For example, if a user inputs the theme "optimal strategy for designing new network infrastructure," the following prompt sentence is input to the generative AI model.

[1021] Example prompt:

[1022] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[1023] Based on this prompt, the server will have each AI start a discussion and propose the optimal solution. The results of the discussion are automatically saved as minutes, so the user can check the contents later.

[1024] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1025] Step 1:

[1026] Generating a Program

[1027] The server generates a program including each AI model: maintenance AI, construction AI, and headquarters AI. Specifically, using programming languages ​​such as Python and Java, each AI model is trained based on a generative AI model (e.g., GPT-4). Training data for each AI model is required as input, and each AI model is generated as output.

[1028] Step 2:

[1029] Enter the discussion topic

[1030] The user uses a terminal to input the topic of discussion into the generative AI model. Specifically, the user enters the topic in the input form on the terminal and clicks the send button. The input is the topic of discussion entered by the user, and the output is the topic sent to the server.

[1031] Step 3:

[1032] Starting the discussion

[1033] The server receives the topic input by the user and starts a discussion with the maintenance AI, construction AI, and headquarters AI. Specifically, the server generates a prompt for each AI model and sends an API request to hold the discussion. The input is the discussion topic received from the user, and the output is the generated prompt to be sent to each AI model.

[1034] Step 4:

[1035] Gathering opinions

[1036] The server collects the opinions returned from each AI model. Specifically, it analyzes the response from each AI model and obtains the opinions in text format. The input is the response from each AI model, and the output is the collected opinions of each AI model in text format.

[1037] Step 5:

[1038] Summarizing the results of discussions

[1039] The server proposes the optimal solution based on the collected opinions. Specifically, it integrates the opinions from each AI model and applies an algorithm to derive the optimal solution. The inputs are the opinions of each AI model, and the output is the proposed optimal solution.

[1040] Step 6:

[1041] Generate minutes

[1042] The server automatically generates the results of the discussion as minutes and saves them in a database. Specifically, the contents of the discussion and the final solution are summarized in text format and saved in a database management system (e.g., MySQL). The input is the results of the discussion, and the output is the minutes that are generated and saved in the database.

[1043] Step 7:

[1044] View minutes

[1045] A user refers to the minutes from a terminal. Specifically, when the user clicks the "Show minutes" button, the server retrieves the minutes from the database and displays them on the terminal. The input is a user request, and the output is the minutes displayed on the terminal.

[1046] (Application example 2)

[1047] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1048] In the conventional system, it took a lot of time and effort for specialists from each department to hold discussions and come up with optimal solutions. It was also time-consuming to accurately record the results of discussions and create meeting minutes for future reference. Furthermore, knowledge was easily tampered with and problems with handover occurred, making efficient operation difficult.

[1049] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1050] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for automatically saving the results of the discussion as minutes, and a means for AIs with multiple specialized knowledge to cooperate and propose optimal solutions. This makes it possible to efficiently and quickly derive optimal solutions, automatically create minutes, and prevent knowledge from being personalized.

[1051] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[1052] "Issues of concern" refer to issues, challenges, risks, etc. that are subject to discussion.

[1053] "Introducing the discussion" is the process of preparing and setting up the discussion.

[1054] "Reducing labor hours" refers to reducing the time and effort required for work.

[1055] "Accelerating discussions" refers to moving discussions forward more quickly.

[1056] "Preventing oversights" refers to making sure that important matters are not overlooked during the discussion or deliberation process.

[1057] A "minutes" is a document that records the content and results of a discussion.

[1058] "Specialist knowledge" refers to advanced knowledge or skills in a particular field.

[1059] "Personalization of knowledge" refers to a state in which knowledge or skills become dependent on specific individuals.

[1060] "Handover" is the process of transferring work and knowledge to the next person in charge.

[1061] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means for automatically saving the results of the discussion as minutes, and a means for multiple AIs with specialized knowledge to work together to propose optimal solutions.

[1062] Hardware and Software Used

[1063] Hardware: Smartphone

[1064] Software: Python, JSON

[1065] Data processing and calculation

[1066] The server uses a generation AI that is assigned a role based on the stance of each department to hold discussions. Specifically, AIs with specialized knowledge, such as maintenance AI, construction AI, and head office AI, work together to identify concerns and introduce discussions. The results of the discussion are automatically saved as minutes and can be viewed on smartphones.

[1067] Examples

[1068] For example, if a factory manager is considering installing a new production line, he or she could begin the discussion with the following prompt:

[1069] Examples of prompt statements

[1070] issue = "Introduction of new production line"

[1071] solutions = manager.conduct_meeting(issue)

[1072] print(solutions)

[1073] By executing this prompt, the maintenance AI will propose "solutions based on network maintenance," and the construction AI will propose "solutions based on network construction." This allows factory managers to check optimal production plans and maintenance schedules on their smartphones.

[1074] This system makes it possible to efficiently and quickly arrive at optimal solutions, automatically create meeting minutes, and prevent knowledge from becoming personalized.

[1075] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1076] Step 1:

[1077] The server initializes the generation AI, which is given the stance of each department as a role. Specifically, it creates instances of maintenance AI, construction AI, headquarters AI, etc., and sets their respective expertise. The name and expertise of each AI are required as input, and the initialized AI instance is obtained as output.

[1078] Step 2:

[1079] The user inputs the concern to be discussed, for example, setting the agenda "Installation of a new production line." As input, the concern to be discussed is required, and as output, the agenda is obtained.

[1080] Step 3:

[1081] The server provides each AI with an agenda and asks them to propose a solution. Each AI generates a solution based on its own expertise. The inputs are the agenda and the expertise of each AI, and the output is the solution proposed by each AI.

[1082] Step 4:

[1083] The server collects the solutions obtained from each AI and selects the optimal solution. As input, it needs the solution of each AI, and as output, it gets the optimal solution.

[1084] Step 5:

[1085] The server automatically saves the results of the discussion as minutes. Specifically, it saves the agenda, each AI's solution, and the optimal solution in JSON format. The inputs are the agenda, each AI's solution, and the optimal solution, and the output is the saved minutes.

[1086] Step 6:

[1087] The user checks the saved minutes using a smartphone. The saved minutes are required as input, and the contents of the minutes are displayed on the smartphone as output.

[1088] In addition, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1089] "Example 1"

[1090] In one embodiment of the present invention, the generating AI means includes an emotion engine that recognizes the user's emotions. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generating AI. The generating AI adjusts the progress of the discussion based on this emotion information. For example, if the user shows anger or dissatisfaction, the generating AI pauses the progress of the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generating AI smoothly progresses the discussion and actively incorporates the user's opinion.

[1091] "Example 2"

[1092] In another embodiment of the present invention, the emotion engine adjusts the results of the discussion according to the user's emotions. For example, if the user is dissatisfied, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user is dissatisfied with the quality of the network, the generation AI prioritizes discussions on improving the quality and reports the results to the user. This enables discussions that take the user's emotions into account.

[1093] The process flow of each embodiment will be described below.

[1094] "Example 1"

[1095] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[1096] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[1097] Step 3: The generation AI adjusts the progress of the discussion based on the emotion information. Specifically, if the user shows anger or dissatisfaction, the generation AI pauses the discussion and prioritizes the user's opinion. Also, if the user shows joy or satisfaction, the generation AI smoothly advances the discussion and actively incorporates the user's opinion.

[1098] "Example 2"

[1099] Step 1: The emotion engine receives information from the user, such as voice, facial expressions, and text input.

[1100] Step 2: The emotion engine analyzes the user's emotions and provides the results to the generative AI.

[1101] Step 3: The generation AI adjusts the results of the discussion based on the emotion information. Specifically, if the user expresses dissatisfaction, the generation AI adjusts the results of the discussion to be more in line with the user's opinion. Specifically, if the user expresses dissatisfaction with the quality of the network, the generation AI prioritizes discussions on improving quality and reports the results to the user.

[1102] Example 1

[1103] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1104] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and move the discussion forward. In addition, the user's feelings cannot be properly reflected during the discussion, which can lead to a decline in the quality of the discussion. Furthermore, it is difficult to quickly collect the results of the discussion and provide them as a report. In order to solve these problems, an efficient discussion system using generative AI that has the stance of each department is needed.

[1105] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1106] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for adjusting the progress of discussions using an emotion engine that recognizes the emotions of users, and a means for aggregating the results of discussions and generating reports. This makes it possible to efficiently aggregate the opinions of each department, progress discussions while appropriately reflecting the emotions of users, and quickly provide the results of discussions as reports.

[1107] "Generative AI means" refers to artificial intelligence generated to hold discussions between roles that are assigned the stance of each department.

[1108] "Means of identifying concerns and initiating discussions without human intervention" refers to a process whereby the system automatically identifies concerns and initiates discussions.

[1109] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to methods for increasing the efficiency of discussions and covering all necessary items without missing anything.

[1110] An "emotion engine" refers to technology that analyzes emotions from a user's voice, facial expressions, text input, etc., and provides the results to generative AI.

[1111] The "means for adjusting the progress of the discussion" refers to a method for appropriately adjusting the progress of the discussion based on the user's emotion information obtained from the emotion engine.

[1112] "Means of aggregating the results of discussions and generating a report" refers to the process for summarizing the results of each AI's discussions and creating a final report.

[1113] "Conservation AI" refers to artificial intelligence that takes the stance of the maintenance department and engages in discussions about maintenance.

[1114] "Construction AI" refers to artificial intelligence that takes a stance on the construction sector and engages in discussions about construction.

[1115] "Headquarters AI" refers to artificial intelligence that takes the stance of the headquarters department and engages in discussions regarding overall management and strategy.

[1116] "User" refers to a person who accesses the system to advance a discussion.

[1117] "Server" refers to a computer system for executing various means such as generative AI means and emotion engines.

[1118] This invention is a system that holds discussions using generation AI that is given a role based on the stance of each department. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs hold discussions with each other. The content of the discussions is related to improving network quality and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[1119] Hardware and Software Used

[1120] The server is a computer system for running the generative AI model. The generative AI model is built using machine learning libraries such as TensorFlow and PyTorch. The server also includes an emotion engine for analyzing the user's emotions. The emotion engine uses voice recognition software and image analysis software to analyze the user's voice, facial expressions, text input, etc.

[1121] A terminal is a device through which a user accesses the system, and includes a PC, a smartphone, etc. A user accesses the system using a web browser and inputs a topic for discussion.

[1122] Program Processing

[1123] The server detects user access and launches a generative AI model. The generative AI model includes pre-trained AI algorithms. The server obtains the necessary information from each department's database to generate an AI with the stance of each department. For example, a maintenance AI is generated based on past maintenance records.

[1124] The server sets the topic of discussion provided by the user. For example, it sets the topic, "What measures are necessary to improve the quality of the network?" The server sends a prompt to each AI to start the discussion. Each AI expresses its opinion based on its own stance.

[1125] The server uses an emotion engine to analyze the user's emotions. Specifically, it analyzes the user's voice, facial expressions, text input, etc. to grasp the user's emotional state. The server adjusts the progress of the discussion based on the user's emotional information obtained from the emotion engine. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[1126] The server aggregates the results of each AI's discussion and generates a report. The report includes each AI's opinion and the conclusion of the discussion. The user can use the terminal to check the generated report and provide feedback if necessary.

[1127] Examples of concrete examples and prompts

[1128] As a concrete example, consider the case of discussing how to improve the quality of a network. The server provides the following topics to the maintenance AI, construction AI, and head office AI:

[1129] "What measures are necessary to improve the quality of the network?"

[1130] Each AI debates as follows:

[1131] Maintenance AI: "The aging of current network equipment is an issue, and regular maintenance is necessary."

[1132] Construction AI: "We need to introduce new network equipment, especially the latest communication technology."

[1133] Headquarters AI: "We need to consider the cost. We need to take the best measures within our budget."

[1134] The emotion engine analyzes the user's emotions, and if the user indicates dissatisfaction, the generative AI pauses the discussion and prioritizes the user's opinion.

[1135] Example prompt:

[1136] "What measures are necessary to improve the quality of the network? Please give us your opinions on maintenance AI, construction AI, and headquarters AI."

[1137] In this way, the server, terminals, and users work together to execute system processing.

[1138] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1139] Step 1:

[1140] A user accesses the system

[1141] A user accesses the system using a terminal (such as a PC or smartphone). Specifically, the user opens a web browser, enters the system's URL, and logs in. The user's authentication information (user name, password) is required as input, and the system's dashboard is displayed as output.

[1142] Step 2:

[1143] The server runs the generative AI model

[1144] The server detects user access and launches the generative AI model. The user's login information is required as input, and the generative AI model is launched as output. The generative AI model is built using machine learning libraries such as TensorFlow and PyTorch.

[1145] Step 3:

[1146] The server generates AI with a stance for each division.

[1147] The server generates AI with the stance of the maintenance department, construction department, head office department, etc. As input, information obtained from the database of each department is required, and as output, a maintenance AI, a construction AI, a head office AI, etc. For example, a maintenance AI is generated based on past maintenance records.

[1148] Step 4:

[1149] The server sets the topic of discussion

[1150] The server sets the topic of discussion provided by the user. As input, it requires the topic entered by the user (e.g., "What measures are necessary to improve the quality of the network?"), and as output, a prompt sentence is generated for each AI.

[1151] Step 5:

[1152] The server starts the discussion for each AI.

[1153] The server has each AI start a discussion based on the set theme. A prompt is required as input, and the opinion of each AI is generated as output. For example, the maintenance AI may say, "The aging of the current network equipment is a problem, and regular maintenance is necessary."

[1154] Step 6:

[1155] The server uses an emotion engine to analyze the user's emotions.

[1156] The server uses an emotion engine to analyze the user's emotions. The inputs required are the user's voice, facial expressions, text input, etc., and the output is the analyzed user's emotional state. Specifically, the analysis is performed using voice recognition software and image analysis software.

[1157] Step 7:

[1158] The server coordinates the progress of the discussion.

[1159] The server adjusts the progress of the discussion based on the user's emotion information obtained from the emotion engine. The user's emotion information is required as input, and the progress of the discussion is adjusted as output. For example, if the user shows dissatisfaction, the server pauses the discussion and prioritizes the user's opinion.

[1160] Step 8:

[1161] The server aggregates the results of the discussions and generates a report.

[1162] The server aggregates the results of each AI's discussion and generates a report. The input is the results of each AI's discussion, and the output is a PDF report. The report contains detailed descriptions of each AI's opinions and the conclusions of the discussion.

[1163] Step 9:

[1164] User checks the report

[1165] The user uses a terminal to review the generated report. As input, the report downloaded through the system interface is required, and as output, the report content is displayed. The user can also provide feedback on the report content.

[1166] (Application example 1)

[1167] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1168] In conventional discussion systems, it is difficult to efficiently collect opinions from each department and advance the discussion. In addition, because the discussion proceeds without considering the user's feelings, the user's dissatisfaction or anger may hinder the progress of the discussion. Furthermore, there is a problem that the efficiency of the discussion decreases due to a lack of means to check the progress of the discussion in real time and intervene if necessary. In order to solve these issues, a system is needed that uses a generative AI that has the stance of each department to hold the discussion, recognizes the user's feelings, and adjusts the progress of the discussion.

[1169] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1170] In this invention, the server includes a generating AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion through a smartphone or a head-mounted display and intervening as necessary. This makes it possible to efficiently collect the opinions of each department and progress the discussion while taking into account the user's emotions, thereby improving the efficiency and effectiveness of the discussion.

[1171] "Generative AI means" is an artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[1172] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically analyzes the content of discussions, extracts important concerns, and starts discussions.

[1173] "Means of reducing labor costs, accelerating discussions, and preventing items from being overlooked" are techniques for increasing the efficiency of discussions and ensuring that all necessary items are not overlooked.

[1174] An "emotion engine" is a technology that analyzes emotions from a user's voice, facial expressions, text input, etc.

[1175] The "means for adjusting the progress of the discussion based on emotional information" is a technique for appropriately controlling the progress of the discussion depending on the emotional state of the users.

[1176] "Means of checking the progress of the discussion through a smartphone or head-mounted display and intervening if necessary" refers to technology that uses mobile devices or wearable devices to monitor the progress of the discussion in real time and allows users to intervene if necessary.

[1177] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means including an emotion engine that recognizes the user's emotions, a means for adjusting the progress of the discussion based on the emotion information, and a means for checking the progress of the discussion via a smartphone or head-mounted display and intervening if necessary.

[1178] System Programming

[1179] The server implements a generative AI model using the OpenAI API to generate an AI with the stance of each department. This allows discussions to be held from the perspectives of the maintenance department, manufacturing department, quality control department, etc. To recognize the user's emotions, the EmotionRecognizer library is used to capture the user's facial expressions through a webcam. The emotion engine analyzes emotions from the user's voice, facial expressions, text input, etc., and provides the results to the generative AI. The generative AI adjusts the progress of the discussion based on this emotional information.

[1180] Hardware and Software

[1181] Hardware: Smartphone, head-mounted display, webcam

[1182] Software: OpenAI API, OpenCV, EmotionRecognizer library

[1183] Data processing and calculation

[1184] The server uses a generative AI model to generate discussions from each department's perspective. To recognize the user's emotions, a webcam is used to capture facial expressions in real time and the emotions are analyzed with the EmotionRecognizer library. The analyzed emotion information is provided to the generative AI and used to adjust the progress of the discussion. Through a smartphone or head-mounted display, the user can check the progress of the discussion in real time and intervene if necessary.

[1185] Examples

[1186] For example, if you are discussing the introduction of a new manufacturing line, you might use the following prompt:

[1187] "Discuss from the maintenance department's perspective: Discuss concerns regarding the implementation of a new production line."

[1188] "Discuss from the manufacturing perspective: Discuss the concerns regarding the implementation of a new manufacturing line."

[1189] "Discuss from the quality control department's perspective: Discuss concerns regarding the implementation of a new production line."

[1190] This makes it possible to efficiently collect opinions from each department and conduct discussions that take into account the feelings of users, thereby improving the efficiency and effectiveness of discussions.

[1191] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1192] Step 1:

[1193] The server initializes the generative AI model using the OpenAI API. As input, it receives a prompt sentence with the stance of each department and passes it to the generative AI model. The generative AI model generates a discussion from the perspective of each department and returns the discussion content in text format as output.

[1194] Step 2:

[1195] The server captures the user's facial expressions in real time using a webcam. It passes the captured image data as input to the EmotionRecognizer library, which analyzes the image data and outputs the user's emotional information.

[1196] Step 3:

[1197] The server provides the generative AI model with emotion information obtained from the EmotionRecognizer library. As input, the generative AI model passes the emotion information and the content of the discussion to adjust the progress of the discussion. The generative AI model appropriately controls the progress of the discussion based on the emotion information and returns the adjusted content of the discussion as output.

[1198] Step 4:

[1199] The terminal (smartphone or head-mounted display) receives the coordinated discussion content from the server. As input, it receives the discussion content from the server and displays it to the user. The user can check the progress of the discussion in real time and intervene if necessary.

[1200] Step 5:

[1201] Users intervene in the discussion through a smartphone or a head-mounted display. As input, users' opinions and feedback are entered into the device and sent to the server. The server receives the users' opinions and passes them back to the generative AI model to update the discussion content. As output, the updated discussion content is displayed again on the device.

[1202] Step 6:

[1203] The server saves the progress of the discussion as a log. As input, it receives the discussion content and emotion information generated at each step and saves them in a database. This allows the discussion history to be checked later.

[1204] Example 2

[1205] Next, a second embodiment of the second form will be described. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal".

[1206] In conventional discussion systems, human intervention was required to hold discussions that reflected the stance of each department, which not only took time but also often slowed down the progress of discussions. In addition, manual work was required to save the results of discussions, which led to problems with personalizing knowledge and handover. Furthermore, since discussions did not take into account the feelings of users, user satisfaction could decline.

[1207] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1208] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stances of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing man-hours, accelerating discussions, and preventing oversights, an emotion engine means for adjusting the results of discussions according to the emotions of users, and a means for saving the results of discussions as meeting minutes. This makes it possible to streamline discussions, prevent knowledge from becoming personal, and hold discussions that take into account the emotions of users.

[1209] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[1210] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically initiates discussions and identifies key issues.

[1211] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[1212] "Emotion engine means" refers to technology for adjusting the outcome of a discussion according to the user's emotions.

[1213] "Means of saving the results of the discussion as meeting minutes" refers to technology for automatically recording and saving the content and results of the discussion.

[1214] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[1215] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[1216] "Headquarters AI" refers to artificial intelligence that understands the overall policies and strategies of the company and holds discussions based on them.

[1217] This invention is a system that includes a generation AI means for holding discussions between roles that are assigned roles that represent the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, an emotion engine means for adjusting the results of the discussion according to the user's emotions, and a means for saving the results of the discussion as meeting minutes.

[1218] Generating a Program

[1219] The server generates the system's programs, which include the maintenance AI, construction AI, headquarters AI, and emotion engine. These programs allow each AI to hold discussions based on their respective expertise and propose optimal solutions.

[1220] Program processing explanation

[1221] The server processes the data in the following steps:

[1222] 1. Data collection:

[1223] The server collects data on network maintenance status, construction plans, and company policies and strategies. This data serves as the basis for each AI's discussion. Specifically, the server obtains data from network monitoring tools (e.g., Nagios) and project management software (e.g., JIRA).

[1224] 2. Start the discussion:

[1225] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson).

[1226] 3. Discussion process:

[1227] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on maintaining the network, while the construction AI proposes a plan to build a new network. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center."

[1228] 4. Save the results of the discussion:

[1229] The server saves the results of the discussion as minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL).

[1230] 5. Emotion engine intervention:

[1231] When a user expresses dissatisfaction, the server activates an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, if a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality.

[1232] 6. Reporting results:

[1233] The server reports the final discussion result to the user. Specifically, the result is displayed on the user's dashboard (e.g., a web application). For example, the report may be in the form of "We will implement the following measures to improve the quality of the network."

[1234] Examples of concrete examples and prompts

[1235] Examples:

[1236] If a user is dissatisfied with the quality of the network, the following prompt sentence is fed into the generative AI model:

[1237] Example prompt:

[1238] "Users are complaining about the quality of their network. Please prioritize discussions on how to improve it and report the results."

[1239] Based on this prompt, the generation AI holds a discussion about improving quality, with the maintenance AI proposing, "We should introduce a new monitoring tool to strengthen network monitoring," and the construction AI proposing, "We should improve network throughput by introducing a new router." The server displays the results of the discussion on the user's dashboard, reporting, "We will introduce a new monitoring tool and router to improve network quality."

[1240] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1241] Step 1:

[1242] The server collects data on network maintenance status, construction plans, and company policies and strategies. Specifically, it obtains data from network monitoring tools (e.g. Nagios) and project management software (e.g. JIRA). As input, it receives data from these tools, and as output, it generates the basic information for each AI to discuss.

[1243] Step 2:

[1244] The server instructs the maintenance AI, construction AI, and headquarters AI to start a discussion based on the collected data. Each AI prepares to exchange opinions using a dedicated discussion engine (e.g. IBM Watson). As input, it provides each AI with the collected data, and as output, it performs initial settings for each AI to start a discussion.

[1245] Step 3:

[1246] The server manages the process in which each AI exchanges opinions. The maintenance AI makes suggestions on network maintenance, while the construction AI proposes a new network construction plan. The headquarters AI adjusts the final solution, taking into account the company's overall policies and strategies. Specifically, the maintenance AI suggests that "a specific patch should be applied to fix the vulnerabilities of the current network," while the construction AI suggests that "network redundancy should be increased by building a new data center." As input, it receives suggestions from each AI, and as output, it generates an optimal solution.

[1247] Step 4:

[1248] The server saves the results of the discussion as the minutes of the meeting. Specifically, it saves the contents of the discussion in text format in a database (e.g., MySQL). It receives the results of the discussion as input, and generates minutes as output and saves them in the database.

[1249] Step 5:

[1250] When a user expresses dissatisfaction, the server launches an emotion engine. The emotion engine analyzes the user's emotion data (e.g., obtained from facial recognition software or feedback forms) and adjusts the outcome of the discussion. Specifically, when a user expresses dissatisfaction with "poor network quality," the emotion engine instructs the generation AI to prioritize discussions on improving quality. It receives the user's emotion data as input and generates instructions to adjust the outcome of the discussion as output.

[1251] Step 6:

[1252] The server reports the final discussion result to the user. Specifically, it displays the result on the user's dashboard (e.g., a web application). For example, it reports in the form of "We will implement the following measures to improve the quality of the network." It receives the final discussion result as input and generates information to report to the user as output.

[1253] (Application example 2)

[1254] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server", and the headset type terminal 314 will be referred to as a "terminal".

[1255] In conventional network security systems, it is difficult to integrate the expertise of each department to derive optimal solutions, and there is a lack of consideration for user feelings, making it difficult to improve user satisfaction. Furthermore, efficient discussion and coordination is required in detecting security risks, proposing countermeasures, and designing and implementing new security infrastructure.

[1256] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1257] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, an emotion engine means for analyzing user emotions and adjusting the results of discussions, a security preservation AI means for detecting security risks and proposing measures, a security construction AI means for proposing the design and implementation of new security infrastructure, and a headquarters AI means for understanding the security policy and strategy of the entire company and adjusting the overall security strategy. This makes it possible to integrate the expertise of each department and propose optimal security measures that take into account the emotions of users.

[1258] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[1259] An "emotion engine" is an engine that has the function of analyzing the user's emotions and adjusting the results of the discussion according to those emotions.

[1260] "Security preservation AI" is an artificial intelligence that has knowledge and experience in network preservation, detects security risks, and proposes countermeasures.

[1261] "Security Construction AI" is an artificial intelligence that has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods.

[1262] "Headquarters AI" is an artificial intelligence that understands the security policies and strategies of the entire company and adjusts the overall security strategy.

[1263] A "role" is a virtual role with expertise and experience that is assigned to each department's stance.

[1264] "Introducing a discussion" is the process of identifying concerns and starting a discussion without human intervention.

[1265] "Reducing labor hours" refers to measures aimed at reducing the time and effort required for work.

[1266] "Accelerating discussions" is a measure aimed at speeding up the progress of discussions.

[1267] "Preventing omissions in consideration" is a measure to ensure that important matters are not overlooked during discussion or consideration.

[1268] A system for implementing this invention is configured as follows: The server includes a generation AI that holds discussions between roles that are assigned the stance of each department, an emotion engine that analyzes user emotions and adjusts the results of the discussion, a security preservation AI that detects security risks and proposes countermeasures, a security construction AI that proposes the design and implementation of new security infrastructure, and a headquarters AI that understands the security policy and strategy of the entire company and adjusts the overall security strategy.

[1269] Hardware and Software Configuration

[1270] The server uses the following hardware and software:

[1271] Hardware: High-performance server (e.g. Dell PowerEdge series)

[1272] Software: Python, sentiment analysis libraries (e.g. TextBlob), generative AI models (e.g. GPT-3)

[1273] Data processing and calculation

[1274] The server performs the following data processing and calculations.

[1275] 1. Generative AI: Each department is assigned a role based on their stance, and the roles are discussed to propose optimal solutions. A generative AI model (e.g. GPT-3) is used to generate the content of the discussion.

[1276] 2. Sentiment Engine: Analyze user sentiment and adjust the discussion outcome depending on the sentiment. Use a sentiment analysis library (e.g. TextBlob) to parse user feedback.

[1277] 3. Security Preservation AI: With knowledge and experience in network preservation, it detects security risks and proposes countermeasures. It uses generative AI models to detect risks and propose countermeasures.

[1278] 4. Security Construction AI: Has knowledge about the design and implementation of new security infrastructure and proposes optimal security construction methods. Uses generative AI models to make design and implementation suggestions.

[1279] 5. HQ AI: Understand the security policies and strategies of the entire enterprise and adjust the overall security strategy. Use generative AI models to make adjustments to the strategy.

[1280] Examples

[1281] For example, if a user provides feedback like "I'm dissatisfied with the current security measures," the emotion engine will analyze this feedback and recognize that the user is expressing dissatisfaction. Based on this, the security preservation AI will suggest updating firewall rules.

[1282] Examples of prompt statements

[1283] User feedback: "I'm not satisfied with the current security measures."

[1284] AI discussion result: "Security Preservation AI proposes updating firewall rules."

[1285] In this way, it becomes possible to integrate the expertise of each department and propose optimal security measures that take into account the user's feelings.

[1286] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1287] Step 1:

[1288] The server receives feedback from the user. The input is the user's feedback text, and the output is the storage of the feedback text. For example, if the user enters "I am not satisfied with the current security measures," the server stores this text in a database.

[1289] Step 2:

[1290] The server parses the user feedback using an emotion engine. The input is the stored feedback text and the output is the user's emotional state. Specifically, it uses a sentiment analysis library (e.g., TextBlob) to extract the emotion "unhappy" from the feedback text.

[1291] Step 3:

[1292] The server sends a prompt to the generation AI based on the analysis results of the emotion engine. The input is the user's emotional state and feedback text, and the output is a prompt sentence for the generation AI. Specifically, it generates the prompt sentence "User feedback: 'I am dissatisfied with the current security measures.'"

[1293] Step 4:

[1294] The server uses a generative AI to hold a discussion and generate an optimal solution. The input is the prompt sentence, and the output is the result of the discussion. Specifically, a generative AI model (e.g., GPT-3) is used to generate the discussion result "Security preservation AI proposes updating firewall rules."

[1295] Step 5:

[1296] The server reports the generated discussion results to the user. The input is the discussion result, and the output is a report message to the user. Specifically, the server sends the message "Security preservation AI proposes updating firewall rules" to the user.

[1297] Step 6:

[1298] The server stores the results of the discussion in a database. The input is the discussion result and the output is the stored discussion result. Specifically, the generated discussion result is stored in the database for future reference and analysis.

[1299] In this way, the server performs a series of processes: receiving user feedback, analyzing emotions, using generative AI to propose optimal solutions, and reporting the results to the user.

[1300] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input for the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1301] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ) and other generation AIs. The data generation model 58 is obtained by making a neural network perform deep learning. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating a voice, text data indicating a text, and image data indicating an image is input. The data generation model 58 performs inference on the input inference data according to the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1302] Another example of generative AI is Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) are mentioned.

[1303] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1304] [Fourth embodiment]

[1305] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1306] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1307] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a wide area network (WAN) and / or a local area network (LAN).

[1308] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. In addition, the microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1309] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs the voice according to instructions from the processor 46.

[1310] Camera 42 is a small digital camera equipped with an optical system including a lens, an aperture, and a shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (e.g., an imaging range defined by an angle of view equivalent to the width of the field of vision of an average healthy person).

[1311] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for transmitting and receiving various types of information between the processor 46 and the processor 28 via the network 54. The transmission and reception of various types of information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is performed in a secure state.

[1312] The control target 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, legs, etc. The posture and behavior of the robot 414 are controlled by controlling the motors of the arms, hands, legs, etc. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1313] Fig. 8 shows an example of main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1314] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32, and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1315] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1316] In the robot 414, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50, and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1317] Next, the specific processing by the specific processing unit 290 of the data processing device 12 will be described.

[1318] "Example 1"

[1319] The present invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[1320] "Example 2"

[1321] As a concrete example, the maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies. These AIs hold discussions based on their respective stances and propose optimal solutions. This discussion is conducted without human intervention, and the results of the discussion are saved as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[1322] The process flow of each embodiment will be described below.

[1323] "Example 1"

[1324] Step 1: The system generates AIs with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.).

[1325] Step 2: The generated AIs begin to discuss with each other. The content of the discussions will be related to improving network quality and achieving the Ministry of Internal Affairs and Communications' objectives.

[1326] Step 3: The results of the discussion are saved as meeting minutes and can be referenced later.

[1327] Step 1: The maintenance AI has knowledge and experience in network maintenance, and the construction AI has knowledge and experience in network construction. The headquarters AI understands the company's overall policies and strategies.

[1328] Step 2: These AIs debate based on their respective stances and propose the optimal solution.

[1329] Step 3: The discussion is conducted without human intervention, and the results are stored as meeting minutes. This eliminates the need for personalization or handover of knowledge.

[1330] Example 1

[1331] Next, a description will be given of Example 1 of the embodiment 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1332] In the conventional discussion system, it was difficult to hold discussions that reflected the stance of each department, and there was a problem that it took a lot of time to identify concerns and start discussions. In addition, the human intervention hindered the acceleration of discussions and increased the risk of overlooking something. Furthermore, the necessity of personalizing knowledge and handover made it difficult to hold efficient discussions.

[1333] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1334] In this invention, the server includes a generation AI means for discussion between roles that are given the stance of each department, a means for a user to input a prompt text, a means for the server to receive the prompt text and select an appropriate generation AI model, a means for the server to send the prompt text to the generation AI model selected by the server, a means for the generation AI model to start a discussion and for each AI to give an opinion from its own perspective, a means for the server to collect and integrate the opinions of each AI, a means for the server to return the integrated discussion result to the user, a means for identifying concerns and introducing discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This enables efficient and comprehensive discussions.

[1335] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[1336] A "prompt sentence" refers to an instruction sentence regarding the topic or subject of discussion that is entered by the user.

[1337] "Server" refers to a computer system for receiving prompts, selecting an appropriate generative AI model, and managing the discussion.

[1338] A "generative AI model" refers to an artificial intelligence model with a specific role or perspective, such as maintenance AI, construction AI, or headquarters AI.

[1339] "Conservation AI" refers to a generative AI model that has a maintenance department stance.

[1340] "Construction AI" refers to a generative AI model that has a construction sector stance.

[1341] "Headquarters AI" refers to a generative AI model that has the stance of the headquarters department.

[1342] "Discussion results" refers to the conclusions and proposals reached through discussions between generative AI models.

[1343] "Reducing labor hours" refers to reducing the time and effort required for work.

[1344] "Accelerating the discussion" refers to speeding up the discussion.

[1345] "Preventing gaps in consideration" refers to making sure that important matters are not overlooked in the discussion.

[1346] "Personalization of knowledge" refers to the concentration of knowledge and information in the hands of specific individuals.

[1347] "Handover" refers to the transfer of work or knowledge to the next person in charge.

[1348] This invention is a system that implements a generation AI in which roles that are given the stance of each department discuss with each other. Specifically, AI with the stances of the maintenance department, construction department, head office department, etc. (hereinafter referred to as maintenance AI, construction AI, head office AI, etc.) is generated, and these AIs discuss with each other. The content of the discussion is related to improving the quality of the network and achieving the Ministry of Internal Affairs and Communications' obliquity, and concerns are identified and discussions are initiated without human intervention. This can reduce labor hours, accelerate discussions, and prevent oversights.

[1349] Hardware and Software Used

[1350] The server uses a high-performance server (e.g., a server equipped with an NVIDIA GPU). The software uses a generative AI model (e.g., GPT-4) and a discussion management system (e.g., a dedicated discussion framework).

[1351] Program Processing

[1352] The server receives a prompt from the user and selects an appropriate generative AI model. It then sends the prompt to the selected generative AI model, starting a discussion. The generative AI model starts a discussion with each AI, such as the maintenance AI, construction AI, and headquarters AI, and gives opinions from their own perspectives. The server collects and integrates the opinions of each AI. Finally, it returns the integrated discussion results to the user.

[1353] Examples

[1354] Using the terminal, the user inputs a prompt sentence related to the topic or agenda of the discussion. For example, the user inputs "Please start a discussion on improving network quality." The server receives this prompt sentence and analyzes it using a natural language processing algorithm. Based on the analysis results, the server selects a maintenance AI, a construction AI, or a headquarters AI. The server sends the prompt sentence to the selected generative AI model via API. The generative AI model receives the prompt sentence and begins preparing for the discussion.

[1355] The generative AI model initiates a discussion between the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to prioritize due to budget constraints." The server collects the opinions sent by each AI and integrates them using a natural language processing algorithm. This results in a comprehensive discussion result.

[1356] The server returns the integrated discussion results to the user, who can check the results on the screen of his / her device. For example, the conclusion "Strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed.

[1357] In this way, the server enables efficient and comprehensive discussions, reducing labor costs, accelerating discussions, and preventing items from being overlooked.

[1358] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1359] Step 1:

[1360] The user inputs a prompt text. The user inputs "Please start a discussion on improving the quality of the network" in the input field of the terminal and clicks the send button. The input prompt text is sent to the server.

[1361] Step 2:

[1362] The server receives the prompt sentence and selects an appropriate generative AI model. The server receives the prompt sentence and analyzes the prompt sentence using a natural language processing algorithm. Based on the analysis result, the maintenance AI, construction AI, or head office AI is selected. The input is the prompt sentence, and the output is the selected generative AI model.

[1363] Step 3:

[1364] The server sends a prompt to the selected generative AI model. The server sends the prompt to the selected generative AI model via API. The generative AI model receives the prompt and starts preparing for discussion. The input is the selected generative AI model and the prompt, and the output is a state ready for discussion.

[1365] Step 4:

[1366] The generative AI model starts a discussion, and each AI gives an opinion from its own perspective. The generative AI model starts a discussion with the maintenance AI, construction AI, and headquarters AI. The maintenance AI points out that "maintenance and inspection of the current network is insufficient," and the construction AI suggests that "new infrastructure needs to be introduced." The headquarters AI states that "we need to set priorities due to budget constraints." The input is the prompt text and the role of each AI, and the output is the opinion of each AI.

[1367] Step 5:

[1368] The server collects and integrates the opinions of each AI. The server collects the opinions sent by each AI and integrates the opinions using a natural language processing algorithm. This results in a comprehensive discussion result. The input is the opinion of each AI, and the output is the integrated discussion result.

[1369] Step 6:

[1370] The server returns the integrated discussion results to the user. The server returns the integrated discussion results to the user. The user can check the results of the discussion on their terminal screen. For example, the conclusion "strengthen network maintenance and inspection, and gradually introduce new infrastructure" is displayed. The input is the integrated discussion results, and the output is the discussion results that are displayed to the user.

[1371] (Application example 1)

[1372] Next, a description will be given of Application Example 1 of Form Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1373] Conventional discussion systems had problems in that it was difficult to gather the opinions of each department and proceed with discussions efficiently, which required a lot of man-hours and made it easy for considerations to be overlooked. Also, in discussions on how to improve production efficiency within factories, there was a lack of a way to generate AI with the stance of each department and have those AIs discuss with each other. Furthermore, the implementation as an application to be installed on smartphones was also insufficient.

[1374] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means. In this invention, the server includes a generating AI means for discussion between roles that are given the stance of each department as a role, a means for identifying concerns and initiating discussions without human intervention, and a means for reducing labor hours, accelerating discussions, and preventing oversights. This makes it possible to generate AI with the stance of each department in a factory, and for these AIs to discuss ways to improve production efficiency. In addition, by implementing it as an application to be installed on a smartphone, users can easily use it.

[1375] A "generative AI means" is an artificial intelligence means that is generated to hold discussions between roles that are assigned the stance of each department.

[1376] "Means for identifying concerns and initiating discussions without human intervention" refers to a means for automatically extracting concerns or issues that arise during the discussion process and starting a discussion.

[1377] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" are measures to improve the efficiency of discussions and cover all necessary items without missing anything.

[1378] "A means of generating AI that has the perspective of each department within a factory and having these AIs discuss with each other how to improve production efficiency" refers to a means of generating artificial intelligence that has the perspective of each department within a factory (production, quality control, maintenance, etc.) and having these AIs discuss with each other how to improve production efficiency.

[1379] "Means including an application installed on a smartphone" refers to means implemented as an application installed on a smartphone.

[1380] "Roles of maintenance AI, construction AI, headquarters AI, production AI, quality control AI, maintenance AI, etc." are the roles of artificial intelligence that have the responsibilities and perspectives of each department.

[1381] As an embodiment of the present invention, we will create an AI that has the stance of each department in a factory, and build a system in which these AIs discuss ways to improve production efficiency. The specific embodiment will be described below.

[1382] System Configuration

[1383] The system is composed of a server, a smartphone, a generative AI model, and a prompt. The server provides the generative AI means for discussions between roles that are assigned the stances of each department. The smartphone is a device on which users can install applications that allow them to easily use the system.

[1384] Hardware and Software Used

[1385] Hardware: Servers, smartphones

[1386] Software: Python, OpenAI API

[1387] Data processing and calculation

[1388] The server processes and calculates data in the following procedure.

[1389] 1. Using generative AI models:

[1390] The server uses the OpenAI API to generate an AI with the stance of each department. Specifically, it inputs a prompt sentence and generates an AI with the stance of each department.

[1391] 2. Conducting the discussion:

[1392] The server holds discussions between the generated AIs of each department, creates discussion prompts, and uses the OpenAI API to obtain the results of the discussion.

[1393] 3. Display the results:

[1394] The results of the discussion are displayed to the user through an application installed on their smartphone.

[1395] Examples

[1396] Below are some examples of prompt sentences.

[1397] Generate an AI with a production sector stance.

[1398] Generate an AI with a quality control department stance.

[1399] Generate an AI with a maintenance department stance.

[1400] Please discuss with the AIs below how to improve production efficiency.

[1401] By using this prompt, it is possible to generate AI with the stance of each department and hold discussions. For example, the production department's AI can propose "what kind of data should be collected to identify bottlenecks in the production line," while the quality control department's AI can discuss "what kind of inspection method should be introduced as a measure to improve quality control."

[1402] In this way, a system is realized that enables efficient discussions on how to improve production efficiency within a factory.

[1403] The flow of the specific process in the application example 1 will be described with reference to FIG.

[1404] Step 1:

[1405] The server creates a prompt sentence to generate an AI with the stance of each department. Specifically, it generates a prompt sentence like the following based on the department specified by the user (e.g. production, quality control, maintenance): "Please generate an AI with the stance of the production department." This prompt sentence is the input, and the generated AI is the output.

[1406] Step 2:

[1407] The server sends the generated prompt sentence to OpenAI's API to generate an AI with the stance of each department. Specifically, it calls OpenAI's API and provides the prompt sentence as input. The API uses the generative AI model to generate an AI with the stance of each department and returns the result to the server. This result becomes the output.

[1408] Step 3:

[1409] The server stores the generated AI for each department in a list. Specifically, the AI ​​returned from the API is saved in list format and used in the subsequent discussion process. This list is the input, and the stored AI list is the output.

[1410] Step 4:

[1411] The server creates a prompt for the AIs in each department to discuss with each other. Specifically, it generates the following prompt based on the AIs stored in the list: "Please hold a discussion between the AIs below to improve production efficiency." This prompt is the input, and the results of the discussion are the output.

[1412] Step 5:

[1413] The server sends the discussion prompt to the OpenAI API and obtains the discussion result. Specifically, it calls the API and provides the discussion prompt as input. The API uses a generative AI model to generate the discussion result and returns the result to the server. This result becomes the output.

[1414] Step 6:

[1415] The server transmits the acquired results of the discussion to an application installed on the smartphone. Specifically, it converts the results of the discussion into a format that the user can view and transmits them to the smartphone. The results of this discussion are the input, and data in a format that the user can view is the output.

[1416] Step 7:

[1417] The user checks the results of the discussion displayed on the smartphone and makes additional discussions or adjustments as necessary. Specifically, the user views the results of the discussion through the smartphone application and decides on the next action. The results of this discussion are the input, and the user's actions are the output.

[1418] Example 2

[1419] Next, a second embodiment of the second embodiment will be described. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1420] In conventional discussion systems, it was necessary for people with specialized knowledge from each department to directly participate in the discussions, which led to problems with personalizing knowledge and handover. In addition, it took a lot of time to progress the discussions and record the results, making it difficult to accelerate the discussions and prevent oversights. This made it difficult to hold efficient discussions and propose optimal solutions.

[1421] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1422] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for a user to input the topic of discussion, a means for the server to start a discussion and collect opinions from each AI model, and a means for automatically generating the results of the discussion as minutes and saving them in a database. This makes it possible to have efficient discussions and propose optimal solutions without the need for personalizing or handing over knowledge.

[1423] "Generative AI means" refers to artificial intelligence technology that allows discussions to be held between roles that are assigned based on the stance of each department.

[1424] "Means of identifying concerns and initiating discussions without human intervention" refers to technology that automatically extracts discussion topics and concerns and starts discussions.

[1425] "Measures to reduce labor costs, accelerate discussions, and prevent items from being overlooked" refers to techniques for making discussions more efficient and covering all necessary items without missing anything.

[1426] "Means for users to input a topic for discussion" refers to an interface that allows users to input a topic for discussion into the system.

[1427] "Means for the server to initiate discussions and aggregate opinions from each AI model" refers to the technology that enables the server to initiate discussions and collect and integrate opinions from each AI model.

[1428] "Means for automatically generating the results of discussions as minutes and storing them in a database" refers to technology for automatically recording the content and results of discussions in text format and storing them in a database.

[1429] "Maintenance AI" refers to artificial intelligence that has knowledge and experience regarding network maintenance.

[1430] "Construction AI" refers to artificial intelligence that has knowledge and experience regarding network construction.

[1431] "Headquarters AI" refers to artificial intelligence that understands the company's overall policies and strategies.

[1432] "Database" refers to an information management system for storing the results of discussions and minutes.

[1433] This invention is a system that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions. Specifically, three AI models, a maintenance AI, a construction AI, and a headquarters AI, hold discussions using their respective expertise, and the results are automatically saved as minutes.

[1434] Hardware and Software Used

[1435] Hardware: Servers, terminals

[1436] Software: Generative AI models (e.g. GPT-4), database management systems (e.g. MySQL)

[1437] Generating a Program

[1438] The server generates programs including the maintenance AI, construction AI, and headquarters AI models. The programs are implemented using programming languages ​​such as Python and Java. Each AI model is based on a generative AI model (e.g., GPT-4) and is trained to have its own specialized knowledge.

[1439] Program Processing

[1440] The user inputs the topic of discussion

[1441] The user inputs the topic of discussion into the generative AI model using the device. For example, if the user wants to input the topic "optimal strategy for designing new network infrastructure," the user writes the topic in the input form on the device and clicks the send button.

[1442] The server starts the discussion.

[1443] The server receives the topic entered by the user and has the maintenance AI, construction AI, and headquarters AI start a discussion. Specifically, the server generates a prompt for each AI model and sends an API request to hold a discussion. For example, the server generates a prompt like the one below.

[1444] Example prompt:

[1445] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[1446] The server aggregates the results of the discussion.

[1447] The server aggregates the opinions returned by each AI model. Specifically, it analyzes the responses from each AI model and applies an algorithm to propose the optimal solution. For example, if the maintenance AI suggests that "a backup system should be installed to ensure network redundancy" and the construction AI suggests that "fiber optic cables should be used for cost-effectiveness," the server integrates these opinions to derive the optimal solution of "using fiber optic cables and installing a backup system."

[1448] The server generates the minutes.

[1449] The server automatically generates the results of the discussion as minutes. Specifically, it summarizes the contents of the discussion and the final solution in text format and saves them in a database management system (e.g., MySQL). Users can later refer to these minutes. For example, when a user clicks the "Show minutes" button on a terminal, the server retrieves the minutes from the database and displays them on the terminal.

[1450] Examples

[1451] For example, if a user inputs the theme "optimal strategy for designing new network infrastructure," the following prompt sentence is input to the generative AI model.

[1452] Example prompt:

[1453] Discuss the optimal strategy for designing a new network infrastructure. The maintenance AI should provide an opinion based on their knowledge and experience in network maintenance, the construction AI should provide an opinion based on their knowledge and experience in network construction, and the head office AI should provide an opinion based on their understanding of the company's overall policies and strategies.

[1454] Based on this prompt, the server will have each AI start a discussion and propose the optimal solution. The results of the discussion are automatically saved as minutes, so the user can check the contents later.

[1455] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1456] Step 1:

[1457] Generating a Program

[1458] The server generates a program including each AI model: maintenance AI, construction AI, and headquarters AI. Specifically, using programming languages ​​such as Python and Java, each AI model is trained based on a generative AI model (e.g., GPT-4). Training data for each AI model is required as input, and each AI model is generated as output.

[1459] Step 2:

[1460] Enter the discussion topic

[1461] The user uses a terminal to input the topic of discussion into the generative AI model. Specifically, the user enters the topic in the input form on the terminal and clicks the send button. The input is the topic of discussion entered by the user, and the output is the topic sent to the server.

[1462] Step 3:

[1463] Starting the discussion

[1464] The server receives the topic input by the user and starts a discussion with the maintenance AI, construction AI, and headquarters AI. Specifically, the server generates a prompt for each AI model and sends an API request to hold the discussion. The input is the discussion topic received from the user, and the output is the generated prompt to be sent to each AI model.

[1465] Step 4:

[1466] Gathering opinions

[1467] The server collects the opinions returned from each AI model. Specifically, it analyzes the response from each AI model and obtains the opinions in text format. The input is the response from each AI model, and the output is the collected opinions of each AI model in text format.

[1468] Step 5:

[1469] Summarizing the results of discussions

[1470] The server proposes the optimal solution based on the collected opinions. Specifically, it integrates the opinions from each AI model and applies an algorithm to derive the optimal solution. The inputs are the opinions of each AI model, and the output is the proposed optimal solution.

[1471] Step 6:

[1472] Generate minutes

[1473] The server automatically generates the results of the discussion as minutes and saves them in a database. Specifically, the contents of the discussion and the final solution are summarized in text format and saved in a database management system (e.g., MySQL). The input is the results of the discussion, and the output is the minutes that are generated and saved in the database.

[1474] Step 7:

[1475] View minutes

[1476] A user refers to the minutes from a terminal. Specifically, when the user clicks the "Show minutes" button, the server retrieves the minutes from the database and displays them on the terminal. The input is a user request, and the output is the minutes displayed on the terminal.

[1477] (Application example 2)

[1478] Next, a description will be given of Application Example 2 of Form Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal".

[1479] In the conventional system, it took a lot of time and effort for specialists from each department to hold discussions and come up with optimal solutions. It was also time-consuming to accurately record the results of discussions and create meeting minutes for future reference. Furthermore, knowledge was easily tampered with and problems with handover occurred, making efficient operation difficult.

[1480] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1481] In this invention, the server includes a generation AI means for discussion between roles that are assigned the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing oversights, a means for automatically saving the results of the discussion as minutes, and a means for AIs with multiple specialized knowledge to cooperate and propose optimal solutions. This makes it possible to efficiently and quickly derive optimal solutions, automatically create minutes, and prevent knowledge from being personalized.

[1482] "Generative AI" is an artificial intelligence that holds discussions among roles that are assigned based on the stance of each department, and proposes optimal solutions.

[1483] "Issues of concern" refer to issues, challenges, risks, etc. that are subject to discussion.

[1484] "Introducing the discussion" is the process of preparing and setting up the discussion.

[1485] "Reducing labor hours" refers to reducing the time and effort required for work.

[1486] "Accelerating discussions" refers to moving discussions forward more quickly.

[1487] "Preventing oversights" refers to making sure that important matters are not overlooked during the discussion or deliberation process.

[1488] A "minutes" is a document that records the content and results of a discussion.

[1489] "Specialist knowledge" refers to advanced knowledge or skills in a particular field.

[1490] "Personalization of knowledge" refers to a state in which knowledge or skills become dependent on specific individuals.

[1491] "Handover" is the process of transferring work and knowledge to the next person in charge.

[1492] A system for implementing this invention includes a generation AI means for holding discussions between roles that are assigned roles based on the stance of each department, a means for identifying concerns and initiating discussions without human intervention, a means for reducing labor hours, accelerating discussions, and preventing items from being overlooked, a means for automatically saving the results of the discussion as minutes, and a means for multiple AIs with specialized knowledge to work together to propose optimal solutions.

[1493] Hardware and Software Used

[1494] Hardware: Smartphone

[1495] Software: Python, JSON

[1496] Data processing and calculation

[1497] The server uses a generation AI that is assigned a role based on the stance of each department to hold discussions. Specifically, AIs with specialized knowledge, such as maintenance AI, construction AI, and head office AI, work together to identify concerns and introduce discussions. The results of the discussion are automatically saved as minutes and can be viewed on smartphones.

[1498] Examples

[1499] For example, if a factory manager is considering installing a new production line, he or she could begin the discussion with the following prompt:

[1500] Examples of prompt statements

[1501] issue = "Introduction of new production line"

[1502] solutions = manager.conduct_meeting(issue)

[1503] print(solutions)

[1504] By executing this prompt, the maintenance AI will propose "solutions based on network maintenance," and the construction AI will propose "solutions based on network construction." This allows factory managers to check optimal production plans and maintenance schedules on their smartphones.

[1505] This system makes it possible to efficiently and quickly arrive at optimal solutions, automatically create meeting minutes, and prevent knowledge from becoming personalized.

[1506] The flow of the specific process in the application example 2 will be described with reference to FIG.

[1507] Step 1:

[1508] The server initializes the generation AI, which is given the stance of each department as a role. Specifically, it creates instances of maintenance AI, construction AI, headquarters AI, etc., and sets their respective expertise. The name and expertise of each AI are required as input, and the initialized AI instance is obtained as output.

[1509] Step 2:

[1510] The user inputs the concern to be discussed, for example, setting the agenda "Installation of a new production line." As input, the concern to be discussed is required, and as output, the agenda is obtained.

[1511] Step 3:

[1512] The server provides each AI with an agenda and asks them to propose a solution. Each AI generates a solution based on its own expertise. The inputs are the agenda and the expertise of each AI, and the output is the solution proposed by each AI.

[1513] Step 4:

[1514] The server collects the solutions obtained from each AI and selects the optimal solution. As input, it needs the solution of each AI, and as output, it gets the o...

Claims

1. A generative AI model for discussion among roles that are given the stance of each department, A means for a user to input a prompt sentence instructing the generative AI model to provide an opinion from each viewpoint; A server receives the prompt sentence and selects an appropriate generative AI model; means for transmitting the prompt sentence to each of the generative AI models selected by the server; A means for each of the generative AI models to initiate a discussion and provide opinions from each of the generative AI models' respective perspectives; A means for the server to collect opinions of each of the generative AI models and integrate them as a discussion result; A means for returning the integrated discussion results to the user; A system including:

2. The system of claim 1 , wherein the generative AI model has the roles of maintenance AI, construction AI, and headquarters AI.

3. The system according to claim 1 , wherein the system does not require personalization and handover of knowledge.

Images