System

Abstract

A system is provided.SOLUTION: A system comprising: means for loading a plurality of trained models from a database or storage; means for adjusting parameters of the plurality of trained models via a graphical user interface; means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output; means for displaying the intermediate output in real time; and means for generating and storing a final model.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 an instruction sentence related to a description of 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] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] While the diversity and range of applications of AI learning models are expanding, there is a lack of tools and platforms for effectively combining multiple learning models to generate unique models. Furthermore, while there is an increasing need to combine the features of different learning models, this process generally requires specialized knowledge and is often time-consuming. Therefore, there is a need for an environment that allows users with diverse knowledge and skills to easily combine models and generate AI tailored to their own purposes. [Means for solving the problem]

[0005] The present invention provides a system for efficiently and intuitively combining multiple trained AI models, which includes the following means:

[0006] A means of loading multiple trained models from a database or storage.

[0007] A means to adjust the parameters of multiple trained models via a graphical user interface.

[0008] A means of combining multiple trained models based on adjusted parameters to generate intermediate outputs.

[0009] A means of displaying intermediate output in real time.

[0010] A means to generate and save the final model.

[0011] This allows users to create their own unique AI model by adjusting and combining the features of different models through an intuitive interface similar to a music equalizer, while checking the output results.

[0012] A "trained model" is an AI model that has been pre-trained using a specific dataset and has learned the patterns and rules for performing a specific task.

[0013] A "database or storage" is a system or device for storing and managing digital data, and is used to store AI models and other data.

[0014] A "graphical user interface" is an interface that allows a user to perform operations visually, and allows settings to be adjusted using visual elements such as sliders and dials.

[0015] "Parameters" are settings or variables that control the behavior and performance of an AI model, and the model's output changes when adjusted by the user.

[0016] "Intermediate output" refers to the intermediate output generated as a result of combining and adjusting multiple trained models, and is intermediate data until the final output result is obtained.

[0017] "Real-time" means that processing and reactions are immediate, and the results are reflected immediately in response to user operations, with little delay in the feel of operation.

[0018] The "final model" is the final AI model generated based on the combination of adjusted parameters and multiple trained models, and is a training model optimized to meet user requirements.

[0019] "Saving" refers to the act of storing the generated final model and related data in a database or storage so that it can be reused later. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 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. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 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. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 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. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0021] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0022] First, the terms used in the following description will be explained.

[0023] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single 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), and an APU (Accelerated Processing Unit).

[0024] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0025] 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.

[0026] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0027] 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. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0028] [First embodiment]

[0029] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0030] 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.

[0031] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0032] 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.

[0033] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. 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.

[0034] 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 (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0035] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0036] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0037] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. 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 process 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.

[0038] 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.

[0039] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. 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.

[0040] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models. This system consists of a server, a terminal, and a user, and allows the user to adjust the parameters of each model using a graphical user interface and generate the final model while checking the results in real time.

[0042] The server is responsible for loading multiple trained models selected by the user from a database or storage, and then sending the loaded model data to the device, making it operable on the device.

[0043] The device analyzes the model data received from the server and displays an intuitive graphical interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model.

[0044] The user adjusts the parameters of each model using the interface on the device. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0045] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. First, the user selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0046] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output result to the user, who can then confirm it.

[0047] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device sends the request to the server, which generates the final combination result and saves it in a database or storage. The final model can then be downloaded by the user or integrated into other systems for use.

[0048] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0049] The processing flow will be explained below.

[0050] Step 1:

[0051] A user logs in to the system and accesses a screen that displays a list of trained models.

[0052] Step 2:

[0053] The terminal sends the user's login information to the server, which then performs authentication.

[0054] Step 3:

[0055] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0056] Step 4:

[0057] The device displays a list of trained models received to the user.

[0058] Step 5:

[0059] The user selects the trained model they want to use and sends that information to their device.

[0060] Step 6:

[0061] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0062] Step 7:

[0063] The server sends the trained model it has loaded to the terminal.

[0064] Step 8:

[0065] The terminal analyzes the received trained model and generates a graphical user interface.

[0066] Step 9:

[0067] The user adjusts the parameters of each trained model using a graphical user interface.

[0068] Step 10:

[0069] The device receives the user's adjustments and transmits them to the server in real time.

[0070] Step 11:

[0071] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0072] Step 12:

[0073] The server generates intermediate output and sends it back to the terminal.

[0074] Step 13:

[0075] The intermediate output received by the terminal is displayed to the user in real time.

[0076] Step 14:

[0077] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0078] Step 15:

[0079] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0080] Step 16:

[0081] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0082] Step 17:

[0083] The terminal sends a "save" request to the server, which generates the final model.

[0084] Step 18:

[0085] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0086] Step 19:

[0087] The device will notify the user that the final model has been saved and provide a download link.

[0088] Step 20:

[0089] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0090] Example 1

[0091] Next, a description will be given of 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."

[0092] Modern machine learning models are highly complex, making it difficult for users to efficiently combine multiple pre-trained models. It is particularly difficult for users without specialized knowledge to intuitively operate the system and adjust models while checking results in real time. Furthermore, there is a lack of efficient ways to adjust and integrate models with different performance metrics, making model optimization time-consuming and labor-intensive.

[0093] 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.

[0094] In this invention, the server includes a means for loading a plurality of trained models from a database or a storage device, a means for adjusting parameters of the plurality of trained models via a graphical user interface, a means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output, and a means for generating and saving a final model. This allows a user, without specialized knowledge, to intuitively and efficiently adjust the plurality of models and generate an optimal model while checking the results in real time.

[0095] "Pre-trained models" are a collection of machine learning algorithms that have been pre-trained using an existing dataset.

[0096] A "database" is a system or device for systematically storing and managing data.

[0097] A "storage device" is hardware or media for storing data and programs.

[0098] A "means" is a method, device, or process for achieving a particular end.

[0099] A "graphical user interface" is a software interface that allows a user to visually interact with the software.

[0100] "Parameters" are setting values ​​that control the operation and behavior of a model.

[0101] "Tuning" means changing settings or configurations to suit a particular purpose.

[0102] "Intermediate outputs" are temporary results produced when combining multiple trained models.

[0103] "Immediate display" means displaying information or data on the screen the moment it is received.

[0104] The "final model" is the final machine learning model that is generated after all tuning and combination is complete.

[0105] "Preservation" refers to the permanent or long-term storage of generated data and models.

[0106] "User" means a person or entity that uses the system to select, adjust, and generate models.

[0107] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained models. This system is composed of a server, a terminal, and a user. The details of the system are described below.

[0108] The server first loads multiple trained models from a database or storage device. For example, the server accesses a database to obtain pre-trained data such as image recognition models and speech recognition models. The server analyzes this model data and sends it to the terminal in an appropriate format.

[0109] The terminal receives and analyzes the model data received from the server. The terminal then displays a graphical user interface that the user can operate intuitively. This interface includes sliders and dials similar to a music equalizer, allowing the user to adjust the parameters of each model. The user can use this to make adjustments while checking the model's behavior in real time.

[0110] The user adjusts the parameters of each model using the interface on the device. Each time this operation is performed, the device sends the adjustment results to the server in real time. The server adjusts the models based on the received parameters and combines multiple models to generate intermediate outputs. These intermediate outputs are then sent back to the device and displayed to the user in real time.

[0111] As a concrete example, consider the case of combining an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0112] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output to the user, who can then confirm it.

[0113] Examples of prompts include:

[0114] "Select an image recognition model and a voice recognition model, and set the image recognition accuracy to 80% and the voice recognition response speed to 90%."

[0115] "Generate intermediate output based on the parameters you set and display the results."

[0116] "Please explain how to hit the save button to generate the final model and save it to the database."

[0117] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device then sends the request to the server, which then generates and saves the final combination result. This final model can then be downloaded by the user or integrated into other systems for use.

[0118] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0119] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0120] Step 1:

[0121] The user accesses the device interface. The user accesses the system's graphical user interface using a PC or mobile device. Input can be a URL or an application launch operation. This launches the interface and displays the model selection screen.

[0122] Step 2:

[0123] The server sends a list of trained models to the device. The server accesses the database and obtains information about the stored trained models (e.g., image recognition models, speech recognition models). The input is a database query. The output is a list of models sent to the device.

[0124] Step 3:

[0125] The user selects the model they want to use from the interface. They select the model they want to use from multiple pre-trained models on their device screen, using checkboxes or drop-down menus. The input is the user's selection operation. The output is the information about the selected model sent to the server.

[0126] Step 4:

[0127] The device requests the user's selection from the server. The device requests the ID and name of the model selected by the user from the server. The input is the model information selected by the user. The output is that the request is sent to the server.

[0128] Step 5:

[0129] The server retrieves the selected model data from the database and sends it to the terminal. The server accesses the database and retrieves the data of the selected model. The input is the selected model information. The output is the corresponding model data sent to the terminal.

[0130] Step 6:

[0131] The terminal displays the model on a graphical user interface. The terminal interprets the model data it receives and displays it in the interface. As input, it has the model data. As output, it presents adjustment tools such as sliders and dials to the user.

[0132] Step 7:

[0133] The user adjusts the model parameters. The user adjusts the parameters of each model by operating sliders and dials. The input is the user's operation information. The output is the adjusted parameters sent to the terminal, which then sends them to the server.

[0134] Step 8:

[0135] The terminal sends the adjusted parameters to the server. The terminal immediately sends the parameters adjusted by the user to the server. The input is the adjusted parameter information. The output is the parameter information sent to the server.

[0136] Step 9:

[0137] The server adjusts the models based on the received parameters and generates intermediate outputs. The server adjusts the behavior of each model based on the adjusted parameters and combines them to generate intermediate outputs. The input is the adjusted parameter information. The output is intermediate output data.

[0138] Step 10:

[0139] The server generates intermediate output and sends it to the terminal. The server generates intermediate output and sends it to the terminal. As input, there is intermediate output data. As output, that data is sent to the terminal.

[0140] Step 11:

[0141] The terminal displays the intermediate output to the user in real time. The terminal displays the intermediate output received from the server to the user in real time. As input, there is intermediate output data. As output, there is intermediate output displayed on the interface.

[0142] Step 12:

[0143] The user fine-tunes the parameters and determines the final model. The user checks the intermediate outputs and fine-tunes the parameters as needed to determine a satisfactory final model. The inputs are the intermediate outputs and user operation information. The output is the final parameter settings.

[0144] Step 13:

[0145] The terminal sends the final parameters to the server. The terminal sends the final parameters determined by the user to the server. As input, there are the final parameter settings. As output, that information is sent to the server.

[0146] Step 14:

[0147] The server generates the final model and stores it in a database or storage device. The server generates the final model based on the final parameters and stores it in a database or storage device. The input is the final parameter settings. The output is the final model data and stores it in a database or storage device.

[0148] Step 15:

[0149] The user downloads the final model or integrates it into another system for use. The user downloads the generated final model via a terminal or integrates it with another system. The inputs are the final model data and the user's download operation. The outputs are the model data stored in the user's environment and the integrated system.

[0150] (Application example 1)

[0151] Next, a description will be given of Application 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."

[0152] In conventional factory robot systems using AI models, optimizing each task required a lot of time and specialized knowledge. Furthermore, there was a lack of means to intuitively combine and use different types of trained models, making it difficult for users to adjust parameters and check the results in real time. To solve this problem, there is a need for a system that efficiently and intuitively combines multiple trained models to optimize the operation of factory robots.

[0153] 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.

[0154] In this invention, the server includes means for loading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for transmitting and displaying the intermediate output to a terminal in the factory in real time, and means for generating and saving a final model. This enables users to intuitively use multiple trained models to optimize the operation of factory robots and effectively utilize complex AI systems without specialized knowledge.

[0155] A "trained model" is an artificial intelligence algorithm that has been pre-trained for a specific task.

[0156] A "database" is an information system that can efficiently manage large amounts of data and search and update it.

[0157] "Storage" is a physical and logical memory device for storing data.

[0158] A "graphical user interface" is an interface that uses visual elements to allow users to operate it intuitively.

[0159] A "parameter" is a setting or variable that affects the behavior or performance of a model.

[0160] "Intermediate output" is a temporary result generated after combining and adjusting multiple trained models.

[0161] "Real time" refers to an operation that responds immediately to a user's operation and displays the result instantly.

[0162] A "factory robot" is an automated mechanical device that performs physical tasks in a factory.

[0163] A "specific task" refers to a specific task or activity that a robot performs within a factory.

[0164] The "final model" is the final AI model that the user has optimized through parameter adjustment and decided to save.

[0165] A "terminal" is a digital device that a user uses to perform operations and check results.

[0166] This invention is a system for generating custom AI models to optimize the operation of robots used in factories. This system generates a unique learning model by efficiently and intuitively combining multiple trained models, and is composed of a server, a terminal, and a user.

[0167] The server runs on AWS EC2 and is configured with a web framework using Django. First, the server loads multiple selected trained models from a database (AWS RDS, PostgreSQL). The loaded model data is sent to the terminal, where it can be operated on the terminal.

[0168] The terminals are iOS / Android devices (smartphones and tablets) that display an intuitive graphical user interface using React Native, which includes sliders and dials similar to a music equalizer, allowing users to adjust the parameters of each model.

[0169] The user adjusts the parameters of each model using the device's interface. For example, they can select an object recognition model and a movement pattern model and set the object recognition accuracy to 90% and the movement reaction speed to 0.5 seconds. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0170] Once the user has further fine-tuned the parameters and is satisfied with the output results, they press the save button to generate the final model. The device then sends the request to the server, which generates the final combination results and stores them in a database or storage. The final model is then downloaded to the robots in the factory and used to perform specific tasks.

[0171] For example, a prompt to a generative AI model might look something like this:

[0172] You need to generate an AI model to deploy optimal robot behavior for specific tasks in your factory. Adjust the AI ​​model to meet the following requirements:

[0173] Object recognition accuracy: 90% or more

[0174] Response time: within 0.5 seconds

[0175] Efficiency of movement patterns: Minimal route and obstacle avoidance

[0176] Use the sliders to adjust these parameters and see the results in real time.

[0177] In this way, users can easily combine multiple pre-trained models to generate custom AI models suited to tasks within their factories, enabling them to effectively utilize complex AI systems without specialized knowledge.

[0178] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0179] Step 1:

[0180] A user accesses the system using a terminal and requests a list of trained models for a specific task of a factory robot. The terminal sends the request to the server. The input to this request is the name of the model selected by the user.

[0181] Step 2:

[0182] The server receives the request, searches for the corresponding trained model in a database (e.g., PostgreSQL on AWS RDS), and retrieves the model data. In this data processing stage, a database query is executed based on the model name to load the appropriate model data. The loaded model data is obtained as output.

[0183] Step 3:

[0184] The server transmits the acquired model data to the terminal, which analyzes the received model data and displays an intuitive graphical user interface that includes sliders and dials that respond in real time.

[0185] Step 4:

[0186] The user adjusts the parameters of each model using the device interface. For example, the accuracy of the object recognition model is set to 90% and the response speed of the movement pattern model is set to 0.5 seconds. These adjustments are sent as data from the device to the server.

[0187] Step 5:

[0188] The server adjusts the behavior of each model based on the received parameters, combines multiple models, and generates intermediate output. Based on the parameter input, it performs data calculations to optimize the internal parameters of the model, and generates an adjusted intermediate model. This intermediate output is obtained as the output.

[0189] Step 6:

[0190] The intermediate output is sent from the server to the terminal, which displays it in real time to the user, who can then check the displayed results and further fine-tune the parameters if necessary. This process can be repeated.

[0191] Step 7:

[0192] When the user is satisfied with the output result, he / she presses the save button to generate the final model. The terminal sends this operation to the server, which then generates the final model. At this stage, the final result of the parameter settings that the user is satisfied with is obtained as input.

[0193] Step 8:

[0194] The server generates the final combination results and saves them in a database or storage. The final model is then downloaded and applied to the robots in the factory, resulting in optimal robot behavior for the specific tasks in the factory.

[0195] Furthermore, an emotion engine that estimates the user's emotion may be 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.

[0196] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0197] The server has the function of loading multiple trained models selected by the user from a database or storage and sending them to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user.

[0198] The device analyzes the model data sent from the server and provides the user with an interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model. The device also includes an emotion engine that recognizes emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0199] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, and the server adjusts the model based on the adjusted parameters and combines them to generate intermediate output. The generated intermediate output is then sent back to the device and displayed to the user in real time.

[0200] The emotion engine analyzes facial expressions, voice, and typing patterns when a user operates the interface to recognize the user's emotional state. Based on this recognized emotion, the emotion engine recommends parameters appropriate for the user. It is also possible to automatically customize the interface according to the user's emotions.

[0201] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the device interface. The device then requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[0202] Next, the user uses the sliders on the interface to set "image recognition accuracy" to 80% and "voice recognition response speed" to 90%. At this point, the emotion engine recognizes the user's emotions and recommends parameter settings that are easier to use if the user is nervous, for example. Alternatively, the interface layout and design can be changed depending on the user's emotions.

[0203] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user, who can check the results and further fine-tune the parameters.

[0204] When the user finally obtains the desired result, they press the "Save" button to generate the final model. The terminal sends the request to the server, which generates the final combination result and saves it in a database or storage. The user can download this final model or integrate it into other systems for use.

[0205] In this way, the present invention makes it possible to effectively combine multiple trained models while taking into account the user's emotional state, thereby intuitively and efficiently generating a unique AI model.

[0206] The processing flow will be explained below.

[0207] Step 1:

[0208] A user logs in to the system and accesses a screen that displays a list of trained models.

[0209] Step 2:

[0210] The terminal sends the user's login information to the server, which then performs authentication.

[0211] Step 3:

[0212] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0213] Step 4:

[0214] The device displays a list of trained models received to the user.

[0215] Step 5:

[0216] The user selects the trained model they want to use and sends that information to their device.

[0217] Step 6:

[0218] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0219] Step 7:

[0220] The server sends the trained model it has loaded to the terminal.

[0221] Step 8:

[0222] The device analyzes the received trained model and generates a graphical user interface, which includes sliders and dials similar to a music equalizer.

[0223] Step 9:

[0224] The user adjusts the parameters of each trained model using a graphical user interface.

[0225] Step 10:

[0226] The emotion engine analyzes the user's facial expressions, voice, and typing patterns in real time while they are operating the device, recognizing their emotional state.

[0227] Step 11:

[0228] The device recommends appropriate parameter settings and customizes the interface based on the user's emotions recognized by the emotion engine.

[0229] Step 12:

[0230] The device receives the user's adjustments and transmits them to the server in real time.

[0231] Step 13:

[0232] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0233] Step 14:

[0234] The server generates intermediate output and sends it back to the terminal.

[0235] Step 15:

[0236] The intermediate output received by the terminal is displayed to the user in real time.

[0237] Step 16:

[0238] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0239] Step 17:

[0240] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0241] Step 18:

[0242] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0243] Step 19:

[0244] The terminal sends a "save" request to the server, which generates the final model.

[0245] Step 20:

[0246] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0247] Step 21:

[0248] The device will notify the user that the final model has been saved and provide a download link.

[0249] Step 22:

[0250] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0251] Example 2

[0252] Next, a description will be given of 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."

[0253] Conventional AI model generation systems have difficulty efficiently and intuitively combining multiple trained models, placing a heavy burden on users. Furthermore, they are unable to adjust parameters or customize interfaces that take the user's emotional state into account, resulting in a lack of improvement in the user experience.

[0254] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating an intermediate output, means for incorporating an emotion engine that recognizes the user's emotional state and recommends parameters, means for automatically customizing the interface according to the user's emotional state, and means for generating and saving a final model. This makes it possible to intuitively and efficiently combine multiple trained models while taking the user's emotional state into consideration to generate a unique AI model.

[0255] A "pre-trained model" is an artificial intelligence model that has already been trained and optimized to perform a specific task.

[0256] A "database" is an electronic system for efficiently storing and retrieving data.

[0257] "Storage" is a physical or virtual space for storing digital data.

[0258] A "graphical user interface" is an interface that includes elements (such as buttons, sliders, and dials) that a user can visually manipulate.

[0259] A "parameter" is a setting value or variable that controls the behavior of an AI model.

[0260] An "emotion engine" is an artificial intelligence engine that identifies a user's emotions and provides appropriate responses and recommendations.

[0261] "Intermediate output" refers to an intermediate result generated before obtaining the final output.

[0262] The "final model" is the AI ​​model that is completed as a result of adjustments and combinations.

[0263] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0264] Hardware and Software Configuration

[0265] server

[0266] The server functions as a means of loading multiple trained models from a database or storage (for example, MySQL or PostgreSQL is used as database software), sending the model data selected by the user to the terminal, and reconstructing the model based on the adjusted parameters.

[0267] Terminal

[0268] The terminal is a means for analyzing the model data sent from the server. The terminal provides a user-operable graphical user interface (GUI). This GUI includes sliders and dials that the user uses to adjust the model parameters. The terminal also includes an emotion engine that uses OpenCV and TensorFlow to recognize emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0269] Emotion Engine

[0270] The emotion engine is a means of recognizing the user's emotions. It is possible to use PyTorch or Keras as a deep learning framework. This emotion engine has the function of recommending appropriate parameters for the user based on the analyzed emotional state and automatically customizing the interface.

[0271] Specific examples

[0272] For example, if a user wants to combine an image recognition model and a voice recognition model, they can use the system by following the steps below.

[0273] 1. The user selects the "image recognition model" and "voice recognition model" from the device interface.

[0274] 2. The device requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[0275] 3. The user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%.

[0276] 4. The emotion engine analyzes the user's facial expressions and voice and recommends easier parameter settings if the user is nervous. It can also change the layout and design of the interface according to the user's emotions.

[0277] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The user can check the results and further fine-tune the parameters.

[0278] Example of input prompt for generative AI model

[0279] 1. User prompt: "I want to combine image and speech recognition models, with image recognition accuracy set to 80% and speech recognition response speed set to 90%."

[0280] 2. Prompt for the generative AI model: "Please combine the image recognition and voice recognition models under the following conditions: Image recognition accuracy: 80%, Voice recognition response speed: 90%. The emotion engine will analyze the user's emotions and recommend suitable parameters."

[0281] This system allows users to generate their own unique AI models by efficiently combining multiple trained models and intuitively operating them while taking into account their own emotional state.

[0282] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0283] Step 1:

[0284] Initialization and model selection

[0285] input:

[0286] A user starts up a device and sends a request to display a list of available trained models.

[0287] output:

[0288] A list of available trained models will be displayed on your device.

[0289] Specific behavior:

[0290] The device initializes the system and requests a list of trained models from the server. The server retrieves the model list from a database (e.g., MySQL or PostgreSQL) and sends it to the device. The device parses it and displays it to the user.

[0291] Step 2:

[0292] Model selection and data loading

[0293] input:

[0294] The user selects the "image recognition model" and "voice recognition model" on the device interface.

[0295] output:

[0296] The selected model data is sent from the server to the terminal.

[0297] Specific behavior:

[0298] Based on the user's operation, the device sends a request for the selected model (e.g., "image recognition model" and "voice recognition model") to the server. The server reads the corresponding model data from the database and sends it to the device.

[0299] Step 3:

[0300] Interface display and parameter adjustment

[0301] input:

[0302] The terminal receives the model data sent from the server.

[0303] output:

[0304] An interface is displayed that allows the user to adjust the parameters.

[0305] Specific behavior:

[0306] The device analyzes the received model data and provides the user with a graphical user interface (GUI) for adjusting the model parameters. It displays UI elements such as sliders and dials that the user can use to adjust the parameters.

[0307] Step 4:

[0308] Parameter submission and model adjustment

[0309] input:

[0310] The user sets parameters on the interface (e.g., "image recognition accuracy" at 80%, "voice recognition response speed" at 90%).

[0311] output:

[0312] The adjusted parameters are sent to the server.

[0313] Specific behavior:

[0314] The parameters set by the user are sent in real time from the device to the server, which then adjusts and combines the selected model based on the received parameters.

[0315] Step 5:

[0316] Generating and displaying intermediate output

[0317] input:

[0318] The server receives the adjusted parameters.

[0319] output:

[0320] An intermediate output is generated and sent to the terminal.

[0321] Specific behavior:

[0322] The server adjusts the AI ​​model based on the adjusted parameters and generates intermediate output, which is sent to the device in JSON format, etc. The device analyzes the received intermediate output and displays it to the user in real time.

[0323] Step 6:

[0324] Analysis and recommendation by emotion engine

[0325] input:

[0326] Emotional data such as the user's facial expression, voice, and typing pattern are input.

[0327] output:

[0328] Emotion-based parameter recommendations and interface customization are performed.

[0329] Specific behavior:

[0330] The device's built-in emotion engine (using OpenCV and TensorFlow) analyzes the user's facial expressions and voice to recognize their emotional state at that time. Based on the recognition results, the emotion engine recommends parameter settings appropriate for the user and customizes the interface.

[0331] Step 7:

[0332] Save and download the final model

[0333] input:

[0334] When the user obtains a satisfactory output result, he or she presses the "save" button.

[0335] output:

[0336] The final model is saved to a database or storage and a download link is provided.

[0337] Specific behavior:

[0338] When the user presses the "Save" button, the device sends a request to the server to save the final model. The server generates the final model and saves it in a database or storage. The server then returns a download link for the final model to the device, and the user clicks the link to download the final model.

[0339] These processing steps enable users to efficiently and intuitively combine multiple pre-trained AI models to generate their own unique AI models, and also use the emotion engine to recommend parameters and customize the interface.

[0340] (Application example 2)

[0341] Next, a description will be given of Application 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."

[0342] In recent years, advances in AI technology have created a demand for combining and using multiple trained models. However, there is no efficient and intuitive way to combine these models. Furthermore, few systems take the user's emotional state into account, and parameter settings and interface customization are cumbersome. There is a need for a system that can solve these issues and provide a better user experience based on the user's emotions.

[0343] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for recommending parameters based on the user's emotional state, and means for customizing the interface according to the user's emotional state. This makes it possible to efficiently and intuitively combine multiple trained models while taking the user's emotional state into consideration.

[0344] A "trained model" is a machine learning algorithm that has been trained in advance using large amounts of data and has the ability to perform a specific task.

[0345] A "database" is a system that organizes and stores data so that it can be accessed and manipulated as needed.

[0346] "Storage" refers to a storage device for long-term storage of data.

[0347] A "graphical user interface" refers to an interface that a user can visually interact with, typically including graphical elements such as windows, buttons, sliders, etc.

[0348] "Parameters" refer to the settings that control the performance and behavior of a model.

[0349] "Intermediate output" refers to the intermediate results obtained when combining multiple trained models.

[0350] "Real-time" means that operations or processes occur immediately, without delay.

[0351] The "final model" refers to the machine learning model that is ultimately created based on the output results desired by the user.

[0352] "Emotional state" refers to a user's current emotional or psychological state.

[0353] "Parameter recommendation" refers to the act of the system suggesting appropriate setting values ​​for parameters set by the user.

[0354] "Interface customization" refers to changing the design and layout of an interface in response to a user's specific requirements or emotional state.

[0355] This invention is a system that efficiently and intuitively combines multiple trained models, recommends parameters taking into account the user's emotional state, and customizes the interface. It is composed of a server, a terminal, and a user. To realize this system, the following processes are performed.

[0356] server

[0357] The server has the function of loading multiple trained models from a database or storage. It loads the model data selected by the user and sends it to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user. It also combines models based on the user's parameter adjustment requests, generates intermediate output, and sends it back to the terminal.

[0358] Terminal

[0359] The device analyzes the model data sent from the server and provides the user with a graphical user interface, which includes sliders and dials similar to a music equalizer to adjust the parameters of each model.The device also includes an emotion engine that analyzes the user's facial expressions, voice, and typing patterns to recognize their emotional state.

[0360] User

[0361] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, which then adjusts and combines the models based on the parameters to generate an intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The emotion engine analyzes the user's emotional state and recommends optimal parameters based on that state. The interface layout and design can also be customized based on the emotional state. When the user is satisfied with the output results, they press the "Save" button to generate the final model. The device sends this request to the server, which generates the final combination results and stores them in a database or storage.

[0362] Specific examples

[0363] For example, consider the use of this technology in a "personalized video viewing application." This application recommends the most suitable videos based on the user's emotional state, improving the viewing experience. When a user requests videos of a specific genre or theme, the emotion engine analyzes the user's emotions and generates a list of recommended videos based on that. The interface also changes dynamically depending on the user's emotional state. Specifically, the following prompts are input to the generative AI model:

[0364] Example prompt sentence:

[0365] Read the user's facial expressions and recommend the following list of movies and documentaries.

[0366] 1. Inspirational Movies

[0367] 2. An interesting documentary

[0368] This system can understand the user's emotional state and provide viewing content that is more optimal for each individual.

[0369] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0370] Step 1:

[0371] The server reads multiple trained models selected by the user from the database. It receives the user's selection, retrieves the model data from the database, and sends it to the device. The input is the user's model selection information, and the output is the model data.

[0372] Step 2:

[0373] The terminal analyzes the model data sent from the server, converts the acquired model data into an appropriate format, and makes it operable by the user. The input is the model data from the server, and the output is the model data converted into a format that can be operated by the user.

[0374] Step 3:

[0375] The terminal provides a graphical user interface, which includes sliders and dials that the user can manipulate directly to adjust the parameters of each model. The input is the converted model data, and the output is a parameter setting screen that can be manipulated in the interface.

[0376] Step 4:

[0377] The user adjusts the parameters of each model using an interface. The emotion engine analyzes the user's facial expressions, voice, and typing patterns to recognize the user's emotional state at that time. The input is the user's actions and emotional data, and the output is the adjusted parameters.

[0378] Step 5:

[0379] The server receives the parameters adjusted by the user, combines multiple trained models based on them, and generates intermediate outputs. The inputs are the parameters adjusted by the user, and the outputs are the intermediate outputs.

[0380] Step 6:

[0381] The terminal displays the intermediate output sent from the server in real time. The interface and intermediate output are updated every time the user changes a parameter. The input is the intermediate output from the server, and the output is the updated interface and display of the intermediate output.

[0382] Step 7:

[0383] The emotion engine recommends optimal parameters based on the user's current emotional state. As the user adjusts parameters, the emotion engine analyzes them in real time and suggests appropriate settings. The input is the user's emotional data, and the output is the recommended parameter settings.

[0384] Step 8:

[0385] The device customizes the interface based on the analysis results of the emotion engine. The color and layout of the interface are changed according to the emotional state. The input is the analyzed emotional data, and the output is a customized interface.

[0386] Step 9:

[0387] When the user is satisfied with the output results, they can click the "Save" button to generate and save the final model. The input is the user's save request, and the output is the generated final model and its saving.

[0388] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[0389] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[0390] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0391] [Second embodiment]

[0392] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0393] 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.

[0394] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0395] 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.

[0396] 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 audio in accordance with instructions from the processor 46.

[0397] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0398] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0399] Fig. 4 shows an example of the 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.

[0400] 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.

[0401] 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.

[0402] In the smart glasses 214, the processor 46 performs the reception output process. 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.

[0403] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0404] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models. This system consists of a server, a terminal, and a user, and allows the user to adjust the parameters of each model using a graphical user interface and generate the final model while checking the results in real time.

[0405] The server is responsible for loading multiple trained models selected by the user from a database or storage, and then sending the loaded model data to the device, making it operable on the device.

[0406] The device analyzes the model data received from the server and displays an intuitive graphical interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model.

[0407] The user adjusts the parameters of each model using the interface on the device. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0408] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. First, the user selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0409] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output result to the user, who can then confirm it.

[0410] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device sends the request to the server, which generates the final combination result and saves it in a database or storage. The final model can then be downloaded by the user or integrated into other systems for use.

[0411] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0412] The processing flow will be explained below.

[0413] Step 1:

[0414] A user logs in to the system and accesses a screen that displays a list of trained models.

[0415] Step 2:

[0416] The terminal sends the user's login information to the server, which then performs authentication.

[0417] Step 3:

[0418] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0419] Step 4:

[0420] The device displays a list of trained models received to the user.

[0421] Step 5:

[0422] The user selects the trained model they want to use and sends that information to their device.

[0423] Step 6:

[0424] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0425] Step 7:

[0426] The server sends the trained model it has loaded to the terminal.

[0427] Step 8:

[0428] The terminal analyzes the received trained model and generates a graphical user interface.

[0429] Step 9:

[0430] The user adjusts the parameters of each trained model using a graphical user interface.

[0431] Step 10:

[0432] The device receives the user's adjustments and transmits them to the server in real time.

[0433] Step 11:

[0434] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0435] Step 12:

[0436] The server generates intermediate output and sends it back to the terminal.

[0437] Step 13:

[0438] The intermediate output received by the terminal is displayed to the user in real time.

[0439] Step 14:

[0440] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0441] Step 15:

[0442] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0443] Step 16:

[0444] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0445] Step 17:

[0446] The terminal sends a "save" request to the server, which generates the final model.

[0447] Step 18:

[0448] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0449] Step 19:

[0450] The device will notify the user that the final model has been saved and provide a download link.

[0451] Step 20:

[0452] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0453] Example 1

[0454] Next, a description will be given of 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."

[0455] Modern machine learning models are highly complex, making it difficult for users to efficiently combine multiple pre-trained models. It is particularly difficult for users without specialized knowledge to intuitively operate the system and adjust models while checking results in real time. Furthermore, there is a lack of efficient ways to adjust and integrate models with different performance metrics, making model optimization time-consuming and labor-intensive.

[0456] 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.

[0457] In this invention, the server includes a means for loading a plurality of trained models from a database or a storage device, a means for adjusting parameters of the plurality of trained models via a graphical user interface, a means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output, and a means for generating and saving a final model. This allows a user, without specialized knowledge, to intuitively and efficiently adjust the plurality of models and generate an optimal model while checking the results in real time.

[0458] "Pre-trained models" are a collection of machine learning algorithms that have been pre-trained using an existing dataset.

[0459] A "database" is a system or device for systematically storing and managing data.

[0460] A "storage device" is hardware or media for storing data and programs.

[0461] A "means" is a method, device, or process for achieving a particular end.

[0462] A "graphical user interface" is a software interface that allows a user to visually interact with the software.

[0463] "Parameters" are setting values ​​that control the operation and behavior of a model.

[0464] "Tuning" means changing settings or configurations to suit a particular purpose.

[0465] "Intermediate outputs" are temporary results produced when combining multiple trained models.

[0466] "Immediate display" means displaying information or data on the screen the moment it is received.

[0467] The "final model" is the final machine learning model that is generated after all tuning and combination is complete.

[0468] "Preservation" refers to the permanent or long-term storage of generated data and models.

[0469] "User" means a person or entity that uses the system to select, adjust, and generate models.

[0470] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained models. This system is composed of a server, a terminal, and a user. The details of the system are described below.

[0471] The server first loads multiple trained models from a database or storage device. For example, the server accesses a database to obtain pre-trained data such as image recognition models and speech recognition models. The server analyzes this model data and sends it to the terminal in an appropriate format.

[0472] The terminal receives and analyzes the model data received from the server. The terminal then displays a graphical user interface that the user can operate intuitively. This interface includes sliders and dials similar to a music equalizer, allowing the user to adjust the parameters of each model. The user can use this to make adjustments while checking the model's behavior in real time.

[0473] The user adjusts the parameters of each model using the interface on the device. Each time this operation is performed, the device sends the adjustment results to the server in real time. The server adjusts the models based on the received parameters and combines multiple models to generate intermediate outputs. These intermediate outputs are then sent back to the device and displayed to the user in real time.

[0474] As a concrete example, consider the case of combining an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0475] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output to the user, who can then confirm it.

[0476] Examples of prompts include:

[0477] "Select an image recognition model and a voice recognition model, and set the image recognition accuracy to 80% and the voice recognition response speed to 90%."

[0478] "Generate intermediate output based on the parameters you set and display the results."

[0479] "Please explain how to hit the save button to generate the final model and save it to the database."

[0480] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device then sends the request to the server, which then generates and saves the final combination result. This final model can then be downloaded by the user or integrated into other systems for use.

[0481] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0482] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0483] Step 1:

[0484] The user accesses the device interface. The user accesses the system's graphical user interface using a PC or mobile device. Input can be a URL or an application launch operation. This launches the interface and displays the model selection screen.

[0485] Step 2:

[0486] The server sends a list of trained models to the device. The server accesses the database and obtains information about the stored trained models (e.g., image recognition models, speech recognition models). The input is a database query. The output is a list of models sent to the device.

[0487] Step 3:

[0488] The user selects the model they want to use from the interface. They select the model they want to use from multiple pre-trained models on their device screen, using checkboxes or drop-down menus. The input is the user's selection operation. The output is the information about the selected model sent to the server.

[0489] Step 4:

[0490] The device requests the user's selection from the server. The device requests the ID and name of the model selected by the user from the server. The input is the model information selected by the user. The output is that the request is sent to the server.

[0491] Step 5:

[0492] The server retrieves the selected model data from the database and sends it to the terminal. The server accesses the database and retrieves the data of the selected model. The input is the selected model information. The output is the corresponding model data sent to the terminal.

[0493] Step 6:

[0494] The terminal displays the model on a graphical user interface. The terminal interprets the model data it receives and displays it in the interface. As input, it has the model data. As output, it presents adjustment tools such as sliders and dials to the user.

[0495] Step 7:

[0496] The user adjusts the model parameters. The user adjusts the parameters of each model by operating sliders and dials. The input is the user's operation information. The output is the adjusted parameters sent to the terminal, which then sends them to the server.

[0497] Step 8:

[0498] The terminal sends the adjusted parameters to the server. The terminal immediately sends the parameters adjusted by the user to the server. The input is the adjusted parameter information. The output is the parameter information sent to the server.

[0499] Step 9:

[0500] The server adjusts the models based on the received parameters and generates intermediate outputs. The server adjusts the behavior of each model based on the adjusted parameters and combines them to generate intermediate outputs. The input is the adjusted parameter information. The output is intermediate output data.

[0501] Step 10:

[0502] The server generates intermediate output and sends it to the terminal. The server generates intermediate output and sends it to the terminal. As input, there is intermediate output data. As output, that data is sent to the terminal.

[0503] Step 11:

[0504] The terminal displays the intermediate output to the user in real time. The terminal displays the intermediate output received from the server to the user in real time. As input, there is intermediate output data. As output, there is intermediate output displayed on the interface.

[0505] Step 12:

[0506] The user fine-tunes the parameters and determines the final model. The user checks the intermediate outputs and fine-tunes the parameters as needed to determine a satisfactory final model. The inputs are the intermediate outputs and user operation information. The output is the final parameter settings.

[0507] Step 13:

[0508] The terminal sends the final parameters to the server. The terminal sends the final parameters determined by the user to the server. As input, there are the final parameter settings. As output, that information is sent to the server.

[0509] Step 14:

[0510] The server generates the final model and stores it in a database or storage device. The server generates the final model based on the final parameters and stores it in a database or storage device. The input is the final parameter settings. The output is the final model data and stores it in a database or storage device.

[0511] Step 15:

[0512] The user downloads the final model or integrates it into another system for use. The user downloads the generated final model via a terminal or integrates it with another system. The inputs are the final model data and the user's download operation. The outputs are the model data stored in the user's environment and the integrated system.

[0513] (Application example 1)

[0514] Next, a description will be given of Application 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."

[0515] In conventional factory robot systems using AI models, optimizing each task required a lot of time and specialized knowledge. Furthermore, there was a lack of means to intuitively combine and use different types of trained models, making it difficult for users to adjust parameters and check the results in real time. To solve this problem, there is a need for a system that efficiently and intuitively combines multiple trained models to optimize the operation of factory robots.

[0516] 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.

[0517] In this invention, the server includes means for loading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for transmitting and displaying the intermediate output to a terminal in the factory in real time, and means for generating and saving a final model. This enables users to intuitively use multiple trained models to optimize the operation of factory robots and effectively utilize complex AI systems without specialized knowledge.

[0518] A "trained model" is an artificial intelligence algorithm that has been pre-trained for a specific task.

[0519] A "database" is an information system that can efficiently manage large amounts of data and search and update it.

[0520] "Storage" is a physical and logical memory device for storing data.

[0521] A "graphical user interface" is an interface that uses visual elements to allow users to operate it intuitively.

[0522] A "parameter" is a setting or variable that affects the behavior or performance of a model.

[0523] "Intermediate output" is a temporary result generated after combining and adjusting multiple trained models.

[0524] "Real time" refers to an operation that responds immediately to a user's operation and displays the result instantly.

[0525] A "factory robot" is an automated mechanical device that performs physical tasks in a factory.

[0526] A "specific task" refers to a specific task or activity that a robot performs within a factory.

[0527] The "final model" is the final AI model that the user has optimized through parameter adjustment and decided to save.

[0528] A "terminal" is a digital device that a user uses to perform operations and check results.

[0529] This invention is a system for generating custom AI models to optimize the operation of robots used in factories. This system generates a unique learning model by efficiently and intuitively combining multiple trained models, and is composed of a server, a terminal, and a user.

[0530] The server runs on AWS EC2 and is configured with a web framework using Django. First, the server loads multiple selected trained models from a database (AWS RDS, PostgreSQL). The loaded model data is sent to the terminal, where it can be operated on the terminal.

[0531] The terminals are iOS / Android devices (smartphones and tablets) that display an intuitive graphical user interface using React Native, which includes sliders and dials similar to a music equalizer, allowing users to adjust the parameters of each model.

[0532] The user adjusts the parameters of each model using the device's interface. For example, they can select an object recognition model and a movement pattern model and set the object recognition accuracy to 90% and the movement reaction speed to 0.5 seconds. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0533] Once the user has further fine-tuned the parameters and is satisfied with the output results, they press the save button to generate the final model. The device then sends the request to the server, which generates the final combination results and stores them in a database or storage. The final model is then downloaded to the robots in the factory and used to perform specific tasks.

[0534] For example, a prompt to a generative AI model might look something like this:

[0535] You need to generate an AI model to deploy optimal robot behavior for specific tasks in your factory. Adjust the AI ​​model to meet the following requirements:

[0536] Object recognition accuracy: 90% or more

[0537] Response time: within 0.5 seconds

[0538] Efficiency of movement patterns: Minimal route and obstacle avoidance

[0539] Use the sliders to adjust these parameters and see the results in real time.

[0540] In this way, users can easily combine multiple pre-trained models to generate custom AI models suited to tasks within their factories, enabling them to effectively utilize complex AI systems without specialized knowledge.

[0541] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0542] Step 1:

[0543] A user accesses the system using a terminal and requests a list of trained models for a specific task of a factory robot. The terminal sends the request to the server. The input to this request is the name of the model selected by the user.

[0544] Step 2:

[0545] The server receives the request, searches for the corresponding trained model in a database (e.g., PostgreSQL on AWS RDS), and retrieves the model data. In this data processing stage, a database query is executed based on the model name to load the appropriate model data. The loaded model data is obtained as output.

[0546] Step 3:

[0547] The server transmits the acquired model data to the terminal, which analyzes the received model data and displays an intuitive graphical user interface that includes sliders and dials that respond in real time.

[0548] Step 4:

[0549] The user adjusts the parameters of each model using the device interface. For example, the accuracy of the object recognition model is set to 90% and the response speed of the movement pattern model is set to 0.5 seconds. These adjustments are sent as data from the device to the server.

[0550] Step 5:

[0551] The server adjusts the behavior of each model based on the received parameters, combines multiple models, and generates intermediate output. Based on the parameter input, it performs data calculations to optimize the internal parameters of the model, and generates an adjusted intermediate model. This intermediate output is obtained as the output.

[0552] Step 6:

[0553] The intermediate output is sent from the server to the terminal, which displays it in real time to the user, who can then check the displayed results and further fine-tune the parameters if necessary. This process can be repeated.

[0554] Step 7:

[0555] When the user is satisfied with the output result, he / she presses the save button to generate the final model. The terminal sends this operation to the server, which then generates the final model. At this stage, the final result of the parameter settings that the user is satisfied with is obtained as input.

[0556] Step 8:

[0557] The server generates the final combination results and saves them in a database or storage. The final model is then downloaded and applied to the robots in the factory, resulting in optimal robot behavior for the specific tasks in the factory.

[0558] Furthermore, 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.

[0559] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0560] The server has the function of loading multiple trained models selected by the user from a database or storage and sending them to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user.

[0561] The device analyzes the model data sent from the server and provides the user with an interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model. The device also includes an emotion engine that recognizes emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0562] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, and the server adjusts the model based on the adjusted parameters and combines them to generate intermediate output. The generated intermediate output is then sent back to the device and displayed to the user in real time.

[0563] The emotion engine analyzes facial expressions, voice, and typing patterns when a user operates the interface to recognize the user's emotional state. Based on this recognized emotion, the emotion engine recommends parameters appropriate for the user. It is also possible to automatically customize the interface according to the user's emotions.

[0564] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the device interface. The device then requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[0565] Next, the user uses the sliders on the interface to set "image recognition accuracy" to 80% and "voice recognition response speed" to 90%. At this point, the emotion engine recognizes the user's emotions and recommends parameter settings that are easier to use if the user is nervous, for example. Alternatively, the interface layout and design can be changed depending on the user's emotions.

[0566] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user, who can check the results and further fine-tune the parameters.

[0567] When the user finally obtains the desired result, they press the "Save" button to generate the final model. The terminal sends the request to the server, which generates the final combination result and saves it in a database or storage. The user can download this final model or integrate it into other systems for use.

[0568] In this way, the present invention makes it possible to effectively combine multiple trained models while taking into account the user's emotional state, thereby intuitively and efficiently generating a unique AI model.

[0569] The processing flow will be explained below.

[0570] Step 1:

[0571] A user logs in to the system and accesses a screen that displays a list of trained models.

[0572] Step 2:

[0573] The terminal sends the user's login information to the server, which then performs authentication.

[0574] Step 3:

[0575] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0576] Step 4:

[0577] The device displays a list of trained models received to the user.

[0578] Step 5:

[0579] The user selects the trained model they want to use and sends that information to their device.

[0580] Step 6:

[0581] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0582] Step 7:

[0583] The server sends the trained model it has loaded to the terminal.

[0584] Step 8:

[0585] The device analyzes the received trained model and generates a graphical user interface, which includes sliders and dials similar to a music equalizer.

[0586] Step 9:

[0587] The user adjusts the parameters of each trained model using a graphical user interface.

[0588] Step 10:

[0589] The emotion engine analyzes the user's facial expressions, voice, and typing patterns in real time while they are operating the device, recognizing their emotional state.

[0590] Step 11:

[0591] The device recommends appropriate parameter settings and customizes the interface based on the user's emotions recognized by the emotion engine.

[0592] Step 12:

[0593] The device receives the user's adjustments and transmits them to the server in real time.

[0594] Step 13:

[0595] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0596] Step 14:

[0597] The server generates intermediate output and sends it back to the terminal.

[0598] Step 15:

[0599] The intermediate output received by the terminal is displayed to the user in real time.

[0600] Step 16:

[0601] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0602] Step 17:

[0603] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0604] Step 18:

[0605] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0606] Step 19:

[0607] The terminal sends a "save" request to the server, which generates the final model.

[0608] Step 20:

[0609] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0610] Step 21:

[0611] The device will notify the user that the final model has been saved and provide a download link.

[0612] Step 22:

[0613] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0614] Example 2

[0615] Next, a description will be given of 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."

[0616] Conventional AI model generation systems have difficulty efficiently and intuitively combining multiple trained models, placing a heavy burden on users. Furthermore, they are unable to adjust parameters or customize interfaces that take the user's emotional state into account, resulting in a lack of improvement in the user experience.

[0617] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating an intermediate output, means for incorporating an emotion engine that recognizes the user's emotional state and recommends parameters, means for automatically customizing the interface according to the user's emotional state, and means for generating and saving a final model. This makes it possible to intuitively and efficiently combine multiple trained models while taking the user's emotional state into consideration to generate a unique AI model.

[0618] A "pre-trained model" is an artificial intelligence model that has already been trained and optimized to perform a specific task.

[0619] A "database" is an electronic system for efficiently storing and retrieving data.

[0620] "Storage" is a physical or virtual space for storing digital data.

[0621] A "graphical user interface" is an interface that includes elements (such as buttons, sliders, and dials) that a user can visually manipulate.

[0622] A "parameter" is a setting value or variable that controls the behavior of an AI model.

[0623] An "emotion engine" is an artificial intelligence engine that identifies a user's emotions and provides appropriate responses and recommendations.

[0624] "Intermediate output" refers to an intermediate result generated before obtaining the final output.

[0625] The "final model" is the AI ​​model that is completed as a result of adjustments and combinations.

[0626] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0627] Hardware and Software Configuration

[0628] server

[0629] The server functions as a means of loading multiple trained models from a database or storage (for example, MySQL or PostgreSQL is used as database software), sending the model data selected by the user to the terminal, and reconstructing the model based on the adjusted parameters.

[0630] Terminal

[0631] The terminal is a means for analyzing the model data sent from the server. The terminal provides a user-operable graphical user interface (GUI). This GUI includes sliders and dials that the user uses to adjust the model parameters. The terminal also includes an emotion engine that uses OpenCV and TensorFlow to recognize emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0632] Emotion Engine

[0633] The emotion engine is a means of recognizing the user's emotions. It is possible to use PyTorch or Keras as a deep learning framework. This emotion engine has the function of recommending appropriate parameters for the user based on the analyzed emotional state and automatically customizing the interface.

[0634] Specific examples

[0635] For example, if a user wants to combine an image recognition model and a voice recognition model, they can use the system by following the steps below.

[0636] 1. The user selects the "image recognition model" and "voice recognition model" from the device interface.

[0637] 2. The device requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[0638] 3. The user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%.

[0639] 4. The emotion engine analyzes the user's facial expressions and voice and recommends easier parameter settings if the user is nervous. It can also change the layout and design of the interface according to the user's emotions.

[0640] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The user can check the results and further fine-tune the parameters.

[0641] Example of input prompt for generative AI model

[0642] 1. User prompt: "I want to combine image and speech recognition models, with image recognition accuracy set to 80% and speech recognition response speed set to 90%."

[0643] 2. Prompt for the generative AI model: "Please combine the image recognition and voice recognition models under the following conditions: Image recognition accuracy: 80%, Voice recognition response speed: 90%. The emotion engine will analyze the user's emotions and recommend suitable parameters."

[0644] This system allows users to generate their own unique AI models by efficiently combining multiple trained models and intuitively operating them while taking into account their own emotional state.

[0645] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0646] Step 1:

[0647] Initialization and model selection

[0648] input:

[0649] A user starts up a device and sends a request to display a list of available trained models.

[0650] output:

[0651] A list of available trained models will be displayed on your device.

[0652] Specific behavior:

[0653] The device initializes the system and requests a list of trained models from the server. The server retrieves the model list from a database (e.g., MySQL or PostgreSQL) and sends it to the device. The device parses it and displays it to the user.

[0654] Step 2:

[0655] Model selection and data loading

[0656] input:

[0657] The user selects the "image recognition model" and "voice recognition model" on the device interface.

[0658] output:

[0659] The selected model data is sent from the server to the terminal.

[0660] Specific behavior:

[0661] Based on the user's operation, the device sends a request for the selected model (e.g., "image recognition model" and "voice recognition model") to the server. The server reads the corresponding model data from the database and sends it to the device.

[0662] Step 3:

[0663] Interface display and parameter adjustment

[0664] input:

[0665] The terminal receives the model data sent from the server.

[0666] output:

[0667] An interface is displayed that allows the user to adjust the parameters.

[0668] Specific behavior:

[0669] The device analyzes the received model data and provides the user with a graphical user interface (GUI) for adjusting the model parameters. It displays UI elements such as sliders and dials that the user can use to adjust the parameters.

[0670] Step 4:

[0671] Parameter submission and model adjustment

[0672] input:

[0673] The user sets parameters on the interface (e.g., "image recognition accuracy" at 80%, "voice recognition response speed" at 90%).

[0674] output:

[0675] The adjusted parameters are sent to the server.

[0676] Specific behavior:

[0677] The parameters set by the user are sent in real time from the device to the server, which then adjusts and combines the selected model based on the received parameters.

[0678] Step 5:

[0679] Generating and displaying intermediate output

[0680] input:

[0681] The server receives the adjusted parameters.

[0682] output:

[0683] An intermediate output is generated and sent to the terminal.

[0684] Specific behavior:

[0685] The server adjusts the AI ​​model based on the adjusted parameters and generates intermediate output, which is sent to the device in JSON format, etc. The device analyzes the received intermediate output and displays it to the user in real time.

[0686] Step 6:

[0687] Analysis and recommendation by emotion engine

[0688] input:

[0689] Emotional data such as the user's facial expression, voice, and typing pattern are input.

[0690] output:

[0691] Emotion-based parameter recommendations and interface customization are performed.

[0692] Specific behavior:

[0693] The device's built-in emotion engine (using OpenCV and TensorFlow) analyzes the user's facial expressions and voice to recognize their emotional state at that time. Based on the recognition results, the emotion engine recommends parameter settings appropriate for the user and customizes the interface.

[0694] Step 7:

[0695] Save and download the final model

[0696] input:

[0697] When the user obtains a satisfactory output result, he or she presses the "save" button.

[0698] output:

[0699] The final model is saved to a database or storage and a download link is provided.

[0700] Specific behavior:

[0701] When the user presses the "Save" button, the device sends a request to the server to save the final model. The server generates the final model and saves it in a database or storage. The server then returns a download link for the final model to the device, and the user clicks the link to download the final model.

[0702] These processing steps enable users to efficiently and intuitively combine multiple pre-trained AI models to generate their own unique AI models, and also use the emotion engine to recommend parameters and customize the interface.

[0703] (Application example 2)

[0704] Next, a description will be given of Application 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."

[0705] In recent years, advances in AI technology have created a demand for combining and using multiple trained models. However, there is no efficient and intuitive way to combine these models. Furthermore, few systems take the user's emotional state into account, and parameter settings and interface customization are cumbersome. There is a need for a system that can solve these issues and provide a better user experience based on the user's emotions.

[0706] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for recommending parameters based on the user's emotional state, and means for customizing the interface according to the user's emotional state. This makes it possible to efficiently and intuitively combine multiple trained models while taking the user's emotional state into consideration.

[0707] A "trained model" is a machine learning algorithm that has been trained in advance using large amounts of data and has the ability to perform a specific task.

[0708] A "database" is a system that organizes and stores data so that it can be accessed and manipulated as needed.

[0709] "Storage" refers to a storage device for long-term storage of data.

[0710] A "graphical user interface" refers to an interface that a user can visually interact with, typically including graphical elements such as windows, buttons, sliders, etc.

[0711] "Parameters" refer to the settings that control the performance and behavior of a model.

[0712] "Intermediate output" refers to the intermediate results obtained when combining multiple trained models.

[0713] "Real-time" means that operations or processes occur immediately, without delay.

[0714] The "final model" refers to the machine learning model that is ultimately created based on the output results desired by the user.

[0715] "Emotional state" refers to a user's current emotional or psychological state.

[0716] "Parameter recommendation" refers to the act of the system suggesting appropriate setting values ​​for parameters set by the user.

[0717] "Interface customization" refers to changing the design and layout of an interface in response to a user's specific requirements or emotional state.

[0718] This invention is a system that efficiently and intuitively combines multiple trained models, recommends parameters taking into account the user's emotional state, and customizes the interface. It is composed of a server, a terminal, and a user. To realize this system, the following processes are performed.

[0719] server

[0720] The server has the function of loading multiple trained models from a database or storage. It loads the model data selected by the user and sends it to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user. It also combines models based on the user's parameter adjustment requests, generates intermediate output, and sends it back to the terminal.

[0721] Terminal

[0722] The device analyzes the model data sent from the server and provides the user with a graphical user interface, which includes sliders and dials similar to a music equalizer to adjust the parameters of each model.The device also includes an emotion engine that analyzes the user's facial expressions, voice, and typing patterns to recognize their emotional state.

[0723] User

[0724] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, which then adjusts and combines the models based on the parameters to generate an intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The emotion engine analyzes the user's emotional state and recommends optimal parameters based on that state. The interface layout and design can also be customized based on the emotional state. When the user is satisfied with the output results, they press the "Save" button to generate the final model. The device sends this request to the server, which generates the final combination results and stores them in a database or storage.

[0725] Specific examples

[0726] For example, consider the use of this technology in a "personalized video viewing application." This application recommends the most suitable videos based on the user's emotional state, improving the viewing experience. When a user requests videos of a specific genre or theme, the emotion engine analyzes the user's emotions and generates a list of recommended videos based on that. The interface also changes dynamically depending on the user's emotional state. Specifically, the following prompts are input to the generative AI model:

[0727] Example prompt sentence:

[0728] Read the user's facial expressions and recommend the following list of movies and documentaries.

[0729] 1. Inspirational Movies

[0730] 2. An interesting documentary

[0731] This system can understand the user's emotional state and provide viewing content that is more optimal for each individual.

[0732] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0733] Step 1:

[0734] The server reads multiple trained models selected by the user from the database. It receives the user's selection, retrieves the model data from the database, and sends it to the device. The input is the user's model selection information, and the output is the model data.

[0735] Step 2:

[0736] The terminal analyzes the model data sent from the server, converts the acquired model data into an appropriate format, and makes it operable by the user. The input is the model data from the server, and the output is the model data converted into a format that can be operated by the user.

[0737] Step 3:

[0738] The terminal provides a graphical user interface, which includes sliders and dials that the user can manipulate directly to adjust the parameters of each model. The input is the converted model data, and the output is a parameter setting screen that can be manipulated in the interface.

[0739] Step 4:

[0740] The user adjusts the parameters of each model using an interface. The emotion engine analyzes the user's facial expressions, voice, and typing patterns to recognize the user's emotional state at that time. The input is the user's actions and emotional data, and the output is the adjusted parameters.

[0741] Step 5:

[0742] The server receives the parameters adjusted by the user, combines multiple trained models based on them, and generates intermediate outputs. The inputs are the parameters adjusted by the user, and the outputs are the intermediate outputs.

[0743] Step 6:

[0744] The terminal displays the intermediate output sent from the server in real time. The interface and intermediate output are updated every time the user changes a parameter. The input is the intermediate output from the server, and the output is the updated interface and display of the intermediate output.

[0745] Step 7:

[0746] The emotion engine recommends optimal parameters based on the user's current emotional state. As the user adjusts parameters, the emotion engine analyzes them in real time and suggests appropriate settings. The input is the user's emotional data, and the output is the recommended parameter settings.

[0747] Step 8:

[0748] The device customizes the interface based on the analysis results of the emotion engine. The color and layout of the interface are changed according to the emotional state. The input is the analyzed emotional data, and the output is a customized interface.

[0749] Step 9:

[0750] When the user is satisfied with the output results, they can click the "Save" button to generate and save the final model. The input is the user's save request, and the output is the generated final model and its saving.

[0751] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[0752] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[0753] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0754] [Third embodiment]

[0755] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0756] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0757] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0758] 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.

[0759] 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 audio in accordance with instructions from the processor 46.

[0760] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0761] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0762] Fig. 6 shows an example of the 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.

[0763] 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.

[0764] 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.

[0765] In the headset type terminal 314, a 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.

[0766] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0767] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models. This system consists of a server, a terminal, and a user, and allows the user to adjust the parameters of each model using a graphical user interface and generate the final model while checking the results in real time.

[0768] The server is responsible for loading multiple trained models selected by the user from a database or storage, and then sending the loaded model data to the device, making it operable on the device.

[0769] The device analyzes the model data received from the server and displays an intuitive graphical interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model.

[0770] The user adjusts the parameters of each model using the interface on the device. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0771] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. First, the user selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0772] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output result to the user, who can then confirm it.

[0773] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device sends the request to the server, which generates the final combination result and saves it in a database or storage. The final model can then be downloaded by the user or integrated into other systems for use.

[0774] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0775] The processing flow will be explained below.

[0776] Step 1:

[0777] A user logs in to the system and accesses a screen that displays a list of trained models.

[0778] Step 2:

[0779] The terminal sends the user's login information to the server, which then performs authentication.

[0780] Step 3:

[0781] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0782] Step 4:

[0783] The device displays a list of trained models received to the user.

[0784] Step 5:

[0785] The user selects the trained model they want to use and sends that information to their device.

[0786] Step 6:

[0787] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0788] Step 7:

[0789] The server sends the trained model it has loaded to the terminal.

[0790] Step 8:

[0791] The terminal analyzes the received trained model and generates a graphical user interface.

[0792] Step 9:

[0793] The user adjusts the parameters of each trained model using a graphical user interface.

[0794] Step 10:

[0795] The device receives the user's adjustments and transmits them to the server in real time.

[0796] Step 11:

[0797] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0798] Step 12:

[0799] The server generates intermediate output and sends it back to the terminal.

[0800] Step 13:

[0801] The intermediate output received by the terminal is displayed to the user in real time.

[0802] Step 14:

[0803] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0804] Step 15:

[0805] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0806] Step 16:

[0807] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0808] Step 17:

[0809] The terminal sends a "save" request to the server, which generates the final model.

[0810] Step 18:

[0811] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0812] Step 19:

[0813] The device will notify the user that the final model has been saved and provide a download link.

[0814] Step 20:

[0815] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0816] Example 1

[0817] Next, a description will be given of 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."

[0818] Modern machine learning models are highly complex, making it difficult for users to efficiently combine multiple pre-trained models. It is particularly difficult for users without specialized knowledge to intuitively operate the system and adjust models while checking results in real time. Furthermore, there is a lack of efficient ways to adjust and integrate models with different performance metrics, making model optimization time-consuming and labor-intensive.

[0819] 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.

[0820] In this invention, the server includes a means for loading a plurality of trained models from a database or a storage device, a means for adjusting parameters of the plurality of trained models via a graphical user interface, a means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output, and a means for generating and saving a final model. This allows a user, without specialized knowledge, to intuitively and efficiently adjust the plurality of models and generate an optimal model while checking the results in real time.

[0821] "Pre-trained models" are a collection of machine learning algorithms that have been pre-trained using an existing dataset.

[0822] A "database" is a system or device for systematically storing and managing data.

[0823] A "storage device" is hardware or media for storing data and programs.

[0824] A "means" is a method, device, or process for achieving a particular end.

[0825] A "graphical user interface" is a software interface that allows a user to visually interact with the software.

[0826] "Parameters" are setting values ​​that control the operation and behavior of a model.

[0827] "Tuning" means changing settings or configurations to suit a particular purpose.

[0828] "Intermediate outputs" are temporary results produced when combining multiple trained models.

[0829] "Immediate display" means displaying information or data on the screen the moment it is received.

[0830] The "final model" is the final machine learning model that is generated after all tuning and combination is complete.

[0831] "Preservation" refers to the permanent or long-term storage of generated data and models.

[0832] "User" means a person or entity that uses the system to select, adjust, and generate models.

[0833] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained models. This system is composed of a server, a terminal, and a user. The details of the system are described below.

[0834] The server first loads multiple trained models from a database or storage device. For example, the server accesses a database to obtain pre-trained data such as image recognition models and speech recognition models. The server analyzes this model data and sends it to the terminal in an appropriate format.

[0835] The terminal receives and analyzes the model data received from the server. The terminal then displays a graphical user interface that the user can operate intuitively. This interface includes sliders and dials similar to a music equalizer, allowing the user to adjust the parameters of each model. The user can use this to make adjustments while checking the model's behavior in real time.

[0836] The user adjusts the parameters of each model using the interface on the device. Each time this operation is performed, the device sends the adjustment results to the server in real time. The server adjusts the models based on the received parameters and combines multiple models to generate intermediate outputs. These intermediate outputs are then sent back to the device and displayed to the user in real time.

[0837] As a concrete example, consider the case of combining an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[0838] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output to the user, who can then confirm it.

[0839] Examples of prompts include:

[0840] "Select an image recognition model and a voice recognition model, and set the image recognition accuracy to 80% and the voice recognition response speed to 90%."

[0841] "Generate intermediate output based on the parameters you set and display the results."

[0842] "Please explain how to hit the save button to generate the final model and save it to the database."

[0843] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device then sends the request to the server, which then generates and saves the final combination result. This final model can then be downloaded by the user or integrated into other systems for use.

[0844] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[0845] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0846] Step 1:

[0847] The user accesses the device interface. The user accesses the system's graphical user interface using a PC or mobile device. Input can be a URL or an application launch operation. This launches the interface and displays the model selection screen.

[0848] Step 2:

[0849] The server sends a list of trained models to the device. The server accesses the database and obtains information about the stored trained models (e.g., image recognition models, speech recognition models). The input is a database query. The output is a list of models sent to the device.

[0850] Step 3:

[0851] The user selects the model they want to use from the interface. They select the model they want to use from multiple pre-trained models on their device screen, using checkboxes or drop-down menus. The input is the user's selection operation. The output is the information about the selected model sent to the server.

[0852] Step 4:

[0853] The device requests the user's selection from the server. The device requests the ID and name of the model selected by the user from the server. The input is the model information selected by the user. The output is that the request is sent to the server.

[0854] Step 5:

[0855] The server retrieves the selected model data from the database and sends it to the terminal. The server accesses the database and retrieves the data of the selected model. The input is the selected model information. The output is the corresponding model data sent to the terminal.

[0856] Step 6:

[0857] The terminal displays the model on a graphical user interface. The terminal interprets the model data it receives and displays it in the interface. As input, it has the model data. As output, it presents adjustment tools such as sliders and dials to the user.

[0858] Step 7:

[0859] The user adjusts the model parameters. The user adjusts the parameters of each model by operating sliders and dials. The input is the user's operation information. The output is the adjusted parameters sent to the terminal, which then sends them to the server.

[0860] Step 8:

[0861] The terminal sends the adjusted parameters to the server. The terminal immediately sends the parameters adjusted by the user to the server. The input is the adjusted parameter information. The output is the parameter information sent to the server.

[0862] Step 9:

[0863] The server adjusts the models based on the received parameters and generates intermediate outputs. The server adjusts the behavior of each model based on the adjusted parameters and combines them to generate intermediate outputs. The input is the adjusted parameter information. The output is intermediate output data.

[0864] Step 10:

[0865] The server generates intermediate output and sends it to the terminal. The server generates intermediate output and sends it to the terminal. As input, there is intermediate output data. As output, that data is sent to the terminal.

[0866] Step 11:

[0867] The terminal displays the intermediate output to the user in real time. The terminal displays the intermediate output received from the server to the user in real time. As input, there is intermediate output data. As output, there is intermediate output displayed on the interface.

[0868] Step 12:

[0869] The user fine-tunes the parameters and determines the final model. The user checks the intermediate outputs and fine-tunes the parameters as needed to determine a satisfactory final model. The inputs are the intermediate outputs and user operation information. The output is the final parameter settings.

[0870] Step 13:

[0871] The terminal sends the final parameters to the server. The terminal sends the final parameters determined by the user to the server. As input, there are the final parameter settings. As output, that information is sent to the server.

[0872] Step 14:

[0873] The server generates the final model and stores it in a database or storage device. The server generates the final model based on the final parameters and stores it in a database or storage device. The input is the final parameter settings. The output is the final model data and stores it in a database or storage device.

[0874] Step 15:

[0875] The user downloads the final model or integrates it into another system for use. The user downloads the generated final model via a terminal or integrates it with another system. The inputs are the final model data and the user's download operation. The outputs are the model data stored in the user's environment and the integrated system.

[0876] (Application example 1)

[0877] Next, a description will be given of Application 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."

[0878] In conventional factory robot systems using AI models, optimizing each task required a lot of time and specialized knowledge. Furthermore, there was a lack of means to intuitively combine and use different types of trained models, making it difficult for users to adjust parameters and check the results in real time. To solve this problem, there is a need for a system that efficiently and intuitively combines multiple trained models to optimize the operation of factory robots.

[0879] 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.

[0880] In this invention, the server includes means for loading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for transmitting and displaying the intermediate output to a terminal in the factory in real time, and means for generating and saving a final model. This enables users to intuitively use multiple trained models to optimize the operation of factory robots and effectively utilize complex AI systems without specialized knowledge.

[0881] A "trained model" is an artificial intelligence algorithm that has been pre-trained for a specific task.

[0882] A "database" is an information system that can efficiently manage large amounts of data and search and update it.

[0883] "Storage" is a physical and logical memory device for storing data.

[0884] A "graphical user interface" is an interface that uses visual elements to allow users to operate it intuitively.

[0885] A "parameter" is a setting or variable that affects the behavior or performance of a model.

[0886] "Intermediate output" is a temporary result generated after combining and adjusting multiple trained models.

[0887] "Real time" refers to an operation that responds immediately to a user's operation and displays the result instantly.

[0888] A "factory robot" is an automated mechanical device that performs physical tasks in a factory.

[0889] A "specific task" refers to a specific task or activity that a robot performs within a factory.

[0890] The "final model" is the final AI model that the user has optimized through parameter adjustment and decided to save.

[0891] A "terminal" is a digital device that a user uses to perform operations and check results.

[0892] This invention is a system for generating custom AI models to optimize the operation of robots used in factories. This system generates a unique learning model by efficiently and intuitively combining multiple trained models, and is composed of a server, a terminal, and a user.

[0893] The server runs on AWS EC2 and is configured with a web framework using Django. First, the server loads multiple selected trained models from a database (AWS RDS, PostgreSQL). The loaded model data is sent to the terminal, where it can be operated on the terminal.

[0894] The terminals are iOS / Android devices (smartphones and tablets) that display an intuitive graphical user interface using React Native, which includes sliders and dials similar to a music equalizer, allowing users to adjust the parameters of each model.

[0895] The user adjusts the parameters of each model using the device's interface. For example, they can select an object recognition model and a movement pattern model and set the object recognition accuracy to 90% and the movement reaction speed to 0.5 seconds. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[0896] Once the user has further fine-tuned the parameters and is satisfied with the output results, they press the save button to generate the final model. The device then sends the request to the server, which generates the final combination results and stores them in a database or storage. The final model is then downloaded to the robots in the factory and used to perform specific tasks.

[0897] For example, a prompt to a generative AI model might look something like this:

[0898] You need to generate an AI model to deploy optimal robot behavior for specific tasks in your factory. Adjust the AI ​​model to meet the following requirements:

[0899] Object recognition accuracy: 90% or more

[0900] Response time: within 0.5 seconds

[0901] Efficiency of movement patterns: Minimal route and obstacle avoidance

[0902] Use the sliders to adjust these parameters and see the results in real time.

[0903] In this way, users can easily combine multiple pre-trained models to generate custom AI models suited to tasks within their factories, enabling them to effectively utilize complex AI systems without specialized knowledge.

[0904] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0905] Step 1:

[0906] A user accesses the system using a terminal and requests a list of trained models for a specific task of a factory robot. The terminal sends the request to the server. The input to this request is the name of the model selected by the user.

[0907] Step 2:

[0908] The server receives the request, searches for the corresponding trained model in a database (e.g., PostgreSQL on AWS RDS), and retrieves the model data. In this data processing stage, a database query is executed based on the model name to load the appropriate model data. The loaded model data is obtained as output.

[0909] Step 3:

[0910] The server transmits the acquired model data to the terminal, which analyzes the received model data and displays an intuitive graphical user interface that includes sliders and dials that respond in real time.

[0911] Step 4:

[0912] The user adjusts the parameters of each model using the device interface. For example, the accuracy of the object recognition model is set to 90% and the response speed of the movement pattern model is set to 0.5 seconds. These adjustments are sent as data from the device to the server.

[0913] Step 5:

[0914] The server adjusts the behavior of each model based on the received parameters, combines multiple models, and generates intermediate output. Based on the parameter input, it performs data calculations to optimize the internal parameters of the model, and generates an adjusted intermediate model. This intermediate output is obtained as the output.

[0915] Step 6:

[0916] The intermediate output is sent from the server to the terminal, which displays it in real time to the user, who can then check the displayed results and further fine-tune the parameters if necessary. This process can be repeated.

[0917] Step 7:

[0918] When the user is satisfied with the output result, he / she presses the save button to generate the final model. The terminal sends this operation to the server, which then generates the final model. At this stage, the final result of the parameter settings that the user is satisfied with is obtained as input.

[0919] Step 8:

[0920] The server generates the final combination results and saves them in a database or storage. The final model is then downloaded and applied to the robots in the factory, resulting in optimal robot behavior for the specific tasks in the factory.

[0921] Furthermore, 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.

[0922] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0923] The server has the function of loading multiple trained models selected by the user from a database or storage and sending them to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user.

[0924] The device analyzes the model data sent from the server and provides the user with an interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model. The device also includes an emotion engine that recognizes emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0925] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, and the server adjusts the model based on the adjusted parameters and combines them to generate intermediate output. The generated intermediate output is then sent back to the device and displayed to the user in real time.

[0926] The emotion engine analyzes facial expressions, voice, and typing patterns when a user operates the interface to recognize the user's emotional state. Based on this recognized emotion, the emotion engine recommends parameters appropriate for the user. It is also possible to automatically customize the interface according to the user's emotions.

[0927] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the device interface. The device then requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[0928] Next, the user uses the sliders on the interface to set "image recognition accuracy" to 80% and "voice recognition response speed" to 90%. At this point, the emotion engine recognizes the user's emotions and recommends parameter settings that are easier to use if the user is nervous, for example. Alternatively, the interface layout and design can be changed depending on the user's emotions.

[0929] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user, who can check the results and further fine-tune the parameters.

[0930] When the user finally obtains the desired result, they press the "Save" button to generate the final model. The terminal sends the request to the server, which generates the final combination result and saves it in a database or storage. The user can download this final model or integrate it into other systems for use.

[0931] In this way, the present invention makes it possible to effectively combine multiple trained models while taking into account the user's emotional state, thereby intuitively and efficiently generating a unique AI model.

[0932] The processing flow will be explained below.

[0933] Step 1:

[0934] A user logs in to the system and accesses a screen that displays a list of trained models.

[0935] Step 2:

[0936] The terminal sends the user's login information to the server, which then performs authentication.

[0937] Step 3:

[0938] The server retrieves a list of available trained models from the database and sends it to the terminal.

[0939] Step 4:

[0940] The device displays a list of trained models received to the user.

[0941] Step 5:

[0942] The user selects the trained model they want to use and sends that information to their device.

[0943] Step 6:

[0944] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[0945] Step 7:

[0946] The server sends the trained model it has loaded to the terminal.

[0947] Step 8:

[0948] The device analyzes the received trained model and generates a graphical user interface, which includes sliders and dials similar to a music equalizer.

[0949] Step 9:

[0950] The user adjusts the parameters of each trained model using a graphical user interface.

[0951] Step 10:

[0952] The emotion engine analyzes the user's facial expressions, voice, and typing patterns in real time while they are operating the device, recognizing their emotional state.

[0953] Step 11:

[0954] The device recommends appropriate parameter settings and customizes the interface based on the user's emotions recognized by the emotion engine.

[0955] Step 12:

[0956] The device receives the user's adjustments and transmits them to the server in real time.

[0957] Step 13:

[0958] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[0959] Step 14:

[0960] The server generates intermediate output and sends it back to the terminal.

[0961] Step 15:

[0962] The intermediate output received by the terminal is displayed to the user in real time.

[0963] Step 16:

[0964] The user checks the intermediate output and makes further fine adjustments to the parameters.

[0965] Step 17:

[0966] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[0967] Step 18:

[0968] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[0969] Step 19:

[0970] The terminal sends a "save" request to the server, which generates the final model.

[0971] Step 20:

[0972] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[0973] Step 21:

[0974] The device will notify the user that the final model has been saved and provide a download link.

[0975] Step 22:

[0976] Users can download the final model and integrate it into other systems or use it on their own platforms.

[0977] Example 2

[0978] Next, a description will be given of 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."

[0979] Conventional AI model generation systems have difficulty efficiently and intuitively combining multiple trained models, placing a heavy burden on users. Furthermore, they are unable to adjust parameters or customize interfaces that take the user's emotional state into account, resulting in a lack of improvement in the user experience.

[0980] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating an intermediate output, means for incorporating an emotion engine that recognizes the user's emotional state and recommends parameters, means for automatically customizing the interface according to the user's emotional state, and means for generating and saving a final model. This makes it possible to intuitively and efficiently combine multiple trained models while taking the user's emotional state into consideration to generate a unique AI model.

[0981] A "pre-trained model" is an artificial intelligence model that has already been trained and optimized to perform a specific task.

[0982] A "database" is an electronic system for efficiently storing and retrieving data.

[0983] "Storage" is a physical or virtual space for storing digital data.

[0984] A "graphical user interface" is an interface that includes elements (such as buttons, sliders, and dials) that a user can visually manipulate.

[0985] A "parameter" is a setting value or variable that controls the behavior of an AI model.

[0986] An "emotion engine" is an artificial intelligence engine that identifies a user's emotions and provides appropriate responses and recommendations.

[0987] "Intermediate output" refers to an intermediate result generated before obtaining the final output.

[0988] The "final model" is the AI ​​model that is completed as a result of adjustments and combinations.

[0989] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[0990] Hardware and Software Configuration

[0991] server

[0992] The server functions as a means of loading multiple trained models from a database or storage (for example, MySQL or PostgreSQL is used as database software), sending the model data selected by the user to the terminal, and reconstructing the model based on the adjusted parameters.

[0993] Terminal

[0994] The terminal is a means for analyzing the model data sent from the server. The terminal provides a user-operable graphical user interface (GUI). This GUI includes sliders and dials that the user uses to adjust the model parameters. The terminal also includes an emotion engine that uses OpenCV and TensorFlow to recognize emotions by analyzing the user's facial expressions, voice, and typing patterns.

[0995] Emotion Engine

[0996] The emotion engine is a means of recognizing the user's emotions. It is possible to use PyTorch or Keras as a deep learning framework. This emotion engine has the function of recommending appropriate parameters for the user based on the analyzed emotional state and automatically customizing the interface.

[0997] Specific examples

[0998] For example, if a user wants to combine an image recognition model and a voice recognition model, they can use the system by following the steps below.

[0999] 1. The user selects the "image recognition model" and "voice recognition model" from the device interface.

[1000] 2. The device requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[1001] 3. The user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%.

[1002] 4. The emotion engine analyzes the user's facial expressions and voice and recommends easier parameter settings if the user is nervous. It can also change the layout and design of the interface according to the user's emotions.

[1003] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The user can check the results and further fine-tune the parameters.

[1004] Example of input prompt for generative AI model

[1005] 1. User prompt: "I want to combine image and speech recognition models, with image recognition accuracy set to 80% and speech recognition response speed set to 90%."

[1006] 2. Prompt for the generative AI model: "Please combine the image recognition and voice recognition models under the following conditions: Image recognition accuracy: 80%, Voice recognition response speed: 90%. The emotion engine will analyze the user's emotions and recommend suitable parameters."

[1007] This system allows users to generate their own unique AI models by efficiently combining multiple trained models and intuitively operating them while taking into account their own emotional state.

[1008] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1009] Step 1:

[1010] Initialization and model selection

[1011] input:

[1012] A user starts up a device and sends a request to display a list of available trained models.

[1013] output:

[1014] A list of available trained models will be displayed on your device.

[1015] Specific behavior:

[1016] The device initializes the system and requests a list of trained models from the server. The server retrieves the model list from a database (e.g., MySQL or PostgreSQL) and sends it to the device. The device parses it and displays it to the user.

[1017] Step 2:

[1018] Model selection and data loading

[1019] input:

[1020] The user selects the "image recognition model" and "voice recognition model" on the device interface.

[1021] output:

[1022] The selected model data is sent from the server to the terminal.

[1023] Specific behavior:

[1024] Based on the user's operation, the device sends a request for the selected model (e.g., "image recognition model" and "voice recognition model") to the server. The server reads the corresponding model data from the database and sends it to the device.

[1025] Step 3:

[1026] Interface display and parameter adjustment

[1027] input:

[1028] The terminal receives the model data sent from the server.

[1029] output:

[1030] An interface is displayed that allows the user to adjust the parameters.

[1031] Specific behavior:

[1032] The device analyzes the received model data and provides the user with a graphical user interface (GUI) for adjusting the model parameters. It displays UI elements such as sliders and dials that the user can use to adjust the parameters.

[1033] Step 4:

[1034] Parameter submission and model adjustment

[1035] input:

[1036] The user sets parameters on the interface (e.g., "image recognition accuracy" at 80%, "voice recognition response speed" at 90%).

[1037] output:

[1038] The adjusted parameters are sent to the server.

[1039] Specific behavior:

[1040] The parameters set by the user are sent in real time from the device to the server, which then adjusts and combines the selected model based on the received parameters.

[1041] Step 5:

[1042] Generating and displaying intermediate output

[1043] input:

[1044] The server receives the adjusted parameters.

[1045] output:

[1046] An intermediate output is generated and sent to the terminal.

[1047] Specific behavior:

[1048] The server adjusts the AI ​​model based on the adjusted parameters and generates intermediate output, which is sent to the device in JSON format, etc. The device analyzes the received intermediate output and displays it to the user in real time.

[1049] Step 6:

[1050] Analysis and recommendation by emotion engine

[1051] input:

[1052] Emotional data such as the user's facial expression, voice, and typing pattern are input.

[1053] output:

[1054] Emotion-based parameter recommendations and interface customization are performed.

[1055] Specific behavior:

[1056] The device's built-in emotion engine (using OpenCV and TensorFlow) analyzes the user's facial expressions and voice to recognize their emotional state at that time. Based on the recognition results, the emotion engine recommends parameter settings appropriate for the user and customizes the interface.

[1057] Step 7:

[1058] Save and download the final model

[1059] input:

[1060] When the user obtains a satisfactory output result, he or she presses the "save" button.

[1061] output:

[1062] The final model is saved to a database or storage and a download link is provided.

[1063] Specific behavior:

[1064] When the user presses the "Save" button, the device sends a request to the server to save the final model. The server generates the final model and saves it in a database or storage. The server then returns a download link for the final model to the device, and the user clicks the link to download the final model.

[1065] These processing steps enable users to efficiently and intuitively combine multiple pre-trained AI models to generate their own unique AI models, and also use the emotion engine to recommend parameters and customize the interface.

[1066] (Application example 2)

[1067] Next, a description will be given of Application 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."

[1068] In recent years, advances in AI technology have created a demand for combining and using multiple trained models. However, there is no efficient and intuitive way to combine these models. Furthermore, few systems take the user's emotional state into account, and parameter settings and interface customization are cumbersome. There is a need for a system that can solve these issues and provide a better user experience based on the user's emotions.

[1069] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for recommending parameters based on the user's emotional state, and means for customizing the interface according to the user's emotional state. This makes it possible to efficiently and intuitively combine multiple trained models while taking the user's emotional state into consideration.

[1070] A "trained model" is a machine learning algorithm that has been trained in advance using a large amount of data and has the ability to perform a specific task.

[1071] A "database" is a system that organizes and stores data so that it can be accessed and manipulated as needed.

[1072] "Storage" refers to a storage device for long-term storage of data.

[1073] A "graphical user interface" refers to an interface that a user can visually interact with, typically including graphical elements such as windows, buttons, sliders, etc.

[1074] "Parameters" refer to the settings that control the performance and behavior of a model.

[1075] "Intermediate output" refers to the intermediate results obtained when combining multiple trained models.

[1076] "Real-time" means that operations or processes occur immediately, without delay.

[1077] The "final model" refers to the machine learning model that is ultimately created based on the output results desired by the user.

[1078] "Emotional state" refers to a user's current emotional or psychological state.

[1079] "Parameter recommendation" refers to the act of the system suggesting appropriate setting values ​​for parameters set by the user.

[1080] "Interface customization" refers to changing the design and layout of an interface in response to a user's specific requirements or emotional state.

[1081] This invention is a system that efficiently and intuitively combines multiple trained models, recommends parameters taking into account the user's emotional state, and customizes the interface. It is composed of a server, a terminal, and a user. To realize this system, the following processes are performed.

[1082] server

[1083] The server has the function of loading multiple trained models from a database or storage. It loads the model data selected by the user and sends it to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user. It also combines models based on the user's parameter adjustment requests, generates intermediate output, and sends it back to the terminal.

[1084] Terminal

[1085] The device analyzes the model data sent from the server and provides the user with a graphical user interface, which includes sliders and dials similar to a music equalizer to adjust the parameters of each model.The device also includes an emotion engine that analyzes the user's facial expressions, voice, and typing patterns to recognize their emotional state.

[1086] User

[1087] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, which then adjusts and combines the models based on the parameters to generate an intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The emotion engine analyzes the user's emotional state and recommends optimal parameters based on that state. The interface layout and design can also be customized based on the emotional state. When the user is satisfied with the output results, they press the "Save" button to generate the final model. The device sends this request to the server, which generates the final combination results and stores them in a database or storage.

[1088] Specific examples

[1089] For example, consider the use of this technology in a "personalized video viewing application." This application recommends the most suitable videos based on the user's emotional state, improving the viewing experience. When a user requests videos of a specific genre or theme, the emotion engine analyzes the user's emotions and generates a list of recommended videos based on that. The interface also changes dynamically depending on the user's emotional state. Specifically, the following prompts are input to the generative AI model:

[1090] Example prompt sentence:

[1091] Read the user's facial expressions and recommend the following list of movies and documentaries.

[1092] 1. Inspirational Movies

[1093] 2. An interesting documentary

[1094] This system can understand the user's emotional state and provide viewing content that is more optimal for each individual.

[1095] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1096] Step 1:

[1097] The server reads multiple trained models selected by the user from the database. It receives the user's selection, retrieves the model data from the database, and sends it to the device. The input is the user's model selection information, and the output is the model data.

[1098] Step 2:

[1099] The terminal analyzes the model data sent from the server, converts the acquired model data into an appropriate format, and makes it operable by the user. The input is the model data from the server, and the output is the model data converted into a format that can be operated by the user.

[1100] Step 3:

[1101] The terminal provides a graphical user interface, which includes sliders and dials that the user can manipulate directly to adjust the parameters of each model. The input is the converted model data, and the output is a parameter setting screen that can be manipulated in the interface.

[1102] Step 4:

[1103] The user adjusts the parameters of each model using an interface. The emotion engine analyzes the user's facial expressions, voice, and typing patterns to recognize the user's emotional state at that time. The input is the user's actions and emotional data, and the output is the adjusted parameters.

[1104] Step 5:

[1105] The server receives the parameters adjusted by the user, combines multiple trained models based on them, and generates intermediate outputs. The inputs are the parameters adjusted by the user, and the outputs are the intermediate outputs.

[1106] Step 6:

[1107] The terminal displays the intermediate output sent from the server in real time. The interface and intermediate output are updated every time the user changes a parameter. The input is the intermediate output from the server, and the output is the updated interface and display of the intermediate output.

[1108] Step 7:

[1109] The emotion engine recommends optimal parameters based on the user's current emotional state. As the user adjusts parameters, the emotion engine analyzes them in real time and suggests appropriate settings. The input is the user's emotional data, and the output is the recommended parameter settings.

[1110] Step 8:

[1111] The device customizes the interface based on the analysis results of the emotion engine. The color and layout of the interface are changed according to the emotional state. The input is the analyzed emotional data, and the output is a customized interface.

[1112] Step 9:

[1113] When the user is satisfied with the output results, they can click the "Save" button to generate and save the final model. The input is the user's save request, and the output is the generated final model and its saving.

[1114] 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 audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio 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 audio data.

[1115] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[1116] 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.

[1117] [Fourth embodiment]

[1118] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1119] 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.

[1120] 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 WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1121] 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. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1122] 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 audio in accordance with instructions from the processor 46.

[1123] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and 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 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1124] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1125] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. 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.

[1126] Fig. 8 shows an example of the 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.

[1127] 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.

[1128] 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.

[1129] In the robot 414, the processor 46 performs the reception output process. 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.

[1130] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. 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."

[1131] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models. This system consists of a server, a terminal, and a user, and allows the user to adjust the parameters of each model using a graphical user interface and generate the final model while checking the results in real time.

[1132] The server is responsible for loading multiple trained models selected by the user from a database or storage, and then sending the loaded model data to the device, making it operable on the device.

[1133] The device analyzes the model data received from the server and displays an intuitive graphical interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model.

[1134] The user adjusts the parameters of each model using the interface on the device. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[1135] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. First, the user selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[1136] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output result to the user, who can then confirm it.

[1137] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device sends the request to the server, which generates the final combination result and saves it in a database or storage. The final model can then be downloaded by the user or integrated into other systems for use.

[1138] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[1139] The processing flow will be explained below.

[1140] Step 1:

[1141] A user logs in to the system and accesses a screen that displays a list of trained models.

[1142] Step 2:

[1143] The terminal sends the user's login information to the server, which then performs authentication.

[1144] Step 3:

[1145] The server retrieves a list of available trained models from the database and sends it to the terminal.

[1146] Step 4:

[1147] The device displays a list of trained models received to the user.

[1148] Step 5:

[1149] The user selects the trained model they want to use and sends that information to their device.

[1150] Step 6:

[1151] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[1152] Step 7:

[1153] The server sends the trained model it has loaded to the terminal.

[1154] Step 8:

[1155] The terminal analyzes the received trained model and generates a graphical user interface.

[1156] Step 9:

[1157] The user adjusts the parameters of each trained model using a graphical user interface.

[1158] Step 10:

[1159] The device receives the user's adjustments and transmits them to the server in real time.

[1160] Step 11:

[1161] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[1162] Step 12:

[1163] The server generates intermediate output and sends it back to the terminal.

[1164] Step 13:

[1165] The intermediate output received by the terminal is displayed to the user in real time.

[1166] Step 14:

[1167] The user checks the intermediate output and makes further fine adjustments to the parameters.

[1168] Step 15:

[1169] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[1170] Step 16:

[1171] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[1172] Step 17:

[1173] The terminal sends a "save" request to the server, which generates the final model.

[1174] Step 18:

[1175] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[1176] Step 19:

[1177] The device will notify the user that the final model has been saved and provide a download link.

[1178] Step 20:

[1179] Users can download the final model and integrate it into other systems or use it on their own platforms.

[1180] Example 1

[1181] Next, a description will be given of 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."

[1182] Modern machine learning models are highly complex, making it difficult for users to efficiently combine multiple pre-trained models. It is particularly difficult for users without specialized knowledge to intuitively operate the system and adjust models while checking results in real time. Furthermore, there is a lack of efficient ways to adjust and integrate models with different performance metrics, making model optimization time-consuming and labor-intensive.

[1183] 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.

[1184] In this invention, the server includes a means for loading a plurality of trained models from a database or a storage device, a means for adjusting parameters of the plurality of trained models via a graphical user interface, a means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output, and a means for generating and saving a final model. This allows a user, without specialized knowledge, to intuitively and efficiently adjust the plurality of models and generate an optimal model while checking the results in real time.

[1185] "Pre-trained models" are a collection of machine learning algorithms that have been pre-trained using an existing dataset.

[1186] A "database" is a system or device for systematically storing and managing data.

[1187] A "storage device" is hardware or media for storing data and programs.

[1188] A "means" is a method, device, or process for achieving a particular end.

[1189] A "graphical user interface" is a software interface that allows a user to visually interact with the software.

[1190] "Parameters" are setting values ​​that control the operation and behavior of a model.

[1191] "Tuning" means changing settings or configurations to suit a particular purpose.

[1192] "Intermediate outputs" are temporary results produced when combining multiple trained models.

[1193] "Immediate display" means displaying information or data on the screen the moment it is received.

[1194] The "final model" is the final machine learning model that is generated after all tuning and combination is complete.

[1195] "Preservation" refers to the permanent or long-term storage of generated data and models.

[1196] "User" means a person or entity that uses the system to select, adjust, and generate models.

[1197] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained models. This system is composed of a server, a terminal, and a user. The details of the system are described below.

[1198] The server first loads multiple trained models from a database or storage device. For example, the server accesses a database to obtain pre-trained data such as image recognition models and speech recognition models. The server analyzes this model data and sends it to the terminal in an appropriate format.

[1199] The terminal receives and analyzes the model data received from the server. The terminal then displays a graphical user interface that the user can operate intuitively. This interface includes sliders and dials similar to a music equalizer, allowing the user to adjust the parameters of each model. The user can use this to make adjustments while checking the model's behavior in real time.

[1200] The user adjusts the parameters of each model using the interface on the device. Each time this operation is performed, the device sends the adjustment results to the server in real time. The server adjusts the models based on the received parameters and combines multiple models to generate intermediate outputs. These intermediate outputs are then sent back to the device and displayed to the user in real time.

[1201] As a concrete example, consider the case of combining an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the interface on the device. The device then requests the model data from the server, which then reads it from the database and sends it to the device.

[1202] Next, the user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%. The device sends these parameters to the server, which adjusts the model based on the parameters, combines them to generate intermediate output, and sends it back to the device. The device immediately displays the new output to the user, who can then confirm it.

[1203] Examples of prompts include:

[1204] "Select an image recognition model and a voice recognition model, and set the image recognition accuracy to 80% and the voice recognition response speed to 90%."

[1205] "Generate intermediate output based on the parameters you set and display the results."

[1206] "Please explain how to hit the save button to generate the final model and save it to the database."

[1207] If the user further fine-tunes the parameters and obtains a satisfactory output result, they press the "Save" button to generate the final model. The device then sends the request to the server, which then generates and saves the final combination result. This final model can then be downloaded by the user or integrated into other systems for use.

[1208] In this way, users can easily combine multiple pre-trained models to generate an AI model that suits their needs, enabling them to effectively utilize complex AI systems even without specialized knowledge.

[1209] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1210] Step 1:

[1211] The user accesses the device interface. The user accesses the system's graphical user interface using a PC or mobile device. Input can be a URL or an application launch operation. This launches the interface and displays the model selection screen.

[1212] Step 2:

[1213] The server sends a list of trained models to the device. The server accesses the database and obtains information about the stored trained models (e.g., image recognition models, speech recognition models). The input is a database query. The output is a list of models sent to the device.

[1214] Step 3:

[1215] The user selects the model they want to use from the interface. They select the model they want to use from multiple pre-trained models on their device screen, using checkboxes or drop-down menus. The input is the user's selection operation. The output is the information about the selected model sent to the server.

[1216] Step 4:

[1217] The device requests the user's selection from the server. The device requests the ID and name of the model selected by the user from the server. The input is the model information selected by the user. The output is that the request is sent to the server.

[1218] Step 5:

[1219] The server retrieves the selected model data from the database and sends it to the terminal. The server accesses the database and retrieves the data of the selected model. The input is the selected model information. The output is the corresponding model data sent to the terminal.

[1220] Step 6:

[1221] The terminal displays the model on a graphical user interface. The terminal interprets the model data it receives and displays it in the interface. As input, it has the model data. As output, it presents adjustment tools such as sliders and dials to the user.

[1222] Step 7:

[1223] The user adjusts the model parameters. The user adjusts the parameters of each model by operating sliders and dials. The input is the user's operation information. The output is the adjusted parameters sent to the terminal, which then sends them to the server.

[1224] Step 8:

[1225] The terminal sends the adjusted parameters to the server. The terminal immediately sends the parameters adjusted by the user to the server. The input is the adjusted parameter information. The output is the parameter information sent to the server.

[1226] Step 9:

[1227] The server adjusts the models based on the received parameters and generates intermediate outputs. The server adjusts the behavior of each model based on the adjusted parameters and combines them to generate intermediate outputs. The input is the adjusted parameter information. The output is intermediate output data.

[1228] Step 10:

[1229] The server generates intermediate output and sends it to the terminal. The server generates intermediate output and sends it to the terminal. As input, there is intermediate output data. As output, that data is sent to the terminal.

[1230] Step 11:

[1231] The terminal displays the intermediate output to the user in real time. The terminal displays the intermediate output received from the server to the user in real time. As input, there is intermediate output data. As output, there is intermediate output displayed on the interface.

[1232] Step 12:

[1233] The user fine-tunes the parameters and determines the final model. The user checks the intermediate outputs and fine-tunes the parameters as needed to determine a satisfactory final model. The inputs are the intermediate outputs and user operation information. The output is the final parameter settings.

[1234] Step 13:

[1235] The terminal sends the final parameters to the server. The terminal sends the final parameters determined by the user to the server. As input, there are the final parameter settings. As output, that information is sent to the server.

[1236] Step 14:

[1237] The server generates the final model and stores it in a database or storage device. The server generates the final model based on the final parameters and stores it in a database or storage device. The input is the final parameter settings. The output is the final model data and stores it in a database or storage device.

[1238] Step 15:

[1239] The user downloads the final model or integrates it into another system for use. The user downloads the generated final model via a terminal or integrates it with another system. The inputs are the final model data and the user's download operation. The outputs are the model data stored in the user's environment and the integrated system.

[1240] (Application example 1)

[1241] Next, a description will be given of Application 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."

[1242] In conventional factory robot systems using AI models, optimizing each task required a lot of time and specialized knowledge. Furthermore, there was a lack of means to intuitively combine and use different types of trained models, making it difficult for users to adjust parameters and check the results in real time. To solve this problem, there is a need for a system that efficiently and intuitively combines multiple trained models to optimize the operation of factory robots.

[1243] 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.

[1244] In this invention, the server includes means for loading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for transmitting and displaying the intermediate output to a terminal in the factory in real time, and means for generating and saving a final model. This enables users to intuitively use multiple trained models to optimize the operation of factory robots and effectively utilize complex AI systems without specialized knowledge.

[1245] A "trained model" is an artificial intelligence algorithm that has been pre-trained for a specific task.

[1246] A "database" is an information system that can efficiently manage large amounts of data and search and update it.

[1247] "Storage" is a physical and logical memory device for storing data.

[1248] A "graphical user interface" is an interface that uses visual elements to allow users to operate it intuitively.

[1249] A "parameter" is a setting or variable that affects the behavior or performance of a model.

[1250] "Intermediate output" is a temporary result generated after combining and adjusting multiple trained models.

[1251] "Real time" refers to an operation that responds immediately to a user's operation and displays the result instantly.

[1252] A "factory robot" is an automated mechanical device that performs physical tasks in a factory.

[1253] A "specific task" refers to a specific task or activity that a robot performs within a factory.

[1254] The "final model" is the final AI model that the user has optimized through parameter adjustment and decided to save.

[1255] A "terminal" is a digital device that a user uses to perform operations and check results.

[1256] This invention is a system for generating custom AI models to optimize the operation of robots used in factories. This system generates a unique learning model by efficiently and intuitively combining multiple trained models, and is composed of a server, a terminal, and a user.

[1257] The server runs on AWS EC2 and is configured with a web framework using Django. First, the server loads multiple selected trained models from a database (AWS RDS, PostgreSQL). The loaded model data is sent to the terminal, where it can be operated on the terminal.

[1258] The terminals are iOS / Android devices (smartphones and tablets) that display an intuitive graphical user interface using React Native, which includes sliders and dials similar to a music equalizer, allowing users to adjust the parameters of each model.

[1259] The user adjusts the parameters of each model using the device's interface. For example, they can select an object recognition model and a movement pattern model and set the object recognition accuracy to 90% and the movement reaction speed to 0.5 seconds. Each time an adjustment is made, the device sends the parameters to the server in real time. The server adjusts the behavior of each model based on the received parameters and combines multiple models to generate intermediate output. This intermediate output is then sent back to the device and displayed to the user in real time.

[1260] Once the user has further fine-tuned the parameters and is satisfied with the output results, they press the save button to generate the final model. The device then sends the request to the server, which generates the final combination results and stores them in a database or storage. The final model is then downloaded to the robots in the factory and used to perform specific tasks.

[1261] For example, a prompt to a generative AI model might look something like this:

[1262] You need to generate an AI model to deploy optimal robot behavior for specific tasks in your factory. Adjust the AI ​​model to meet the following requirements:

[1263] Object recognition accuracy: 90% or more

[1264] Response time: within 0.5 seconds

[1265] Efficiency of movement patterns: Minimal route and obstacle avoidance

[1266] Use the sliders to adjust these parameters and see the results in real time.

[1267] In this way, users can easily combine multiple pre-trained models to generate custom AI models suited to tasks within their factories, enabling them to effectively utilize complex AI systems without specialized knowledge.

[1268] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1269] Step 1:

[1270] A user accesses the system using a terminal and requests a list of trained models for a specific task of a factory robot. The terminal sends the request to the server. The input to this request is the name of the model selected by the user.

[1271] Step 2:

[1272] The server receives the request, searches for the corresponding trained model in a database (e.g., PostgreSQL on AWS RDS), and retrieves the model data. In this data processing stage, a database query is executed based on the model name to load the appropriate model data. The loaded model data is obtained as output.

[1273] Step 3:

[1274] The server transmits the acquired model data to the terminal, which analyzes the received model data and displays an intuitive graphical user interface that includes sliders and dials that respond in real time.

[1275] Step 4:

[1276] The user adjusts the parameters of each model using the device interface. For example, the accuracy of the object recognition model is set to 90% and the response speed of the movement pattern model is set to 0.5 seconds. These adjustments are sent as data from the device to the server.

[1277] Step 5:

[1278] The server adjusts the behavior of each model based on the received parameters, combines multiple models, and generates intermediate output. Based on the parameter input, it performs data calculations to optimize the internal parameters of the model, and generates an adjusted intermediate model. This intermediate output is obtained as the output.

[1279] Step 6:

[1280] The intermediate output is sent from the server to the terminal, which displays it in real time to the user, who can then check the displayed results and further fine-tune the parameters if necessary. This process can be repeated.

[1281] Step 7:

[1282] When the user is satisfied with the output result, he / she presses the save button to generate the final model. The terminal sends this operation to the server, which then generates the final model. At this stage, the final result of the parameter settings that the user is satisfied with is obtained as input.

[1283] Step 8:

[1284] The server generates the final combination results and saves them in a database or storage. The final model is then downloaded and applied to the robots in the factory, resulting in optimal robot behavior for the specific tasks in the factory.

[1285] Furthermore, 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.

[1286] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[1287] The server has the function of loading multiple trained models selected by the user from a database or storage and sending them to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user.

[1288] The device analyzes the model data sent from the server and provides the user with an interface that includes sliders and dials similar to a music equalizer to adjust the parameters of each model. The device also includes an emotion engine that recognizes emotions by analyzing the user's facial expressions, voice, and typing patterns.

[1289] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, and the server adjusts the model based on the adjusted parameters and combines them to generate intermediate output. The generated intermediate output is then sent back to the device and displayed to the user in real time.

[1290] The emotion engine analyzes facial expressions, voice, and typing patterns when a user operates the interface to recognize the user's emotional state. Based on this recognized emotion, the emotion engine recommends parameters appropriate for the user. It is also possible to automatically customize the interface according to the user's emotions.

[1291] As a concrete example, consider the case where a user wants to combine an image recognition model and a voice recognition model. The user first selects an "image recognition model" and a "voice recognition model" from the device interface. The device then requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[1292] Next, the user uses the sliders on the interface to set "image recognition accuracy" to 80% and "voice recognition response speed" to 90%. At this point, the emotion engine recognizes the user's emotions and recommends parameter settings that are easier to use if the user is nervous, for example. Alternatively, the interface layout and design can be changed depending on the user's emotions.

[1293] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user, who can check the results and further fine-tune the parameters.

[1294] When the user finally obtains the desired result, they press the "Save" button to generate the final model. The terminal sends the request to the server, which generates the final combination result and saves it in a database or storage. The user can download this final model or integrate it into other systems for use.

[1295] In this way, the present invention makes it possible to effectively combine multiple trained models while taking into account the user's emotional state, thereby intuitively and efficiently generating a unique AI model.

[1296] The processing flow will be explained below.

[1297] Step 1:

[1298] A user logs in to the system and accesses a screen that displays a list of trained models.

[1299] Step 2:

[1300] The terminal sends the user's login information to the server, which then performs authentication.

[1301] Step 3:

[1302] The server retrieves a list of available trained models from the database and sends it to the terminal.

[1303] Step 4:

[1304] The device displays a list of trained models received to the user.

[1305] Step 5:

[1306] The user selects the trained model they want to use and sends that information to their device.

[1307] Step 6:

[1308] The device sends the user's selection to the server, which then loads the corresponding trained model from the database.

[1309] Step 7:

[1310] The server sends the trained model it has loaded to the terminal.

[1311] Step 8:

[1312] The device analyzes the received trained model and generates a graphical user interface, which includes sliders and dials similar to a music equalizer.

[1313] Step 9:

[1314] The user adjusts the parameters of each trained model using a graphical user interface.

[1315] Step 10:

[1316] The emotion engine analyzes the user's facial expressions, voice, and typing patterns in real time while they are operating the device, recognizing their emotional state.

[1317] Step 11:

[1318] The device recommends appropriate parameter settings and customizes the interface based on the user's emotions recognized by the emotion engine.

[1319] Step 12:

[1320] The device receives the user's adjustments and transmits them to the server in real time.

[1321] Step 13:

[1322] Based on the parameters received by the server, each trained model is adjusted and combined to generate an intermediate output.

[1323] Step 14:

[1324] The server generates intermediate output and sends it back to the terminal.

[1325] Step 15:

[1326] The intermediate output received by the terminal is displayed to the user in real time.

[1327] Step 16:

[1328] The user checks the intermediate output and makes further fine adjustments to the parameters.

[1329] Step 17:

[1330] The terminal sends the readjusted parameters to the server, and the server again generates an intermediate output and sends it back to the terminal.

[1331] Step 18:

[1332] When the user finally obtains a satisfactory output result, he or she presses the "Save" button.

[1333] Step 19:

[1334] The terminal sends a "save" request to the server, which generates the final model.

[1335] Step 20:

[1336] The server saves the generated final model in a database or storage and transmits the information to the terminal.

[1337] Step 21:

[1338] The device will notify the user that the final model has been saved and provide a download link.

[1339] Step 22:

[1340] Users can download the final model and integrate it into other systems or use it on their own platforms.

[1341] Example 2

[1342] Next, a description will be given of 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."

[1343] Conventional AI model generation systems have difficulty efficiently and intuitively combining multiple trained models, placing a heavy burden on users. Furthermore, they are unable to adjust parameters or customize interfaces that take the user's emotional state into account, resulting in a lack of improvement in the user experience.

[1344] The identification process by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating an intermediate output, means for incorporating an emotion engine that recognizes the user's emotional state and recommends parameters, means for automatically customizing the interface according to the user's emotional state, and means for generating and saving a final model. This makes it possible to intuitively and efficiently combine multiple trained models while taking the user's emotional state into consideration to generate a unique AI model.

[1345] A "pre-trained model" is an artificial intelligence model that has already been trained and optimized to perform a specific task.

[1346] A "database" is an electronic system for efficiently storing and retrieving data.

[1347] "Storage" is a physical or virtual space for storing digital data.

[1348] A "graphical user interface" is an interface that includes elements (such as buttons, sliders, and dials) that a user can visually manipulate.

[1349] A "parameter" is a setting value or variable that controls the behavior of an AI model.

[1350] An "emotion engine" is an artificial intelligence engine that identifies a user's emotions and provides appropriate responses and recommendations.

[1351] "Intermediate output" refers to an intermediate result generated before obtaining the final output.

[1352] The "final model" is the AI ​​model that is completed as a result of adjustments and combinations.

[1353] This invention is a system that generates a unique learning model by efficiently and intuitively combining multiple trained AI models and further combining them with an emotion engine that recognizes user emotions. This system consists of components such as a server, a terminal, and a user. The user adjusts the parameters of each model using a graphical user interface, and the emotion engine recognizes the user's emotions and recommends parameters and customizes the interface.

[1354] Hardware and Software Configuration

[1355] server

[1356] The server functions as a means of loading multiple trained models from a database or storage (for example, MySQL or PostgreSQL is used as database software), sending the model data selected by the user to the terminal, and reconstructing the model based on the adjusted parameters.

[1357] Terminal

[1358] The terminal is a means for analyzing the model data sent from the server. The terminal provides a user-operable graphical user interface (GUI). This GUI includes sliders and dials that the user uses to adjust the model parameters. The terminal also includes an emotion engine that uses OpenCV and TensorFlow to recognize emotions by analyzing the user's facial expressions, voice, and typing patterns.

[1359] Emotion Engine

[1360] The emotion engine is a means of recognizing the user's emotions. It is possible to use PyTorch or Keras as a deep learning framework. This emotion engine has the function of recommending appropriate parameters for the user based on the analyzed emotional state and automatically customizing the interface.

[1361] Specific examples

[1362] For example, if a user wants to combine an image recognition model and a voice recognition model, they can use the system by following the steps below.

[1363] 1. The user selects the "image recognition model" and "voice recognition model" from the device interface.

[1364] 2. The device requests the selection from the server, and the server reads the corresponding model data from the database and sends it to the device.

[1365] 3. The user uses the sliders on the interface to set "Image Recognition Accuracy" to 80% and "Voice Recognition Response Speed" to 90%.

[1366] 4. The emotion engine analyzes the user's facial expressions and voice and recommends easier parameter settings if the user is nervous. It can also change the layout and design of the interface according to the user's emotions.

[1367] The parameters adjusted by the user are sent to the server in real time, and the server adjusts and combines the model based on the parameters to generate intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The user can check the results and further fine-tune the parameters.

[1368] Example of input prompt for generative AI model

[1369] 1. User prompt: "I want to combine image and speech recognition models, with image recognition accuracy set to 80% and speech recognition response speed set to 90%."

[1370] 2. Prompt for the generative AI model: "Please combine the image recognition and voice recognition models under the following conditions: Image recognition accuracy: 80%, Voice recognition response speed: 90%. The emotion engine will analyze the user's emotions and recommend suitable parameters."

[1371] This system allows users to generate their own unique AI models by efficiently combining multiple trained models and intuitively operating them while taking into account their own emotional state.

[1372] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1373] Step 1:

[1374] Initialization and model selection

[1375] input:

[1376] A user starts up a device and sends a request to display a list of available trained models.

[1377] output:

[1378] A list of available trained models will be displayed on your device.

[1379] Specific behavior:

[1380] The device initializes the system and requests a list of trained models from the server. The server retrieves the model list from a database (e.g., MySQL or PostgreSQL) and sends it to the device. The device parses it and displays it to the user.

[1381] Step 2:

[1382] Model selection and data loading

[1383] input:

[1384] The user selects the "image recognition model" and "voice recognition model" on the device interface.

[1385] output:

[1386] The selected model data is sent from the server to the terminal.

[1387] Specific behavior:

[1388] Based on the user's operation, the device sends a request for the selected model (e.g., "image recognition model" and "voice recognition model") to the server. The server reads the corresponding model data from the database and sends it to the device.

[1389] Step 3:

[1390] Interface display and parameter adjustment

[1391] input:

[1392] The terminal receives the model data sent from the server.

[1393] output:

[1394] An interface is displayed that allows the user to adjust the parameters.

[1395] Specific behavior:

[1396] The device analyzes the received model data and provides the user with a graphical user interface (GUI) for adjusting the model parameters. It displays UI elements such as sliders and dials that the user can use to adjust the parameters.

[1397] Step 4:

[1398] Parameter submission and model adjustment

[1399] input:

[1400] The user sets parameters on the interface (e.g., "image recognition accuracy" at 80%, "voice recognition response speed" at 90%).

[1401] output:

[1402] The adjusted parameters are sent to the server.

[1403] Specific behavior:

[1404] The parameters set by the user are sent in real time from the device to the server, which then adjusts and combines the selected model based on the received parameters.

[1405] Step 5:

[1406] Generating and displaying intermediate output

[1407] input:

[1408] The server receives the adjusted parameters.

[1409] output:

[1410] An intermediate output is generated and sent to the terminal.

[1411] Specific behavior:

[1412] The server adjusts the AI ​​model based on the adjusted parameters and generates intermediate output, which is sent to the device in JSON format, etc. The device analyzes the received intermediate output and displays it to the user in real time.

[1413] Step 6:

[1414] Analysis and recommendation by emotion engine

[1415] input:

[1416] Emotional data such as the user's facial expression, voice, and typing pattern are input.

[1417] output:

[1418] Emotion-based parameter recommendations and interface customization are performed.

[1419] Specific behavior:

[1420] The device's built-in emotion engine (using OpenCV and TensorFlow) analyzes the user's facial expressions and voice to recognize their emotional state at that time. Based on the recognition results, the emotion engine recommends parameter settings appropriate for the user and customizes the interface.

[1421] Step 7:

[1422] Save and download the final model

[1423] input:

[1424] When the user obtains a satisfactory output result, he or she presses the "save" button.

[1425] output:

[1426] The final model is saved to a database or storage and a download link is provided.

[1427] Specific behavior:

[1428] When the user presses the "Save" button, the device sends a request to the server to save the final model. The server generates the final model and saves it in a database or storage. The server then returns a download link for the final model to the device, and the user clicks the link to download the final model.

[1429] These processing steps enable users to efficiently and intuitively combine multiple pre-trained AI models to generate their own unique AI models, and also use the emotion engine to recommend parameters and customize the interface.

[1430] (Application example 2)

[1431] Next, a description will be given of Application 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."

[1432] In recent years, advances in AI technology have created a demand for combining and using multiple trained models. However, there is no efficient and intuitive way to combine these models. Furthermore, few systems take the user's emotional state into account, and parameter settings and interface customization are cumbersome. There is a need for a system that can solve these issues and provide a better user experience based on the user's emotions.

[1433] The identification processing by the identification processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for reading multiple trained models from a database or storage, means for adjusting parameters of the multiple trained models via a graphical user interface, means for combining the multiple trained models based on the adjusted parameters and generating intermediate output, means for recommending parameters based on the user's emotional state, and means for customizing the interface according to the user's emotional state. This makes it possible to efficiently and intuitively combine multiple trained models while taking the user's emotional state into consideration.

[1434] A "trained model" is a machine learning algorithm that has been trained in advance using large amounts of data and has the ability to perform a specific task.

[1435] A "database" is a system that organizes and stores data so that it can be accessed and manipulated as needed.

[1436] "Storage" refers to a storage device for long-term storage of data.

[1437] A "graphical user interface" refers to an interface that a user can visually interact with, typically including graphical elements such as windows, buttons, sliders, etc.

[1438] "Parameters" refer to the settings that control the performance and behavior of a model.

[1439] "Intermediate output" refers to the intermediate results obtained when combining multiple trained models.

[1440] "Real-time" means that operations or processes occur immediately, without delay.

[1441] The "final model" refers to the machine learning model that is ultimately created based on the output results desired by the user.

[1442] "Emotional state" refers to a user's current emotional or psychological state.

[1443] "Parameter recommendation" refers to the act of the system suggesting appropriate setting values ​​for parameters set by the user.

[1444] "Interface customization" refers to changing the design and layout of an interface in response to a user's specific requirements or emotional state.

[1445] This invention is a system that efficiently and intuitively combines multiple trained models, recommends parameters taking into account the user's emotional state, and customizes the interface. It is composed of a server, a terminal, and a user. To realize this system, the following processes are performed.

[1446] server

[1447] The server has the function of loading multiple trained models from a database or storage. It loads the model data selected by the user and sends it to the terminal. The loaded model data is analyzed on the terminal and becomes operable by the user. It also combines models based on the user's parameter adjustment requests, generates intermediate output, and sends it back to the terminal.

[1448] Terminal

[1449] The device analyzes the model data sent from the server and provides the user with a graphical user interface, which includes sliders and dials similar to a music equalizer to adjust the parameters of each model.The device also includes an emotion engine that analyzes the user's facial expressions, voice, and typing patterns to recognize their emotional state.

[1450] User

[1451] The user adjusts the parameters of each model using the interface on their device. The adjusted parameters are sent to the server in real time, which then adjusts and combines the models based on the parameters to generate an intermediate output. This intermediate output is sent back to the device and displayed to the user in real time. The emotion engine analyzes the user's emotional state and recommends optimal parameters based on that state. The interface layout and design can also be customized based on the emotional state. When the user is satisfied with the output results, they press the "Save" button to generate the final model. The device sends this request to the server, which generates the final combination results and stores them in a database or storage.

[1452] Specific examples

[1453] For example, consider the use of this technology in a "personalized video viewing application." This application recommends the most suitable videos based on the user's emotional state, improving the viewing experience. When a user requests videos of a specific genre or theme, the emotion engine analyzes the user's emotions and generates a list of recommended videos based on that. The interface also changes dynamically depending on the user's emotional state. Specifically, the following prompts are input to the generative AI model:

[1454] Example prompt sentence:

[1455] Read the user's facial expressions and recommend the following list of movies and documentaries.

[1456] 1. Inspirational Movies

[1457] 2. An interesting documentary

[1458] This system can understand the user's emotional state and provide viewing content that is more optimal for each individual.

[1459] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1460] Step 1:

[1461] The server reads multiple trained models selected by the user from the database. It receives the user's selection, retrieves the model data from the database, and sends it to the device. The input is the user's model selection information, and the output is the model data.

[1462] Step 2:

[1463] The terminal analyzes the model data sent from the server, converts the acquired model data into an appropriate format, and makes it operable by the user. The input is the model data from the server, and the output is the model data converted into a format that can be operated by the user.

[1464] Step 3:

[1465] The terminal provides a graphical user interface, which includes sliders and dials that the user can manipulate directly to adjust the parameters of each model. The input is the converted model data, and the output is a parameter setting screen that can be manipulated in the interface.

[1466] Step 4:

[1467] The user adjusts the parameters of each model using an interface. The emotion engine analyzes the user's facial expressions, voice, and typing patterns to recognize the user's emotional state at that time. The input is the user's actions and emotional data, and the output is the adjusted parameters.

[1468] Step 5:

[1469] The server receives the parameters adjusted by the user, combines multiple trained models based on them, and generates intermediate outputs. The inputs are the parameters adjusted by the user, and the outputs are the intermediate outputs.

[1470] Step 6:

[1471] The terminal displays the intermediate output sent from the server in real time. The interface and intermediate output are updated every time the user changes a parameter. The input is the intermediate output from the server, and the output is the updated interface and display of the intermediate output.

[1472] Step 7:

[1473] The emotion engine recommends optimal parameters based on the user's current emotional state. As the user adjusts parameters, the emotion engine analyzes them in real time and suggests appropriate settings. The input is the user's emotional data, and the output is the recommended parameter settings.

[1474] Step 8:

[1475] The device customizes the interface based on the analysis results of the emotion engine. The color and layout of the interface are changed according to the emotional state. The input is the analyzed emotional data, and the output is a customized interface.

[1476] Step 9:

[1477] When the user is satisfied with the output results, they can click the "Save" button to generate and save the final model. The input is the user's save request, and the output is the generated final model and its saving.

[1478] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding 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.

[1479] 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> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with 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.

[1480] 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 robot 414.

[1481] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1482] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1483] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1484] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1485] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1486] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1487] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1488] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1489] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1490] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1491] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1492] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1493] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1494] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1495] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1496] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1497] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1498] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1499] The following is further disclosed regarding the above embodiment.

[1500] (Claim 1)

[1501] A means of loading multiple trained models from a database or storage;

[1502] means for adjusting parameters of the plurality of trained models via a graphical user interface;

[1503] a means for combining multiple trained models based on the adjusted parameters to generate intermediate outputs;

[1504] a means for displaying the intermediate output in real time;

[1505] A means to generate and save the final model

[1506] A system including:

[1507] (Claim 2)

[1508] 10. The system of claim 1, wherein the means for displaying the intermediate output in real time is updated each time a user changes a parameter.

[1509] (Claim 3)

[1510] 2. The system according to claim 1, wherein the means for generating and saving the final model operates in response to the user pressing a "Save" button when a satisfactory output result is obtained.

[1511] "Example 1"

[1512] (Claim 1)

[1513] means for loading a plurality of trained models from a database or storage device;

[1514] means for adjusting parameters of the plurality of trained models via a graphical user interface;

[1515] means for combining the plurality of trained models based on the adjusted parameters to generate an intermediate output;

[1516] means for immediately displaying intermediate output;

[1517] A means to generate and save the final model

[1518] A system including:

[1519] (Claim 2)

[1520] 2. The system of claim 1, wherein the means for real-time display of intermediate output is updated each time a user changes a parameter.

[1521] (Claim 3)

[1522] 2. The system according to claim 1, wherein the means for generating and saving the final model operates in response to the user pressing a "Save" button when a satisfactory output result is obtained.

[1523] "Application Example 1"

[1524] (Claim 1)

[1525] A means of loading multiple trained models from a database or storage;

[1526] means for adjusting parameters of the plurality of trained models via a graphical user interface;

[1527] a means for combining multiple trained models based on the adjusted parameters to generate intermediate outputs;

[1528] a means for displaying the intermediate output in real time;

[1529] a means for generating and saving the final model;

[1530] a means for adjusting parameters to provide optimal robot behavior for specific tasks within a factory;

[1531] means for transmitting and displaying intermediate outputs in real time on a terminal in the factory based on the adjusted robot operation;

[1532] A system including:

[1533] (Claim 2)

[1534] 10. The system of claim 1, wherein the means for displaying the intermediate output in real time is updated each time a user changes a parameter.

[1535] (Claim 3)

[1536] 2. The system according to claim 1, wherein the means for generating and saving the final model operates in response to the user pressing a "Save" button when a satisfactory output result is obtained.

[1537] "Example 2: Combining Emotion Engines"

[1538] (Claim 1)

[1539] A means of loading multiple trained models from a database or storage;

[1540] means for adjusting parameters of the plurality of trained models via a graphical user interface;

[1541] a means for combining multiple trained models based on the adjusted parameters to generate intermediate outputs;

[1542] a means for displaying the intermediate output in real time;

[1543] means for incorporating an emotion engine that recognizes the user's emotional state and recommends parameters;

[1544] means for automatically customizing the interface according to the user's emotional state;

[1545] A means to generate and save the final model

[1546] A system including:

[1547] (Claim 2)

[1548] 10. The system of claim 1, wherein the means for displaying the intermediate output in real time is updated each time a user changes a parameter.

[1549] (Claim 3)

[1550] 2. The system according to claim 1, wherein the means for generating and saving the final model operates in response to the user pressing a "Save" button when a satisfactory output result is obtained.

[1551] "Application example 2 when combining emotion engines"

[1552] (Claim 1)

[1553] A means of loading multiple trained models from a database or storage;

[1554] means for adjusting parameters of the plurality of trained models via a graphical user interface;

[1555] a means for combining multiple trained models based on the adjusted parameters to generate intermediate outputs;

[1556] a means for displaying the intermediate output in real time;

[1557] a means for generating and saving the final model;

[1558] means for recommending parameters based on the emotional state of the user;

[1559] A system including means for customizing an interface according to a user's emotional state.

[1560] (Claim 2)

[1561] 10. The system of claim 1, wherein the intermediate output is updated in real time whenever a user changes a parameter.

[1562] (Claim 3)

[1563] The system according to claim 1, wherein a final model is generated and saved in response to the user pressing a "save" button when a satisfactory output result is obtained. [Explanation of symbols]

[1564] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of loading multiple trained models from a database or storage; means for adjusting parameters of the plurality of trained models via a graphical user interface; a means for combining multiple trained models based on the adjusted parameters to generate intermediate outputs; a means for displaying the intermediate output in real time; A means to generate and save the final model A system including:

2. 2. The system of claim 1, wherein the means for displaying the intermediate output in real time is updated each time a user changes a parameter.

3. 2. The system according to claim 1, wherein the means for generating and saving the final model operates in response to a user pressing a "save" button when a satisfactory output result is obtained.

Images